<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:googleplay="http://www.google.com/schemas/play-podcasts/1.0"><channel><title><![CDATA[PHUONGLIB: Business Research]]></title><description><![CDATA[A complete series on quantitative data analysis for theses and journal articles, from research design and sample size to SPSS and PSPP, CB-SEM in JASP and AMOS and PLS-SEM in SmartPLS, with sample data on GitHub.]]></description><link>https://www.phuonglib.com/s/business-research</link><image><url>https://substackcdn.com/image/fetch/$s_!tUFo!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F68ab7f3d-2ea9-431f-a018-11cffe5264ac_1000x1000.png</url><title>PHUONGLIB: Business Research</title><link>https://www.phuonglib.com/s/business-research</link></image><generator>Substack</generator><lastBuildDate>Sun, 22 Nov 2026 21:57:11 GMT</lastBuildDate><atom:link href="https://www.phuonglib.com/feed" rel="self" type="application/rss+xml"/><copyright><![CDATA[Phuong Nguyen]]></copyright><language><![CDATA[en]]></language><webMaster><![CDATA[phuonglib@substack.com]]></webMaster><itunes:owner><itunes:email><![CDATA[phuonglib@substack.com]]></itunes:email><itunes:name><![CDATA[PHUONGLIB]]></itunes:name></itunes:owner><itunes:author><![CDATA[PHUONGLIB]]></itunes:author><googleplay:owner><![CDATA[phuonglib@substack.com]]></googleplay:owner><googleplay:email><![CDATA[phuonglib@substack.com]]></googleplay:email><googleplay:author><![CDATA[PHUONGLIB]]></googleplay:author><itunes:block><![CDATA[Yes]]></itunes:block><item><title><![CDATA[IPMA and Higher-Order Constructs in SmartPLS 4]]></title><description><![CDATA[Use an importance-performance map to find which factors to prioritize, then build and evaluate a reflective-reflective higher-order construct with repeated indicators]]></description><link>https://www.phuonglib.com/p/ipma-higher-order-constructs-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/ipma-higher-order-constructs-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:50:57 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ygf-!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ygf-!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!ygf-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:1016154,&quot;alt&quot;:&quot;A gardener with an orange watering can choosing among four young trees, behind them one large tree whose canopy grows from two trunks.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A gardener with an orange watering can choosing among four young trees, behind them one large tree whose canopy grows from two trunks." title="A gardener with an orange watering can choosing among four young trees, behind them one large tree whose canopy grows from two trunks." srcset="https://substackcdn.com/image/fetch/$s_!ygf-!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!ygf-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F60ec38a3-89e4-4a53-a529-4673de447e50_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>IPMA combines importance (total effect on the target construct) with performance (the mean score rescaled to 0 to 100) to show which factor to prioritize.</p></li><li><p>In the Marketplace X model, confirmation of expectations has the highest importance (0.312) and the lowest performance (72.08), with item CONF4 as the item to prioritize.</p></li><li><p>IPMA uses total effects, so a factor whose direct effect is not significant can still be the most important; read the scale of the vertical axis before comparing performance.</p></li><li><p>In a reflective-reflective higher-order construct, the paths to the lower-order constructs are read as loadings; here AVE is 0.778 and rho_c 0.875.</p></li><li><p>A higher-order construct needs a theoretical basis established beforehand, not just a wish for a more parsimonious model.</p></li></ul></blockquote><div><hr></div><p>This is the last post of the 42-post series. It introduces two extensions of PLS-SEM (Hair et al., 2022):</p><ul><li><p><strong>IPMA</strong> (importance-performance map analysis): turns the model results into <strong>managerial recommendations</strong>, that is, which factors to prioritize to increase repurchase intention.</p></li><li><p><strong>Higher-order construct model</strong> (HCM): a construct formed from other constructs, instead of being measured directly by items.</p></li></ul><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch42-ipma-higher-order-constructs-smartpls">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code> (IPMA part) and <code>hcm-model</code>, a copy edited into a higher-order model (HCM part). Data: casewise deletion, 254 respondents. According to the SmartPLS license comparison page, the Student license does not include IPMA. The numbers in this post come from the SmartPLS reports; R (scripts <code>ipma.R</code> and <code>higher_order.R</code> in the repository's <code>scripts/</code> folder, package cSEM) gives the same results to the third decimal place, except where noted. SmartPLS lists constructs and paths in alphabetical order (CONFIRM, INTENT, SATISF, TRUST, VALUE).</p></blockquote><h2>Part 1. IPMA</h2><h3>1. The idea</h3><p>Standard PLS-SEM results tell you the <strong>importance</strong> of each factor: its total effect on the target construct. IPMA adds a second dimension, <strong>performance</strong>: the mean score of that factor, rescaled to a range of 0 to 100 (Ringle &amp; Sarstedt, 2016).</p><p>Putting the two dimensions together gives a map: the horizontal axis is importance, the vertical axis is performance. A factor in the <strong>high importance, low performance</strong> corner is the one to prioritize: improving it has the largest effect on the target construct, and it has the most room for improvement.</p><p><strong>Data requirements</strong> (Ringle &amp; Sarstedt, 2016):</p><ol><li><p>All items point in the <strong>same direction</strong>: a high score is good. Reverse-worded items must be recoded first (TRU4R in this series was recoded in <a href="https://www.phuonglib.com/p/recode-compute-composite-scores-pspp-spss">Post 10</a>).</p></li><li><p>Items are measured on a <strong>metric or quasi-metric</strong> scale (a 1 to 5 Likert scale here).</p></li><li><p>All outer weights are <strong>positive</strong>.</p></li></ol><h3>2. How performance is computed</h3><p><strong>Step 1. Rescale each item to 0 to 100:</strong> new value = (value - 1) / (5 - 1) x 100. For example, the mean of CONF2 is 3.921, so the performance of CONF2 is (3.921 - 1) / 4 x 100 = <strong>73.0</strong>.</p><p><strong>Step 2. Rescale the outer weights so that they sum to 1</strong> within each construct (rescaled outer weights). For CONFIRM: 0.246 + 0.276 + 0.204 + 0.274 = 1.000.</p><p><strong>Step 3. Compute the performance of the construct</strong> = sum of (rescaled weight x item performance). For CONFIRM: 0.246 x 71.85 + 0.276 x 73.03 + 0.204 x 72.05 + 0.274 x 71.36 = <strong>72.08</strong>.</p><h3>3. Running IPMA in SmartPLS 4</h3><p><strong>Step 1.</strong> Open the model <code>marketplace-x-model</code> and choose <strong>Calculate &gt; Importance-performance map analysis (IPMA)</strong>.</p><p><strong>Step 2.</strong> On the <strong>IPMA setup</strong> tab (Figure 1):</p><ul><li><p>Set <strong>Target construct</strong> to <strong>INTENT</strong>. Until a target is selected, the Start calculation button is disabled and the dialog shows "A target construct needs to be selected".</p></li><li><p>Keep <strong>IPMA results</strong> at <strong>All predecessors of the selected target construct</strong>, to get results for every construct that precedes INTENT.</p></li><li><p>Check the <strong>Indicator scales</strong> table: <strong>Scale min</strong> and <strong>Scale max</strong> are 1 and 5 for every item. If the actual scale differs from the observed minimum and maximum (for example, nobody chose 1), correct the scale with the link "adjust these scales in the data file".</p></li></ul><p>Check that the <strong>Data</strong> tab is set to <strong>Casewise deletion</strong>, then click <strong>Start calculation</strong>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!sfki!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!sfki!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!sfki!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!sfki!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!sfki!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!sfki!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:196591,&quot;alt&quot;:&quot;Figure 1. IPMA dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. IPMA dialog" title="Figure 1. IPMA dialog" srcset="https://substackcdn.com/image/fetch/$s_!sfki!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!sfki!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!sfki!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!sfki!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c6c5e1f-ac1c-4abd-ac79-cfa37de52f0d_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. IPMA dialog: (1) target construct INTENT, (2) results for all predecessors, (3) the indicator scales table, (4) the Start calculation button.</figcaption></figure></div><h3>4. Reading the construct-level results</h3><p><strong>Step 3.</strong> The results graph (Figure 2) shows the <strong>performance</strong> inside each circle and the <strong>rescaled outer weight</strong> on each line to an item. The <strong>Graphical output</strong> panel at the bottom left shows what the graph is displaying (Rescaled outer weights, LV performance).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!EdgV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!EdgV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!EdgV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png" width="1100" height="486" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:486,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:167336,&quot;alt&quot;:&quot;Figure 2. IPMA results graph&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. IPMA results graph" title="Figure 2. IPMA results graph" srcset="https://substackcdn.com/image/fetch/$s_!EdgV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!EdgV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4be8b9b8-9182-4abb-9e01-b3fd647f093c_1100x486.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. IPMA results graph: (1) items display the rescaled outer weights, (2) constructs display their performance, (3) the performance of INTENT, 71.443.</figcaption></figure></div><p><strong>Step 4.</strong> Open <strong>Quality criteria &gt; Importance-performance map [INTENT] (constructs) &gt; Importance-performance map</strong> (Figure 3), then open <strong>Construct total effects for [INTENT]</strong> to read the numbers (Figure 4).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!9Tyd!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!9Tyd!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!9Tyd!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png" width="1100" height="486" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:486,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:203038,&quot;alt&quot;:&quot;Figure 3. Construct-level IPMA map&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Construct-level IPMA map" title="Figure 3. Construct-level IPMA map" srcset="https://substackcdn.com/image/fetch/$s_!9Tyd!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!9Tyd!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0324fb1c-a0af-4ac4-8708-ee8599343757_1100x486.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. The SmartPLS IPMA map: (1) the item Importance-performance map, (2) CONFIRM, the highest importance, (3) SATISF, the lowest importance.</figcaption></figure></div><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!0ZDc!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 424w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 848w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 1272w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!0ZDc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png" width="1100" height="268" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:268,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:81471,&quot;alt&quot;:&quot;Figure 4. Importance of the constructs&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Importance of the constructs" title="Figure 4. Importance of the constructs" srcset="https://substackcdn.com/image/fetch/$s_!0ZDc!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 424w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 848w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 1272w, https://substackcdn.com/image/fetch/$s_!0ZDc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2891968-0648-4c95-8c51-2058fd6cbafe_1100x268.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 4. Importance (total effect on INTENT): (1) the item Construct total effects for [INTENT], (2) the total effects table.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/cbOfR/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0d176b61-6a88-4c3a-953a-5fa98eaec126_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c871a229-fc6f-4b88-b68f-54d8789d6144_1220x666.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 1. Importance and performance of the four constructs for INTENT&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/cbOfR/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>On the SmartPLS map the vertical axis runs from 0 to 100, so the four points lie almost on one horizontal line. Figure 5 zooms in on the vertical axis and adds two mean lines that divide the map into four zones.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Ze8e!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Ze8e!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Ze8e!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png" width="1900" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:126890,&quot;alt&quot;:&quot;Figure 5. Zoomed-in IPMA map&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Zoomed-in IPMA map" title="Figure 5. Zoomed-in IPMA map" srcset="https://substackcdn.com/image/fetch/$s_!Ze8e!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!Ze8e!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F570c73b9-93e2-4ad2-b4a9-cce3120b4718_1900x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. IPMA map of the four constructs, vertical axis zoomed in; the dashed lines are the mean importance (0.271) and the mean performance (72.44). Values from SmartPLS.</figcaption></figure></div><p><strong>Interpretation:</strong></p><ul><li><p><strong>Confirmation of expectations is the factor to prioritize.</strong> It has the highest importance (0.312) and the lowest performance (72.08). Following Ringle and Sarstedt (2016), with an unstandardized total effect of 0.298 (see the note below), a 1-point increase in the performance of CONFIRM (on the 0 to 100 scale) goes with an increase of about 0.30 points in the performance of INTENT, more than any other factor.</p></li><li><p><strong>The importance of CONFIRM is mostly indirect.</strong> The direct effect of CONFIRM on INTENT is not significant (<a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a>), but the total effect also includes the part that runs through VALUE and SATISF (<a href="https://www.phuonglib.com/p/mediation-analysis-smartpls">Post 40</a>). So CONFIRM is not important in terms of its direct effect, but it is the most important in terms of its total effect.</p></li><li><p><strong>The performance of the four factors is almost the same</strong> (72.08 to 72.79, a difference of less than 1 point on a 100-point scale). Such small differences should not be read as real differences; the priority order here comes mainly from importance.</p></li><li><p><strong>SATISF has the lowest importance</strong> (0.221), because it only has a direct effect on INTENT.</p></li></ul><blockquote><p><strong>Note on how importance is computed.</strong> SmartPLS 4.1.2.1 reports importance as the <strong>standardized total effect</strong>, which equals the Total effects table of Post 40 (CONFIRM 0.312; VALUE 0.270). Ringle and Sarstedt (2016) describe importance as the <strong>unstandardized</strong> total effect, computed on the 0 to 100 scores. Computed this way in R, it is 0.298 (CONFIRM), 0.257 (TRUST), 0.249 (VALUE) and 0.211 (SATISF), the same values as the <code>doIPMA</code> function of cSEM. Both approaches give <strong>the same ranking</strong>. When reporting, state the software and the type of total effect used.</p></blockquote><h3>5. Reading the indicator-level results</h3><p><strong>Step 5.</strong> Open <strong>Importance-performance map [INTENT] (indicators) &gt; Importance-performance map</strong> (Figure 6), then open <strong>Indicator total effects for [INTENT]</strong> (Figure 7).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!zCKF!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!zCKF!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!zCKF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png" width="1100" height="486" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:486,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:220556,&quot;alt&quot;:&quot;Figure 6. Indicator-level IPMA map&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Indicator-level IPMA map" title="Figure 6. Indicator-level IPMA map" srcset="https://substackcdn.com/image/fetch/$s_!zCKF!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 424w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 848w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 1272w, https://substackcdn.com/image/fetch/$s_!zCKF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4d7bd32-5dde-48fe-8d90-0970b42e9023_1100x486.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. IPMA map of the 16 items: (1) the item Importance-performance map for indicators, (2) CONF4, the highest importance, (3) SAT2, the lowest importance.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!6hfj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!6hfj!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 424w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 848w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 1272w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!6hfj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png" width="1000" height="489" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:489,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:146893,&quot;alt&quot;:&quot;Figure 7. Importance of the indicators&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Importance of the indicators" title="Figure 7. Importance of the indicators" srcset="https://substackcdn.com/image/fetch/$s_!6hfj!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 424w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 848w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 1272w, https://substackcdn.com/image/fetch/$s_!6hfj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11985da4-32d6-4b10-bc7b-1f4e50bfceec_1000x489.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Importance of each item: (1) the item Indicator total effects for [INTENT], (2) CONF4, 0.109.</figcaption></figure></div><p>Importance of an item = standardized outer weight of the item x importance of its construct. The R cross-check matches SmartPLS for all 16 items, for example CONF4: 0.349 x 0.312 = <strong>0.109</strong>.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/8X8yp/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/87a80a72-f802-4b05-b434-1701309e2c25_1220x620.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f1f29d3c-bb4b-4566-a250-25663ea4e264_1220x728.png&quot;,&quot;height&quot;:358,&quot;title&quot;:&quot;Table 2. Importance and performance of selected items&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/8X8yp/1/" width="730" height="358" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>CONF4</strong> is the item to prioritize: it has the highest importance and the lowest performance of the four CONFIRM items. In a real study, the wording of this item tells you which specific aspect of the experience falls short of customers' expectations.</p><h2>Part 2. Higher-order construct models</h2><h3>6. What is a higher-order construct?</h3><p>A <strong>higher-order construct</strong> (HOC) is measured through <strong>lower-order constructs</strong> (LOCs) rather than directly by items. For example, "service quality" may consist of the dimensions reliability, responsiveness and empathy, each measured by a few items.</p><p>Sarstedt et al. (2019) distinguish four types, according to the measurement of the lower-order constructs (first term) and the relationship between the higher-order and lower-order constructs (second term):</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/nnTwd/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a9c6132a-c161-4ea8-924f-5b873e7a44ac_1220x540.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/840115e0-9542-4feb-a37e-2eb5d13f9bb5_1220x660.png&quot;,&quot;height&quot;:283,&quot;title&quot;:&quot;Table 3. Four types of higher-order constructs (Sarstedt et al., 2019)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/nnTwd/1/" width="730" height="283" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Three estimation approaches are commonly used: <strong>repeated indicators</strong> (the HOC reuses all the items of its LOCs), <strong>embedded two-stage</strong> and <strong>disjoint two-stage</strong> (Becker et al., 2012; Sarstedt et al., 2019).</p><h3>7. The illustrative model</h3><p>To illustrate the technique, this post builds a reflective-reflective higher-order construct <strong>EVALUATION</strong> (Post-purchase evaluation) with two lower-order constructs, VALUE and SATISF (Figure 8). The new structure: CONFIRM to EVALUATION to INTENT, keeping the two paths CONFIRM to INTENT and TRUST to INTENT.</p><blockquote><p><strong>This is a technical example, not the theoretical model of the series.</strong> The Marketplace X model treats VALUE and SATISF as two separate constructs with a causal relationship (H3). Merging them into one higher-order construct removes hypothesis H3. In a real study, a higher-order construct needs a theoretical basis established beforehand, and the lower-order constructs must be dimensions of the same general concept.</p></blockquote><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!t2j4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!t2j4!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 424w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 848w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 1272w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!t2j4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png" width="1900" height="880" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:880,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:110712,&quot;alt&quot;:&quot;Figure 8. Illustrative higher-order model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Illustrative higher-order model" title="Figure 8. Illustrative higher-order model" srcset="https://substackcdn.com/image/fetch/$s_!t2j4!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 424w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 848w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 1272w, https://substackcdn.com/image/fetch/$s_!t2j4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72e66960-3e03-4a37-af39-51d670111e9e_1900x880.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. Illustrative higher-order model, repeated indicators approach. The orange arrows are the relationships between the higher-order and lower-order constructs, read as loadings. Coefficients from SmartPLS (Figure 10).</figcaption></figure></div><h3>8. Building the higher-order model in SmartPLS 4</h3><p><strong>Step 6.</strong> Make a copy of <code>marketplace-x-model</code>, name it <code>hcm-model</code> and open the copy.</p><p><strong>Step 7.</strong> Select the five paths CONFIRM to VALUE, CONFIRM to SATISF, VALUE to SATISF, VALUE to INTENT and SATISF to INTENT one by one, and click <strong>Delete</strong> to remove them.</p><p><strong>Step 8.</strong> Click the <strong>Latent variable</strong> button, click on the canvas to create a new construct and name it <strong>EVALUATION</strong>.</p><p><strong>Step 9.</strong> Select the eight items <strong>PV1 to PV4</strong> and <strong>SAT1 to SAT4</strong> in the Indicators list and drag them onto EVALUATION. These items stay with VALUE and SATISF as well: this is the <strong>repeated indicators</strong> approach.</p><p><strong>Step 10.</strong> Click <strong>Connect</strong> and draw four paths: EVALUATION to VALUE, EVALUATION to SATISF (reflective-reflective, so the arrows point <strong>from</strong> the higher-order construct <strong>to</strong> the lower-order constructs), CONFIRM to EVALUATION and EVALUATION to INTENT. Click <strong>Save</strong> (Figure 9).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ekC4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ekC4!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 424w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 848w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 1272w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ekC4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png" width="1100" height="604" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/aa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:604,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:186367,&quot;alt&quot;:&quot;Figure 9. The higher-order model in SmartPLS&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. The higher-order model in SmartPLS" title="Figure 9. The higher-order model in SmartPLS" srcset="https://substackcdn.com/image/fetch/$s_!ekC4!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 424w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 848w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 1272w, https://substackcdn.com/image/fetch/$s_!ekC4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa7375dd-bb8d-4e38-b359-60c646b924ca_1100x604.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. The higher-order model in SmartPLS: (1) the Latent variable button, (2) the higher-order construct EVALUATION, (3) the eight items repeated from VALUE and SATISF.</figcaption></figure></div><p><strong>Step 11.</strong> Choose <strong>Calculate &gt; PLS-SEM algorithm</strong> (same settings as in the previous posts), click <strong>Start calculation</strong>, then open <strong>Final results &gt; Path coefficients &gt; Matrix</strong> (Figure 10).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2CY-!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!2CY-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:68382,&quot;alt&quot;:&quot;Figure 10. Coefficients of the higher-order model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. Coefficients of the higher-order model" title="Figure 10. Coefficients of the higher-order model" srcset="https://substackcdn.com/image/fetch/$s_!2CY-!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!2CY-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62aafd1a-8706-4f3f-9e4f-a10a2cda071d_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 10. Path coefficient matrix: (1) the item Matrix, (2) the relationships between EVALUATION and the two lower-order constructs, 0.888 (SATISF) and 0.875 (VALUE).</figcaption></figure></div><h3>9. Evaluating the higher-order construct</h3><p>For the reflective-reflective type, the two coefficients EVALUATION to VALUE (0.875) and EVALUATION to SATISF (0.888) are <strong>read as the loadings</strong> of the higher-order construct (Sarstedt et al., 2019). Assess reliability and convergent validity as for the reflective scales in <a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">Post 34</a>, but using these two "loadings":</p><ul><li><p><strong>AVE</strong> = (0.875&#178; + 0.888&#178;) / 2 = (0.766 + 0.789) / 2 = <strong>0.778</strong>, above 0.50.</p></li><li><p><strong>Composite reliability</strong> rho_c = (0.875 + 0.888)&#178; / [(0.875 + 0.888)&#178; + (1 - 0.766) + (1 - 0.789)] = 3.108 / 3.553 = <strong>0.875</strong>, above 0.70.</p></li></ul><p>The R&#178; values of VALUE (0.766) and SATISF (0.789) in the report are simply the squares of these two coefficients, not the explanatory power of a causal model. Do not interpret them as ordinary R&#178; values.</p><h3>10. Testing the structural model</h3><p><strong>Step 12.</strong> Run <strong>Calculate &gt; Bootstrapping</strong> with the settings of Post 39, then open <strong>Path coefficients &gt; Mean, STDEV, T values, p values</strong> (Figure 11).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!VtV0!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!VtV0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/df061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:116319,&quot;alt&quot;:&quot;Figure 11. Bootstrapping of the higher-order model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 11. Bootstrapping of the higher-order model" title="Figure 11. Bootstrapping of the higher-order model" srcset="https://substackcdn.com/image/fetch/$s_!VtV0!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!VtV0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdf061608-e4cc-408f-ac90-0bc6ffcec802_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 11. Significance tests of the higher-order model: (1) the item Mean, STDEV, T values, p values, (2) the two higher-order loadings, (3) CONFIRM to INTENT, p = 0.081.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/bIGMD/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0d79f2ed-7d04-426a-b72a-77f1fb64498f_1220x568.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/456a447e-32ab-4352-a86e-4aa2261cdf06_1220x638.png&quot;,&quot;height&quot;:297,&quot;title&quot;:&quot;Table 4. Bootstrapping results of the higher-order model&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/bIGMD/1/" width="730" height="297" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Interpretation:</strong></p><ul><li><p><strong>Post-purchase evaluation has a strong effect on INTENT</strong> (0.357; <em>p</em> &lt; 0.001), larger than the separate effects of VALUE (0.185) or SATISF (0.221) in the original model.</p></li><li><p><strong>The R&#178; of INTENT is still 0.318</strong>, the same as in the original model: merging the two constructs into one makes the model more parsimonious without reducing its explanatory power.</p></li><li><p><strong>The direct effect of CONFIRM on INTENT is still not significant</strong> (<em>p</em> = 0.081), as in Post 39.</p></li></ul><p>R (script <code>higher_order.R</code>) gives exactly the same coefficients; the standard errors and <em>p</em> values differ in the third decimal place because of the bootstrap samples (for example, CONFIRM to INTENT: <em>p</em> = 0.082).</p><p><strong>When not to use repeated indicators?</strong> For the <strong>reflective-formative</strong> or <strong>formative-formative</strong> types, if the higher-order construct has a predecessor (like CONFIRM here), the repeated indicators approach pulls the coefficient from that predecessor to the higher-order construct toward zero, because the lower-order constructs already explain almost all of the variance of the higher-order construct. In that case, use a two-stage approach (Becker et al., 2012; Sarstedt et al., 2019).</p><h2>11. Reporting the results</h2><blockquote><p><strong>IPMA.</strong> The IPMA for the target construct repurchase intention shows that confirmation of expectations has the highest importance (total effect 0.312) but the lowest performance (72.08 on a 100-point scale), followed by trust (0.280; 72.63), perceived value (0.270; 72.79) and satisfaction (0.221; 72.24). Because the performance of the factors is almost the same, managerial priority should go to confirmation of expectations, especially item CONF4 (importance 0.109; performance 71.4).</p><p><strong>Higher-order model.</strong> The higher-order construct post-purchase evaluation (reflective-reflective, estimated with repeated indicators; Sarstedt et al., 2019) has loadings on perceived value and satisfaction of 0.875 and 0.888 respectively (<em>p</em> &lt; 0.001), AVE = 0.778 and composite reliability 0.875. Post-purchase evaluation has a positive effect on repurchase intention (coefficient = 0.357; <em>t</em> = 6.334; <em>p</em> &lt; 0.001).</p></blockquote><h2>Five common mistakes</h2><ol><li><p><strong>Running IPMA while reverse-worded items have not been recoded.</strong> Negative weights make the performance values wrong.</p></li><li><p><strong>Reading the IPMA map without looking at the scale of the vertical axis.</strong> On a 0 to 100 map, the points can look identical; on a zoomed-in map, small differences look very large.</p></li><li><p><strong>Concluding "priority for improvement" from direct effects only.</strong> IPMA uses total effects; a factor whose direct effect is not significant can still be the most important (CONFIRM in this post).</p></li><li><p><strong>Building a higher-order construct without a theoretical basis</strong>, just to make the model more parsimonious or to avoid discriminant validity problems between the lower-order constructs.</p></li><li><p><strong>Interpreting the R&#178; of the lower-order constructs</strong> in a repeated indicators model as explanatory power.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Does the IPMA target construct have to be the final construct of the model?</strong></p><p>No, but you usually choose the outcome that matters most for management (repurchase intention here). You can run IPMA for several different target constructs.</p><p><strong>Is a performance of 72 high or low?</strong></p><p>IPMA has no general threshold. Performance is meaningful when compared across factors, across customer groups or across survey waves.</p><p><strong>Does a higher-order construct need a discriminant validity assessment?</strong></p><p>Yes. Discriminant validity is assessed between the higher-order construct and the other constructs in the model (here CONFIRM, TRUST and INTENT), not between the higher-order construct and its own lower-order constructs (Sarstedt et al., 2019).</p><p><strong>What does the repeated indicators approach require?</strong></p><p>The lower-order constructs should have roughly the same number of items; otherwise the lower-order construct with more items dominates the higher-order construct (Becker et al., 2012). In this post, VALUE and SATISF both have four items.</p><div><hr></div><p>This brings the 42-post series to a close: from research design and data cleaning in SPSS, through CFA and CB-SEM in JASP, to PLS-SEM in SmartPLS. Each post comes with data, scripts and screenshots so that readers can rerun the analyses and check every number.</p><h2>References</h2><p>Becker, J.-M., Klein, K., &amp; Wetzels, M. (2012). Hierarchical latent variable models in PLS-SEM: Guidelines for using reflective-formative type models. <em>Long Range Planning, 45</em>(5&#8211;6), 359&#8211;394. https://doi.org/10.1016/j.lrp.2012.10.001</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Ringle, C. M., &amp; Sarstedt, M. (2016). Gain more insight from your PLS-SEM results: The importance-performance map analysis. <em>Industrial Management &amp; Data Systems, 116</em>(9), 1865&#8211;1886. https://doi.org/10.1108/IMDS-10-2015-0449</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>Sarstedt, M., Hair, J. F., Cheah, J.-H., Becker, J.-M., &amp; Ringle, C. M. (2019). How to specify, estimate, and validate higher-order constructs in PLS-SEM. <em>Australasian Marketing Journal, 27</em>(3), 197&#8211;211. https://doi.org/10.1016/j.ausmj.2019.05.003</p>]]></content:encoded></item><item><title><![CDATA[Moderation Analysis in SmartPLS 4: Interaction, f² and Simple Slopes]]></title><description><![CDATA[Draw a moderating effect, read the interaction coefficient, f&#178; and simple slope plot, test it with bootstrapping and report a non-significant result]]></description><link>https://www.phuonglib.com/p/moderation-analysis-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/moderation-analysis-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:50:30 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Dk7z!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Dk7z!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!Dk7z!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:845752,&quot;alt&quot;:&quot;A beam of light from a lamp to a potted plant, a figure holding a translucent filter in the middle of the beam, the plant in orange light.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A beam of light from a lamp to a potted plant, a figure holding a translucent filter in the middle of the beam, the plant in orange light." title="A beam of light from a lamp to a potted plant, a figure holding a translucent filter in the middle of the beam, the plant in orange light." srcset="https://substackcdn.com/image/fetch/$s_!Dk7z!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Dk7z!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F289bcf86-096a-4ee2-b52f-1621cc931919_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>A moderator changes the strength of a path; in SmartPLS you draw it by connecting the moderator to the middle of that arrow.</p></li><li><p>SmartPLS 4 uses the two-stage approach: the interaction term is the product of the latent variable scores from the main-effects model.</p></li><li><p>With an interaction term in the model, a main-effect coefficient is the effect when the moderator is at its mean, not a general effect.</p></li><li><p>Trust does not moderate the effect of satisfaction on repurchase intention: coefficient &#8722;0.020 (p = 0.670) and f&#178; = 0.001, with almost parallel simple slopes.</p></li><li><p>A non-significant moderation result should still be reported in full, with the coefficient, p, confidence interval and f&#178;.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/mediation-analysis-smartpls">Post 40</a> showed <strong>through what</strong> expectation confirmation affects repurchase intention (mediation). This post asks a different question: does the effect of one construct become <strong>stronger or weaker depending on the level of a third construct</strong>? That third construct is a <strong>moderator</strong>.</p><p><a href="https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp">Post 29</a> tested a moderation hypothesis with CB-SEM: <strong>Does trust moderate the effect of satisfaction on repurchase intention?</strong> The reasoning: customers who trust Marketplace X may buy again even when they are not very satisfied, so the effect of satisfaction should be weaker when trust is high. This post tests the same hypothesis again with SmartPLS 4.</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch41-moderation-analysis-smartpls">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026; 30-day trial) for the model <code>moderation-model</code>, a copy of <code>marketplace-x-model</code> (<a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>), with casewise deletion on 254 respondents. Figure 3 (the PLS-SEM algorithm dialog) comes from an earlier capture with the same SmartPLS version; the dialog shows no project or model names. According to the SmartPLS license comparison page, the Student license does not include moderation analysis. The numbers in this post come from the SmartPLS report. Section 6 checks them in R (script <code>moderation.R</code> in the repository's <code>scripts/</code> folder, package cSEM) and against CB-SEM in AMOS and JASP (Post 29). SmartPLS lists constructs and paths in alphabetical order (CONFIRM, INTENT, SATISF, TRUST, VALUE, TRUST x SATISF).</p></blockquote><h2>1. Moderator and mediator</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/lDZf8/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a1429d06-64b0-472e-96a2-9508b59feb30_1220x462.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1a1d6bdc-7b1e-4ae8-81b1-e0dcdc278b0a_1220x532.png&quot;,&quot;height&quot;:260,&quot;title&quot;:&quot;Table 1. Mediator and moderator compared&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:false}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/lDZf8/1/" width="730" height="260" frameborder="0" scrolling="no"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>In the Marketplace X model, TRUST has a direct effect on INTENT (H7) and is also hypothesized to moderate the path from SATISF to INTENT (Figure 1).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Nc0E!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Nc0E!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 424w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 848w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 1272w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Nc0E!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png" width="1900" height="720" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:720,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:97694,&quot;alt&quot;:&quot;Figure 1. The moderation model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. The moderation model" title="Figure 1. The moderation model" srcset="https://substackcdn.com/image/fetch/$s_!Nc0E!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 424w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 848w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 1272w, https://substackcdn.com/image/fetch/$s_!Nc0E!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91c720c1-7194-4380-b3d4-f506fbf9ad86_1900x720.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Left: the conceptual model, TRUST acts on the path from SATISF to INTENT. Right: the statistical model, with the added interaction term TRUST x SATISF; the coefficients are from SmartPLS (Figure 4).</figcaption></figure></div><h2>2. How does SmartPLS estimate a moderating effect?</h2><p>SmartPLS 4 uses the <strong>two-stage approach</strong> (Becker et al., 2018; Hair et al., 2022):</p><ol><li><p><strong>Stage 1.</strong> Estimate the main-effects model (without the interaction term) and obtain the <strong>latent variable scores</strong> of SATISF and TRUST.</p></li><li><p><strong>Stage 2.</strong> Create the interaction term as the <strong>product of these two scores</strong>, use it as a single-item construct and estimate the model with the interaction term.</p></li></ol><p>Two points to know:</p><ul><li><p><strong>The path from the moderator to the dependent construct.</strong> SmartPLS adds the path from the moderator to the dependent construct automatically during the calculation, because the interaction test is only correct when the model also contains the main effect of the moderator. In the Marketplace X model, the path TRUST to INTENT (H7) is already there.</p></li><li><p><strong>The scale of the interaction term.</strong> With <strong>Type of results</strong> set to <strong>Standardized</strong>, the latent variable scores that are multiplied are standardized. Their product is kept as it is; it is not standardized again.</p></li></ul><h2>3. Adding the moderating effect to the model</h2><p><strong>Step 1.</strong> Create a copy of <code>marketplace-x-model</code> in the workspace: select the model in the Workspace pane and choose <strong>Files &gt; Duplicate</strong>. Rename the copy <code>moderation-model</code> and open it. Keep the original model unchanged so that the earlier posts can still be reproduced.</p><p><strong>Step 2.</strong> Click the <strong>Connect</strong> button on the toolbar. Click <strong>TRUST</strong>, drag to <strong>the middle of the arrow from SATISF to INTENT</strong> and release the mouse button (Figure 2). SmartPLS draws a <strong>dashed line</strong> from TRUST that touches that arrow; there is no separate node for the interaction term on the canvas.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Bvyb!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Bvyb!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 424w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 848w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 1272w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Bvyb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png" width="1100" height="597" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:597,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:168556,&quot;alt&quot;:&quot;Figure 2. Drawing the moderating effect&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Drawing the moderating effect" title="Figure 2. Drawing the moderating effect" srcset="https://substackcdn.com/image/fetch/$s_!Bvyb!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 424w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 848w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 1272w, https://substackcdn.com/image/fetch/$s_!Bvyb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5e9fd21d-2d4b-4869-a2ff-5e2cd5199857_1100x597.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. Drawing the moderating effect: (1) the Connect button, (2) the dashed line from TRUST that touches the arrow from SATISF to INTENT.</figcaption></figure></div><p><strong>Step 3.</strong> Click <strong>Save</strong>. Choose <strong>Calculate &gt; PLS-SEM algorithm</strong>, check that <strong>Type of results</strong> is <strong>Standardized</strong> and that the <strong>Data</strong> tab is set to <strong>Casewise deletion</strong>, then click <strong>Start calculation</strong> (Figure 3).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!2iUg!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2iUg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!2iUg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e9af1489-250a-48d8-a558-459246de0631_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:90188,&quot;alt&quot;:&quot;Figure 3. The PLS-SEM algorithm dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. The PLS-SEM algorithm dialog" title="Figure 3. The PLS-SEM algorithm dialog" srcset="https://substackcdn.com/image/fetch/$s_!2iUg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!2iUg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe9af1489-250a-48d8-a558-459246de0631_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. The PLS-SEM algorithm dialog: (1) Type of results set to Standardized, (2) the Start calculation button.</figcaption></figure></div><h2>4. Reading the results of the PLS-SEM algorithm</h2><h3>The interaction coefficient</h3><p>The results graph has an extra path from <strong>TRUST x SATISF</strong> to INTENT, shown as a label on the dashed line (Figure 4).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!K3nc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!K3nc!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!K3nc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png" width="1100" height="479" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:479,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:159335,&quot;alt&quot;:&quot;Figure 4. Results of the PLS-SEM algorithm&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Results of the PLS-SEM algorithm" title="Figure 4. Results of the PLS-SEM algorithm" srcset="https://substackcdn.com/image/fetch/$s_!K3nc!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!K3nc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd46e79f7-1372-45cf-8d6e-c998161f7e1d_1100x479.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Results of the PLS-SEM algorithm: (1) the coefficient of the moderating effect, &#8722;0.020, (2) the R&#178; of INTENT, 0.318.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/HluBH/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5639691b-52ab-4065-a196-d34682951539_1220x656.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ee299e4e-ede4-4f76-a0b2-9280e2f39858_1220x776.png&quot;,&quot;height&quot;:358,&quot;title&quot;:&quot;Table 2. Path coefficients to INTENT without and with the interaction term&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/HluBH/1/" width="730" height="358" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Reading the coefficients:</strong></p><ul><li><p><strong>The interaction coefficient of &#8722;0.020</strong> is negative, in the hypothesized direction (the higher the trust, the weaker the effect of satisfaction), but it is <strong>very small</strong>.</p></li><li><p><strong>With the interaction term in the model, the coefficient of SATISF (0.222) is the effect of satisfaction when trust is at its mean</strong>, no longer a "general effect" as in Post 37. Likewise, 0.275 is the effect of trust when satisfaction is at its mean.</p></li><li><p><strong>R&#178; does not change</strong> (0.318, rounded to three decimals): the interaction term explains almost no additional variance of INTENT.</p></li></ul><h3>Effect size f&#178;</h3><p><strong>Step 4.</strong> Open <strong>Quality criteria &gt; f-square &gt; Matrix</strong> (Figure 5).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Y6qL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Y6qL!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 424w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 848w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 1272w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Y6qL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png" width="1100" height="355" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:355,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:86842,&quot;alt&quot;:&quot;Figure 5. f&#178; of the interaction term&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. f&#178; of the interaction term" title="Figure 5. f&#178; of the interaction term" srcset="https://substackcdn.com/image/fetch/$s_!Y6qL!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 424w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 848w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 1272w, https://substackcdn.com/image/fetch/$s_!Y6qL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcdc932a3-83d6-4a9a-ad30-cc3c5899fef5_1100x355.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. The f&#178; matrix: (1) the item Matrix, (2) the f&#178; of TRUST x SATISF on INTENT, 0.001, which SmartPLS shows in red because it is below 0.02.</figcaption></figure></div><p>The f&#178; of the interaction term is <strong>0.001</strong>. For moderating effects, Cohen's threshold of 0.02 (Cohen, 1988) is too strict: in a 30-year review, Aguinis et al. (2005) found that the average f&#178; of moderating effects (with categorical moderators) in published studies was only about 0.009. Even against this benchmark, 0.001 is very small.</p><h3>Simple slopes</h3><p><strong>Step 5.</strong> Open <strong>Final results &gt; Simple slope analysis</strong> (Figure 6).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!nXcZ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!nXcZ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!nXcZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png" width="1100" height="479" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:479,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:136048,&quot;alt&quot;:&quot;Figure 6. Simple slope plot&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Simple slope plot" title="Figure 6. Simple slope plot" srcset="https://substackcdn.com/image/fetch/$s_!nXcZ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!nXcZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F05ab83ac-1e04-4008-a861-c47945f82a5c_1100x479.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. The simple slope plot: (1) the item Simple slope analysis, (2) the three lines for low trust (&#8722;1 standard deviation), mean trust and high trust (+1 standard deviation).</figcaption></figure></div><p>Each line shows the relationship between SATISF (horizontal axis) and INTENT (vertical axis) at one level of TRUST. Reading the plot:</p><ul><li><p><strong>The three lines are almost parallel.</strong> The slopes, computed by hand from the SmartPLS coefficients as 0.222 - 0.020 x (level of TRUST), are <strong>0.242</strong> when trust is low, <strong>0.222</strong> when it is at the mean and <strong>0.202</strong> when it is high. The slope decreases very little.</p></li><li><p><strong>The distance between the lines</strong> is the main effect of TRUST (0.275): at the same level of satisfaction, respondents with high trust have a higher repurchase intention.</p></li></ul><p>If there were a substantial moderating effect, the lines would <strong>not be parallel</strong>: for example, the line for high trust would be clearly flatter than the line for low trust.</p><h2>5. Testing significance with bootstrapping</h2><p><strong>Step 6.</strong> Click <strong>Edit</strong> to return to the model, choose <strong>Calculate &gt; Bootstrapping</strong> with the settings of <a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a> (5000 subsamples, Percentile bootstrap, Two-tailed, Fixed seed), then click <strong>Start calculation</strong>.</p><p><strong>Step 7.</strong> Open <strong>Final results &gt; Path coefficients &gt; Mean, STDEV, T values, p values</strong> (Figure 7).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!sK4c!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!sK4c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:128058,&quot;alt&quot;:&quot;Figure 7. Testing the coefficients with the interaction term&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Testing the coefficients with the interaction term" title="Figure 7. Testing the coefficients with the interaction term" srcset="https://substackcdn.com/image/fetch/$s_!sK4c!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!sK4c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4fc06dfe-f087-4ccc-b708-aab1afc5d453_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 7. Significance tests of the path coefficients in the moderation model: (1) the item Mean, STDEV, T values, p values, (2) TRUST x SATISF to INTENT, p = 0.670, (3) CONFIRM to INTENT, p = 0.105.</figcaption></figure></div><p><strong>Step 8.</strong> Open <strong>Path coefficients &gt; Confidence intervals</strong> (Figure 8).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!r1rv!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!r1rv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:111766,&quot;alt&quot;:&quot;Figure 8. Confidence intervals with the interaction term&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Confidence intervals with the interaction term" title="Figure 8. Confidence intervals with the interaction term" srcset="https://substackcdn.com/image/fetch/$s_!r1rv!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!r1rv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42ab47cc-a2a6-4476-9b80-2b468f6f5c3c_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 8. Percentile confidence intervals: (1) the item Confidence intervals, (2) TRUST x SATISF to INTENT, [&#8722;0.115, 0.070].</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/DwDxt/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/973f0dec-0285-4c88-b925-c41e44d68dbe_1220x682.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d5ec639c-66a3-4683-aa37-9515f99c0078_1220x752.png&quot;,&quot;height&quot;:355,&quot;title&quot;:&quot;Table 3. Bootstrapping results of the moderation model&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/DwDxt/1/" width="730" height="355" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Conclusion:</strong> the interaction coefficient is <strong>not significant</strong> (<em>t</em> = 0.427, <em>p</em> = 0.670), and the confidence interval includes 0. <strong>Trust does not moderate</strong> the effect of satisfaction on repurchase intention. The moderation hypothesis is <strong>not supported</strong>.</p><p>The other paths keep the conclusions of Post 39. H4 (CONFIRM to INTENT) is still not significant; <em>p</em> rises from 0.085 to 0.105 because the model has changed.</p><h2>6. Comparison with CB-SEM and R</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/FlYld/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3cf8b475-095d-497c-98ba-1857021c1c3b_1220x802.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ab1a48b7-d999-437d-a3bd-905392e75772_1220x922.png&quot;,&quot;height&quot;:382,&quot;title&quot;:&quot;Table 4. The interaction coefficient across software and methods&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/FlYld/1/" width="730" height="382" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>All four approaches lead to <strong>the same conclusion</strong>. The CB-SEM coefficients differ because AMOS and JASP use a latent interaction term measured by four product indicators and report a standardized coefficient, while SmartPLS uses the product of two latent variable scores. In R, the standardized coefficient of the interaction term is &#8722;0.024; dividing it by the standard deviation of the score product (1.149) gives &#8722;0.021, on the same scale as SmartPLS. The remaining difference of 0.001 may come from computational details in stage 2 and does not affect the conclusion.</p><h2>7. Reporting the results</h2><blockquote><p>The moderating effect of trust on the relationship between satisfaction and repurchase intention was tested with the two-stage approach (Becker et al., 2018) and bootstrapping with 5000 subsamples. The coefficient of the interaction term was negative, in the hypothesized direction, but very small and not statistically significant (coefficient = &#8722;0.020, <em>t</em> = 0.427, <em>p</em> = 0.670, 95% CI [&#8722;0.115, 0.070], f&#178; = 0.001). The R&#178; of repurchase intention did not change (0.318). The simple slope plot shows that the effect of satisfaction is almost constant across levels of trust (slopes of 0.242, 0.222 and 0.202 at low, mean and high trust). Therefore, the hypothesis that trust moderates the effect of satisfaction is not supported.</p></blockquote><p>When a moderating effect is not significant, still report the coefficient, <em>p</em>, confidence interval and f&#178;, and discuss possible reasons in the discussion section (for example, the variation of trust in the sample is not large enough, or the sample is too small to detect a small effect).</p><h2>Five common mistakes</h2><ol><li><p><strong>Reading a main-effect coefficient as a general effect.</strong> With an interaction term in the model, the SATISF coefficient of 0.222 is the effect when trust is at its mean.</p></li><li><p><strong>Leaving out the path from the moderator to the dependent construct.</strong> SmartPLS adds this path automatically during the calculation; when reporting, present the main effect of the moderator as well.</p></li><li><p><strong>Splitting a continuous moderator at the median</strong> and comparing the two groups. This discards information and reduces statistical power (Post 29).</p></li><li><p><strong>Concluding that there is moderation from the simple slope plot alone.</strong> The plot only illustrates; the conclusion rests on the test of the interaction coefficient.</p></li><li><p><strong>Testing many moderators and reporting only the significant ones.</strong> A moderation hypothesis needs a theoretical basis stated in advance.</p></li></ol><h2>Frequently asked questions</h2><p><strong>What if the moderator is categorical (for example, gender)?</strong></p><p>According to the SmartPLS documentation, a binary moderator should be coded 0 and 1, entered into the model as a single-item construct and the moderating effect drawn as in this post. An alternative is multigroup analysis (MGA), which compares the path coefficients between the two groups, similar to Post 29.</p><p><strong>Why does R&#178; not increase while the interaction coefficient is not 0?</strong></p><p>The interaction coefficient of &#8722;0.020 differs from 0 in the sample, but the additional variance it explains is tiny (f&#178; = 0.001), so R&#178; rounded to three decimals stays the same. The bootstrap test shows that this coefficient is not significantly different from 0.</p><p><strong>Can I use another way of creating the interaction term?</strong></p><p>The SmartPLS 4 documentation describes the two-stage approach. Becker et al. (2018) compared ways of creating interaction terms and recommend the two-stage approach in most situations, especially when the goal is to test the significance of the moderating effect.</p><p><strong>Is a sample of 254 large enough to detect a moderating effect?</strong></p><p>Moderating effects are usually small (Aguinis et al., 2005), so detecting them requires large samples. With an f&#178; of 0.001, as in this post, the required sample is much larger. Mention this in the limitations section.</p><div><hr></div><h2>References</h2><p>Aguinis, H., Beaty, J. C., Boik, R. J., &amp; Pierce, C. A. (2005). Effect size and power in assessing moderating effects of categorical variables using multiple regression: A 30-year review. <em>Journal of Applied Psychology, 90</em>(1), 94&#8211;107. https://doi.org/10.1037/0021-9010.90.1.94</p><p>Becker, J.-M., Ringle, C. M., &amp; Sarstedt, M. (2018). Estimating moderating effects in PLS-SEM and PLSc-SEM: Interaction term generation &#215; data treatment. <em>Journal of Applied Structural Equation Modeling, 2</em>(2), 1&#8211;21. https://doi.org/10.47263/JASEM.2(2)01</p><p>Cohen, J. (1988). <em>Statistical power analysis for the behavioral sciences</em> (2nd ed.). Lawrence Erlbaum Associates.</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p>]]></content:encoded></item><item><title><![CDATA[Mediation Analysis in SmartPLS 4: Indirect Effects and Mediation Types]]></title><description><![CDATA[Read specific and total indirect effects from the bootstrapping report, compute them by hand and classify mediation following Zhao et al. (2010)]]></description><link>https://www.phuonglib.com/p/mediation-analysis-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/mediation-analysis-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:50:04 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/209f7347-329f-441e-a608-51fb76176567_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!NBIW!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!NBIW!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!NBIW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/209f7347-329f-441e-a608-51fb76176567_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:1136354,&quot;alt&quot;:&quot;Orange water flowing from a mountain lake through two terraced pools down to a valley, beside a nearly dry direct channel.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Orange water flowing from a mountain lake through two terraced pools down to a valley, beside a nearly dry direct channel." title="Orange water flowing from a mountain lake through two terraced pools down to a valley, beside a nearly dry direct channel." srcset="https://substackcdn.com/image/fetch/$s_!NBIW!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!NBIW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F209f7347-329f-441e-a608-51fb76176567_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>A specific indirect effect is the product of the path coefficients along that path; the total effect is the direct effect plus the total indirect effect.</p></li><li><p>SmartPLS computes every indirect effect automatically each time you run Bootstrapping; an indirect effect is significant when its confidence interval excludes 0.</p></li><li><p>Expectation confirmation affects repurchase intention only indirectly: the total indirect effect is 0.211 [0.141, 0.292], while the direct effect is not significant (p = 0.085).</p></li><li><p>Classify mediation by the significance of the indirect and direct effects (Zhao et al., 2010), not by the Baron and Kenny conditions or by VAF.</p></li><li><p>PLS-SEM indirect effects are smaller than CB-SEM ones on the same model, but the conclusions are the same.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a> showed that the <strong>direct</strong> effect of expectation confirmation on repurchase intention is not significant (H4: <em>p</em> = 0.085). But in the Marketplace X model, confirmation also affects perceived value and satisfaction, and these two constructs in turn affect intention. Does confirmation then influence intention <strong>through</strong> these two constructs? That is the question of <strong>mediation analysis</strong>.</p><p><a href="https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp">Post 28</a> ran this analysis with CB-SEM in JASP. This post repeats it in SmartPLS 4, using the same bootstrapping report as Post 39, and compares the results.</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch40-mediation-analysis-smartpls">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of the SmartPLS 4.1.2.1 bootstrapping report (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code>: 5000 samples, percentile confidence intervals, two-tailed test, casewise deletion, 254 respondents (Post 39). According to the SmartPLS license comparison page, the Student license does not include mediation analysis. Section 5 checks the results in R (script <code>mediation.R</code> in the repository's <code>scripts/</code> folder, package cSEM). SmartPLS uses its own row order in these tables, so the order of the rows differs from the tables in this post.</p></blockquote><h2>1. Types of effects in the model</h2><p>The Marketplace X model has two mediators (VALUE and SATISF) in sequence. From CONFIRM to INTENT there are <strong>four paths</strong>:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/N1KTa/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/43a635d7-5458-4cf1-a512-b1f5ac2dee61_1220x568.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1237eb51-f954-442d-8c9a-a422d0433e44_1220x638.png&quot;,&quot;height&quot;:297,&quot;title&quot;:&quot;Table 1. The four paths from CONFIRM to INTENT&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/N1KTa/1/" width="730" height="297" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><ul><li><p>A <strong>specific indirect effect</strong> is the <strong>product</strong> of the coefficients along that path.</p></li><li><p>The <strong>total indirect effect</strong> is the sum of the specific indirect effects between two constructs.</p></li><li><p>The <strong>total effect</strong> = direct effect + total indirect effect.</p></li></ul><h3>Hand calculation</h3><p>Using the path coefficients from Post 39 (H1 = 0.501; H2 = 0.345; H3 = 0.383; H4 = 0.101; H5 = 0.185; H6 = 0.221):</p><ol><li><p><strong>Via VALUE:</strong> 0.501 x 0.185 = <strong>0.093</strong>.</p></li><li><p><strong>Via SATISF:</strong> 0.345 x 0.221 = <strong>0.076</strong>.</p></li><li><p><strong>Via the serial path VALUE then SATISF:</strong> 0.501 x 0.383 x 0.221 = <strong>0.042</strong>.</p></li><li><p><strong>Total indirect effect:</strong> 0.093 + 0.076 + 0.042 = <strong>0.211</strong>.</p></li><li><p><strong>Total effect:</strong> 0.101 + 0.211 = <strong>0.312</strong>.</p></li></ol><p>Multiplying rounded coefficients can differ from SmartPLS by one unit in the third decimal place, because SmartPLS multiplies the unrounded coefficients.</p><h3>Why bootstrapping?</h3><p>The product of two coefficients is not normally distributed, so the Sobel test (which assumes a normal distribution) is inaccurate (Preacher &amp; Hayes, 2008). Bootstrapping computes the product in each of the 5000 bootstrap samples and takes the 2.5% and 97.5% cut-off points as the confidence interval. An indirect effect is significant when the <strong>confidence interval does not contain 0</strong>.</p><p>In SmartPLS, <strong>no separate setup is needed</strong> for indirect effects. Every time you run Bootstrapping, SmartPLS automatically computes all specific indirect effects, total indirect effects and total effects of the model.</p><h2>2. Reading the indirect effects in SmartPLS 4</h2><p><strong>Step 1.</strong> Run <strong>Calculate &gt; Bootstrapping</strong> with the settings of Post 39 (5000 samples, Percentile bootstrap, Two-tailed, Fixed seed, Casewise deletion). If the report from Post 39 is still open, use that report.</p><p><strong>Step 2.</strong> Open <strong>Final results &gt; Specific indirect effects &gt; Mean, STDEV, T values, p values</strong> (Figure 1).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!4l_D!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!4l_D!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:114194,&quot;alt&quot;:&quot;Figure 1. Specific indirect effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Specific indirect effects" title="Figure 1. Specific indirect effects" srcset="https://substackcdn.com/image/fetch/$s_!4l_D!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!4l_D!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1db3f55b-2db7-44c8-8d79-9ae0f744b350_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 1. Specific indirect effects: (1) the item Mean, STDEV, T values, p values, (2) the p values, (3) the path CONFIRM to VALUE to INTENT, p = 0.014.</figcaption></figure></div><p>The table has five specific indirect effects. The name in the second row is truncated; the full name is CONFIRM -&gt; VALUE -&gt; SATISF -&gt; INTENT (serial mediation). The row CONFIRM -&gt; VALUE -&gt; SATISF is the indirect effect of CONFIRM on <strong>SATISF</strong> through VALUE; it has nothing to do with INTENT.</p><p><strong>Step 3.</strong> Open <strong>Specific indirect effects &gt; Confidence intervals</strong> (Figure 2).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!91iJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!91iJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:99675,&quot;alt&quot;:&quot;Figure 2. Confidence intervals of the specific indirect effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Confidence intervals of the specific indirect effects" title="Figure 2. Confidence intervals of the specific indirect effects" srcset="https://substackcdn.com/image/fetch/$s_!91iJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!91iJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee7b8a01-6de8-46fe-b9b9-8762d8f9c4c0_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 2. 95% percentile confidence intervals of the specific indirect effects: (1) the item Confidence intervals, (2) the 2.5% and 97.5% limits.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/mEISp/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/749be4f1-27fe-48a0-8a64-d4017080d68d_1220x970.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/51b06f90-c26d-4c5f-ac5e-0d7f0f79f533_1220x1090.png&quot;,&quot;height&quot;:520,&quot;title&quot;:&quot;Table 2. Specific indirect effects with bootstrap confidence intervals&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/mEISp/1/" width="730" height="520" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!X0t5!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!X0t5!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 424w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 848w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 1272w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!X0t5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png" width="1900" height="800" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:800,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:98620,&quot;alt&quot;:&quot;Figure 3. Indirect effects and confidence intervals&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Indirect effects and confidence intervals" title="Figure 3. Indirect effects and confidence intervals" srcset="https://substackcdn.com/image/fetch/$s_!X0t5!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 424w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 848w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 1272w, https://substackcdn.com/image/fetch/$s_!X0t5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F089b7dab-c4f1-4059-856b-0aaaf7dccb18_1900x800.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. Specific indirect effects (dots), the total indirect effect of CONFIRM on INTENT (diamond) and their 95% confidence intervals, numbers from SmartPLS. None of the intervals contains 0.</figcaption></figure></div><p><strong>Step 4.</strong> Open <strong>Final results &gt; Total indirect effects &gt; Mean, STDEV, T values, p values</strong> (Figure 4).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!4EMZ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!4EMZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/9649c43c-598b-440d-8a05-ca471307729e_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:90604,&quot;alt&quot;:&quot;Figure 4. Total indirect effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Total indirect effects" title="Figure 4. Total indirect effects" srcset="https://substackcdn.com/image/fetch/$s_!4EMZ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!4EMZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9649c43c-598b-440d-8a05-ca471307729e_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 4. Total indirect effects: (1) the item Mean, STDEV, T values, p values, (2) CONFIRM to INTENT, 0.211.</figcaption></figure></div><p><strong>Step 5.</strong> Open <strong>Total indirect effects &gt; Confidence intervals</strong> to get the confidence interval of the total indirect effect (Figure 5).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!sdSc!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!sdSc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:80536,&quot;alt&quot;:&quot;Figure 5. Confidence intervals of the total indirect effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Confidence intervals of the total indirect effects" title="Figure 5. Confidence intervals of the total indirect effects" srcset="https://substackcdn.com/image/fetch/$s_!sdSc!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!sdSc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9b2c57dd-60df-453d-ade2-4d6ee564fe31_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 5. Confidence intervals of the total indirect effects: (1) the item Confidence intervals, (2) CONFIRM to INTENT, [0.141, 0.292].</figcaption></figure></div><p><strong>Step 6.</strong> Open <strong>Final results &gt; Total effects &gt; Mean, STDEV, T values, p values</strong> (Figure 6).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!1LPV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!1LPV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:119492,&quot;alt&quot;:&quot;Figure 6. Total effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Total effects" title="Figure 6. Total effects" srcset="https://substackcdn.com/image/fetch/$s_!1LPV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!1LPV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f4bef61-db9c-4cb9-8051-391b2f3a4f34_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 6. Total effects: (1) the item Mean, STDEV, T values, p values, (2) CONFIRM to INTENT, 0.312, (3) VALUE to INTENT, 0.270.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/t3Ey6/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b7053bf7-dc7d-4a8b-ab7a-778f140d28cd_1220x526.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fd058010-c033-4780-b7d5-6c628feceb90_1220x596.png&quot;,&quot;height&quot;:276,&quot;title&quot;:&quot;Table 3. Direct, total indirect and total effects&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/t3Ey6/1/" width="730" height="276" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!v_TV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!v_TV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 424w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 848w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 1272w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!v_TV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png" width="1900" height="640" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:640,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:62973,&quot;alt&quot;:&quot;Figure 7. Decomposing the total effect&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Decomposing the total effect" title="Figure 7. Decomposing the total effect" srcset="https://substackcdn.com/image/fetch/$s_!v_TV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 424w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 848w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 1272w, https://substackcdn.com/image/fetch/$s_!v_TV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe00ebf9c-7f6c-43ac-8cfc-85f7a2bb7141_1900x640.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. The total effects of CONFIRM and of VALUE on INTENT, split into the direct part and the indirect parts, numbers from SmartPLS.</figcaption></figure></div><h2>3. Types of mediation</h2><p>The approach of Baron and Kenny (1986) requires a significant total effect before mediation is examined, and uses the Sobel test. Zhao et al. (2010) showed that this condition is not necessary: what matters is <strong>whether the indirect effect is significant</strong>. Nitzl et al. (2016) brought this classification into PLS-SEM, and Hair et al. (2022) follow it:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/1gwtq/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3c31ab69-b397-49ba-b010-853e347539dd_1220x578.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ecc2d1a0-760f-4877-a6af-4fcb9defbe7a_1220x698.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 4. Types of mediation (Zhao et al., 2010; Nitzl et al., 2016)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/1gwtq/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Applied to the Marketplace X model:</strong></p><ul><li><p><strong>CONFIRM to INTENT: indirect-only.</strong> The total indirect effect of 0.211 is significant, all three indirect paths are significant and the direct effect is not significant. Expectation confirmation affects repurchase intention <strong>through</strong> perceived value and satisfaction. So the fact that H4 is not supported does not mean that confirmation is unimportant.</p></li><li><p><strong>VALUE to INTENT: complementary.</strong> The indirect effect through SATISF (0.084) and the direct effect (0.185) are both significant and both positive. Satisfaction is a <strong>partial</strong> mediator: perceived value acts both directly and through satisfaction.</p></li><li><p><strong>CONFIRM to SATISF: complementary.</strong> The indirect effect through VALUE (0.192) and the direct effect (0.345) are both significant and have the same sign.</p></li></ul><h3>Variance accounted for (VAF)</h3><p>Some theses also report the <strong>VAF</strong> (variance accounted for) = total indirect effect / total effect, to describe <strong>what share</strong> of the effect runs through the mediators:</p><ul><li><p>CONFIRM to INTENT: 0.211 / 0.312 = <strong>67.6%</strong>.</p></li><li><p>VALUE to INTENT: 0.084 / 0.270 = <strong>31.1%</strong>.</p></li><li><p>CONFIRM to SATISF: 0.192 / 0.537 = <strong>35.8%</strong>.</p></li></ul><p>VAF is descriptive only. The type of mediation is decided by the significance of the indirect and direct effects, as in the table above (Nitzl et al., 2016). For example, for CONFIRM to INTENT the VAF is 67.6%, not 100%, because in the sample the direct effect is 0.101, not 0. Even so, this direct effect is not significantly different from 0, so the type of mediation is still indirect-only.</p><h2>4. Comparison with CB-SEM (Post 28)</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/doN7Z/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4f827c6e-ce38-4d61-a38e-4376bbf5d02f_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/56f458a7-8559-4623-a2c8-5d667512d9c7_1220x616.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 5. Indirect effects in PLS-SEM and CB-SEM compared&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/doN7Z/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Both columns are standardized coefficients (the CB-SEM column comes from Post 28, <em>Standardized estimates = All</em>). Post 28 used FIML on 267 respondents and 2000 bootstrap samples; this post uses the 254 complete responses and 5000 samples, so the two columns cannot be compared exactly. Even so, <strong>the conclusions are the same</strong>: all three indirect paths are significant, and CONFIRM to INTENT is indirect-only mediation. The PLS-SEM indirect effects are smaller, because PLS-SEM path coefficients are smaller than CB-SEM coefficients when the constructs are common factors (Posts <a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">31</a> and <a href="https://www.phuonglib.com/p/evaluate-structural-model-smartpls">37</a>).</p><p>Post 28 also noted that with CB-SEM on the 254 complete responses (unstandardized coefficients, percentile bootstrap), the confidence interval of the path via VALUE is [0.003, 0.250], with a lower bound very close to 0; an earlier bootstrap run even gave an interval touching 0. With PLS-SEM on the same 254 respondents, this interval is [0.024, 0.171] and does not contain 0. This path still has the largest <em>p</em> value (0.014) of the three indirect paths to INTENT, so interpret it with caution.</p><h2>5. Checking the results in R</h2><p>The script <code>mediation.R</code> computes each specific indirect effect as the product of the coefficients in each of the 5000 bootstrap samples of the cSEM package, then takes the 2.5% and 97.5% cut-off points:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/bxJNW/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/28194e91-166b-4008-af14-41d20baf62f7_1220x650.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0c6762a2-522f-42f0-9a2c-74aac9efe229_1220x720.png&quot;,&quot;height&quot;:320,&quot;title&quot;:&quot;Table 6. Indirect effects from R and SmartPLS compared&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/bxJNW/1/" width="730" height="320" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The coefficients match exactly; the confidence intervals differ by at most 0.002 because the random bootstrap samples are different.</p><h2>6. Reporting the results</h2><blockquote><p>Indirect effects were tested with bootstrapping using 5000 samples and 95% percentile confidence intervals (Nitzl et al., 2016). Expectation confirmation has positive, significant indirect effects on repurchase intention through perceived value (0.093; 95% CI [0.024, 0.171]), through satisfaction (0.076; [0.031, 0.130]) and through the serial path from perceived value to satisfaction (0.042; [0.017, 0.074]). The total indirect effect is 0.211 ([0.141, 0.292]), while the direct effect is not significant (0.101; <em>p</em> = 0.085), indicating indirect-only mediation (Zhao et al., 2010): the effect of expectation confirmation on repurchase intention runs through perceived value and satisfaction. Satisfaction is a complementary mediator in the relationship between perceived value and repurchase intention (indirect 0.084; [0.035, 0.144]; direct 0.185; <em>p</em> = 0.008).</p></blockquote><p>The mediation table in a thesis should include, for each path: the specific indirect effect, <em>t</em> or <em>p</em>, the 95% confidence interval, the corresponding direct effect and the type of mediation.</p><h2>Five common mistakes</h2><ol><li><p><strong>Stopping the analysis because the direct or total effect is not significant</strong> (the Baron and Kenny condition). For CONFIRM to INTENT, stopping at H4 would miss the most important result of this post.</p></li><li><p><strong>Reporting only the total indirect effect.</strong> When there are several mediators, report each specific indirect effect to show which paths are significant.</p></li><li><p><strong>Misreading the row CONFIRM -&gt; VALUE -&gt; SATISF</strong> as an effect on INTENT. It is an indirect effect on SATISF.</p></li><li><p><strong>Concluding that "full mediation" means there are no other paths.</strong> Zhao et al. (2010) point out that there may be mediators not included in the model; write "indirect-only mediation" instead.</p></li><li><p><strong>Using the Sobel test</strong> instead of bootstrap confidence intervals.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Do I need to declare the mediators in SmartPLS?</strong></p><p>No. SmartPLS computes all indirect effects from the structure of the model. You only need to draw the paths correctly and run Bootstrapping.</p><p><strong>Why does the p value of an indirect effect sometimes disagree with the confidence interval?</strong></p><p>SmartPLS computes <em>p</em> from <em>t</em> (coefficient / standard deviation), while the percentile confidence interval is taken directly from the bootstrap distribution, which is skewed for products of coefficients. When the two give different conclusions for a value close to the threshold, rely on the confidence interval (Preacher &amp; Hayes, 2008). In this post, both give the same conclusion.</p><p><strong>Should I use bias-corrected confidence intervals?</strong></p><p>SmartPLS also offers the item Bias-corrected confidence intervals. This post uses percentile intervals to stay consistent with Post 39. Post 28 compared the two types of confidence interval in CB-SEM.</p><p><strong>Do the mediators have to be measured with reflective scales?</strong></p><p>No. The analysis is the same for reflective and formative measurement, as long as the scales met the requirements of the measurement model evaluation (Posts <a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">34</a> to <a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">36</a>).</p><div><hr></div><h2>References</h2><p>Baron, R. M., &amp; Kenny, D. A. (1986). The moderator&#8211;mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. <em>Journal of Personality and Social Psychology, 51</em>(6), 1173&#8211;1182. https://doi.org/10.1037/0022-3514.51.6.1173</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Nitzl, C., Rold&#225;n, J. L., &amp; Cepeda, G. (2016). Mediation analysis in partial least squares path modeling: Helping researchers discuss more sophisticated models. <em>Industrial Management &amp; Data Systems, 116</em>(9), 1849&#8211;1864. https://doi.org/10.1108/IMDS-07-2015-0302</p><p>Preacher, K. J., &amp; Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. <em>Behavior Research Methods, 40</em>(3), 879&#8211;891. https://doi.org/10.3758/BRM.40.3.879</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>Zhao, X., Lynch, J. G., Jr., &amp; Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. <em>Journal of Consumer Research, 37</em>(2), 197&#8211;206. https://doi.org/10.1086/651257</p>]]></content:encoded></item><item><title><![CDATA[How to Install SmartPLS 4 and Find Your Way Around the Interface]]></title><description><![CDATA[Install SmartPLS 4 on Windows or macOS, pick the right Mac installer, activate a license and learn every area of the interface]]></description><link>https://www.phuonglib.com/p/install-smartpls-4-interface</link><guid isPermaLink="false">https://www.phuonglib.com/p/install-smartpls-4-interface</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:49:40 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/a3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!NQQT!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!NQQT!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:null,&quot;width&quot;:null,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:null,&quot;alt&quot;:&quot;A small figure opening a large wooden crate from which a tidy workbench with drawers and an orange drafting lamp unfolds.&quot;,&quot;title&quot;:null,&quot;type&quot;:null,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A small figure opening a large wooden crate from which a tidy workbench with drawers and an orange drafting lamp unfolds." title="A small figure opening a large wooden crate from which a tidy workbench with drawers and an orange drafting lamp unfolds." srcset="https://substackcdn.com/image/fetch/$s_!NQQT!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!NQQT!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3fed38d-6920-4b24-a4bd-3da02719274a_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div></div></div></div></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>SmartPLS 4.1.2.1 runs on Windows, macOS and Linux; Apple silicon and Intel Macs need different installers.</p></li><li><p>The free Student license accepts at most 100 rows of CSV data, so it cannot run the 267-respondent Marketplace X data set.</p></li><li><p>The 30-day trial has every Professional feature but is available only once per computer; start it when you are ready to work with real data.</p></li><li><p>A project holds the data and its models; this series creates the project Marketplace X with the model marketplace-x-model.</p></li><li><p>Close the license dialog with Start using SmartPLS, never with Deactivate license.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31, "What Is SmartPLS? PLS-SEM vs CB-SEM"</a> answered the question of <em>when</em> to use PLS-SEM. This post prepares the tool: you install SmartPLS 4, activate a license, create the first project for the running example of the series (repurchase intention on an online marketplace, called <em>Marketplace X</em>) and walk through each area of the interface. By the end, you have a project called <strong>Marketplace X</strong> with an empty PLS-SEM model, ready for importing data and drawing the model in <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33, "Importing Data and Drawing a Model in SmartPLS 4"</a>.</p><blockquote><p><strong>About the screenshots.</strong> All figures are real screenshots of SmartPLS 4.1.2.1 on macOS (Apple silicon Mac). Figures 2, 3, 8, 9, 11 and 12 were taken on 5 November 2026, when the computer had the Student license; the other screenshots were taken in November 2026 with the 30-day trial. The steps inside the installer were not captured and are described in words. In the license dialog, the licensee name and the license key are masked. Information on installers, system requirements and licenses comes from the official SmartPLS website, accessed on 5 and 6 November 2026.</p></blockquote><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!WKJL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!WKJL!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 424w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 848w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 1272w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!WKJL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png" width="1675" height="540" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:540,&quot;width&quot;:1675,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:88424,&quot;alt&quot;:&quot;Figure 1. Four steps to get started with SmartPLS 4&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Four steps to get started with SmartPLS 4" title="Figure 1. Four steps to get started with SmartPLS 4" srcset="https://substackcdn.com/image/fetch/$s_!WKJL!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 424w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 848w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 1272w, https://substackcdn.com/image/fetch/$s_!WKJL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe325415f-6c5c-4ded-97f7-a96113b9570e_1675x540.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Four steps to get started with SmartPLS 4, matching Sections 2 to 5 of this post.</figcaption></figure></div><h2>1. Before you start: version and system requirements</h2><p>The current version is <strong>SmartPLS 4.1.2.1</strong>, released on 25 October 2026 (SmartPLS GmbH, n.d.-b). It fixes a single bug of version 4.1.2.0, released the same day: dragging indicators into the model did not work (SmartPLS GmbH, n.d.-f). If you have 4.1.2.0, update before working through Post 33.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/ciKIR/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/151b104f-b809-4361-96af-bdb59912cbd5_1220x668.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4e8dfdf2-7a75-4a7a-ad92-b5a4d7e5cb3d_1220x738.png&quot;,&quot;height&quot;:330,&quot;title&quot;:&quot;Table 1. Installers and system requirements of SmartPLS 4&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:false}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/ciKIR/" width="730" height="330" frameborder="0" scrolling="no"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Three points to check before downloading:</p><ul><li><p><strong>Memory and disk:</strong> at least 2 GB of RAM and about 200 MB of disk space (SmartPLS GmbH, n.d.-h). Large models and bootstrapping with many subsamples use more memory.</p></li><li><p><strong>Mac computers:</strong> pick the installer that matches the chip. To see which chip you have, open the Apple menu in the top-left corner and choose <strong>About This Mac</strong>: a <em>Chip</em> line reading Apple M1, M2 and so on means Apple silicon; a <em>Processor</em> line reading Intel means an Intel Mac.</p></li><li><p><strong>Older Macs running a macOS version before 11:</strong> they can only use SmartPLS 4.0.9.6, which cannot be updated to the current SmartPLS 4 (SmartPLS GmbH, n.d.-h).</p></li></ul><p>SmartPLS 3 installs into its own folder and runs side by side with SmartPLS 4 (SmartPLS GmbH, n.d.-b), so you do not need to remove the old version if you still have old projects.</p><h2>2. Download and install</h2><p><strong>Step 1.</strong> Open the official download page <code>https://www.smartpls.com/downloads</code> and download the installer for your computer (table in Section 1). Download only from the SmartPLS website, not from third-party sites.</p><p><strong>Step 2.</strong> Run the installer:</p><ul><li><p><strong>Windows:</strong> double-click the <code>.exe</code> file and accept the default steps of the installer.</p></li><li><p><strong>macOS:</strong> double-click the <code>.dmg</code> file. A window opens with the <strong>SmartPLS 4 Installer</strong> icon (Figure 2). Double-click this icon to start the installer, then accept the default steps. Unlike many Mac apps, you do <strong>not</strong> drag the icon into the Applications folder; the installer puts SmartPLS in the right place (on the computer used for the screenshots, the folder <code>/Applications/SmartPLS 4</code>).</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!_Y6-!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!_Y6-!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 424w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 848w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 1272w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!_Y6-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png" width="700" height="445" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:445,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:131705,&quot;alt&quot;:&quot;Figure 2. Window of the .dmg file on macOS&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Window of the .dmg file on macOS" title="Figure 2. Window of the .dmg file on macOS" srcset="https://substackcdn.com/image/fetch/$s_!_Y6-!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 424w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 848w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 1272w, https://substackcdn.com/image/fetch/$s_!_Y6-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b75313b-1569-45cf-961d-da82a84b0bee_700x445.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. Window that opens when you double-click <code>smartpls4_macos_arm.dmg</code>: (1) the SmartPLS 4 Installer, (2) the instruction to double-click it.</figcaption></figure></div><p><strong>Step 3.</strong> Open SmartPLS. On Windows, open it from the Start menu; on macOS, open it from Launchpad or the <code>SmartPLS 4</code> folder in Applications.</p><p>After installation, SmartPLS <strong>updates itself</strong>. If an update fails, download the installer again and install over the existing version (SmartPLS GmbH, n.d.-b).</p><h2>3. Activate a license</h2><p>The first time you open SmartPLS, it asks you to choose a license type. There are three options (SmartPLS GmbH, n.d.-a):</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Egl7Z/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dce680cb-efcb-4d79-a85b-c2209167f1df_1220x750.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cf30b22b-c11a-43e8-87c8-ea3afc8a393a_1220x820.png&quot;,&quot;height&quot;:375,&quot;title&quot;:&quot;Table 2. SmartPLS 4 license types&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Egl7Z/" width="730" height="375" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>To activate the trial, choose <strong>Free 30-day trial</strong>, enter your email address, open your mailbox to get the confirmation code and enter the code in SmartPLS (SmartPLS GmbH, n.d.-c). A seat license needs an Internet connection for activation; after that, SmartPLS runs offline (SmartPLS GmbH, n.d.-g). A computer without Internet access can be activated offline with an activation certificate (SmartPLS GmbH, n.d.-d).</p><p>To view or change the current license, choose <strong>SmartPLS &gt; Switch license</strong> (Figure 3).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!eOA5!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!eOA5!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 424w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 848w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 1272w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!eOA5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png" width="640" height="432" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:432,&quot;width&quot;:640,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:69817,&quot;alt&quot;:&quot;Figure 3. Current license dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Current license dialog" title="Figure 3. Current license dialog" srcset="https://substackcdn.com/image/fetch/$s_!eOA5!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 424w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 848w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 1272w, https://substackcdn.com/image/fetch/$s_!eOA5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdbb2a0fc-3fcf-4cb4-9ef6-710a4ff9026d_640x432.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. The Your current license dialog: (1) license type, here Student with reduced functionality, (2) expiry date, (3) the button that deactivates the license, (4) the button that closes the dialog so you can keep working. The licensee name and the license key are masked.</figcaption></figure></div><p>Two things to know about this dialog:</p><ul><li><p><strong>Do not click Deactivate license</strong> unless you want to move the license to another computer or change the license type. This button removes the license from the computer. To close the dialog, click <strong>Start using SmartPLS</strong>.</p></li><li><p><strong>Do not share your license key</strong> when you send screenshots to other people. This is why the key in Figure 3 is masked.</p></li></ul><p><strong>Can the Student license run the Marketplace X data?</strong> No. The comparison table on the SmartPLS pricing page states that the Student license accepts at most 100 rows of data, reads only CSV files and lacks tools used later in this series, such as HTMT (<a href="https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls">Post 35, "Discriminant Validity in SmartPLS 4"</a>), mediation analysis (<a href="https://www.phuonglib.com/p/mediation-analysis-smartpls">Post 40, "Mediation Analysis in SmartPLS 4"</a>), moderation (<a href="https://www.phuonglib.com/p/moderation-analysis-smartpls">Post 41, "Moderation Analysis in SmartPLS 4"</a>), PLSpredict (<a href="https://www.phuonglib.com/p/plspredict-cvpat-smartpls">Post 38, "Assessing Predictive Power in SmartPLS 4"</a>), IPMA and higher-order models (<a href="https://www.phuonglib.com/p/ipma-higher-order-constructs-smartpls">Post 42, "IPMA and Higher-Order Constructs in SmartPLS 4"</a>) (SmartPLS GmbH, n.d.-e). The Marketplace X data set has 267 respondents, so with the Student license SmartPLS imports the file but marks it as <em>not licensed</em> and does not let you use it (Post 33).</p><p><strong>Which license does this series use?</strong> The license dialog and the Calculate menu in this post were captured with the Student license, because installing and exploring the interface do not need data. From Post 33 onwards, the screenshots use the <strong>30-day trial</strong>, which has all features of the Professional license. If you only have the Student license, you can still learn how to draw the model, and you can cross-check the numbers with JASP or the R package cSEM as introduced in Post 31. Because each computer gets only one trial, activate it when you are ready to work through the data-based posts in one go, or when you run the real data of your thesis.</p><h2>4. Welcome screen and workspace</h2><p>After activation, SmartPLS opens the welcome screen (Figure 4).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!yK_8!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!yK_8!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!yK_8!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:186212,&quot;alt&quot;:&quot;Figure 4. Welcome screen of SmartPLS 4&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Welcome screen of SmartPLS 4" title="Figure 4. Welcome screen of SmartPLS 4" srcset="https://substackcdn.com/image/fetch/$s_!yK_8!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!yK_8!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F67ffe21a-97a2-4a5d-b981-bba0b3aa7937_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Welcome screen of SmartPLS 4.1.2.1 with an empty workspace. The numbers are explained right below.</figcaption></figure></div><ol><li><p><strong>Menu bar:</strong> three menus, SmartPLS, Files and Themes. When a model is open, the menu bar also shows Edit, Calculate and Accessibility (Section 6).</p></li><li><p><strong>Toolbar:</strong> the <strong>New project</strong> button and five buttons that create models: <strong>PLS-SEM</strong>, <strong>CB-SEM</strong>, <strong>GSCA</strong>, <strong>PROCESS</strong> and <strong>REGRESSION</strong>. The model buttons are grayed out while the workspace has no project.</p></li><li><p><strong>Workspace pane:</strong> the list of projects. The line below the word Workspace is the folder that stores the projects, here <code>~/Documents/smartpls4_workspace_en</code> (~ stands for your home folder). The <strong>Archive</strong> entry holds archived projects.</p></li><li><p><strong>License message:</strong> here "You are using a Professional license, all features enabled", which is what the trial shows. With the Student license, the message says that some features are disabled.</p></li><li><p><strong>Sample projects:</strong> ready-made projects for many methods (PLS-SEM, CB-SEM, regression, GSCA and others). One of them, <em>Corporate reputation &#8211; PLS-SEM book (primer)</em>, is the example of the book by Hair et al. (2022).</p></li><li><p><strong>Version in use:</strong> here v. 4.1.2.1. Report this version number in the methods section of your thesis.</p></li></ol><p>The <strong>workspace</strong> is a folder on your computer where SmartPLS stores all projects (SmartPLS GmbH, n.d.-b). Each project is a subfolder named after the project. To use another folder, choose <strong>Files &gt; Choose workspace</strong>. When you move to another computer or make a backup, use <strong>Files &gt; Export project</strong> instead of copying folders by hand. If you keep the workspace in a cloud-synchronized folder such as Dropbox or OneDrive, use SmartPLS 4.1.1.9 or later, which made workspaces in such folders more stable (SmartPLS GmbH, n.d.-f).</p><h2>5. Create the first project and a PLS-SEM model</h2><p>In SmartPLS, a <strong>project</strong> holds the data and the models that use those data. A project can have several models, for example the base model and a model with a moderator. We create the project <strong>Marketplace X</strong> for the running example of the series.</p><p><strong>Step 1.</strong> Click <strong>New project</strong> on the toolbar. The <em>Create a new project</em> dialog opens.</p><p><strong>Step 2.</strong> Type the project name in the <strong>Name</strong> box (Figure 5), then click <strong>Create</strong>. Keep the name short and plain, because the project name is also the folder name in the workspace.</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!svZo!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 424w, https://substackcdn.com/image/fetch/$s_!svZo!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 848w, https://substackcdn.com/image/fetch/$s_!svZo!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 1272w, https://substackcdn.com/image/fetch/$s_!svZo!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!svZo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png" width="560" height="238" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:238,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:27429,&quot;alt&quot;:&quot;Figure 5. Create a new project dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Create a new project dialog" title="Figure 5. Create a new project dialog" srcset="https://substackcdn.com/image/fetch/$s_!svZo!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 424w, https://substackcdn.com/image/fetch/$s_!svZo!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 848w, https://substackcdn.com/image/fetch/$s_!svZo!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 1272w, https://substackcdn.com/image/fetch/$s_!svZo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6941599d-43ca-4b50-8cb9-c6c14a712f20_560x238.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 5. The Create a new project dialog: (1) project name box, (2) Create button.</figcaption></figure></div><p>The new project appears in the Workspace pane (Figure 6) with two links: <strong>Import data file</strong> to import data and <strong>Create model</strong> to create a model. The model buttons on the toolbar become active.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!lPWL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!lPWL!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 424w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 848w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 1272w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!lPWL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png" width="760" height="304" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:304,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:91689,&quot;alt&quot;:&quot;Figure 6. The new project Marketplace X&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. The new project Marketplace X" title="Figure 6. The new project Marketplace X" srcset="https://substackcdn.com/image/fetch/$s_!lPWL!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 424w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 848w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 1272w, https://substackcdn.com/image/fetch/$s_!lPWL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1733c50-3158-4d8d-862a-3b6c8f0e22cd_760x304.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. (1) Project Marketplace X with the two links Import data file and Create model, (2) the model buttons, now active.</figcaption></figure></div><p><strong>Step 3.</strong> Select the project <strong>Marketplace X</strong>, then click the <strong>PLS-SEM</strong> button on the toolbar. The <em>New model</em> dialog opens (Figure 7).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ZFEH!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ZFEH!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 424w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 848w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 1272w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ZFEH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png" width="360" height="448" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b32785c8-521e-46b0-ab9f-47334cee1127_360x448.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:448,&quot;width&quot;:360,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:42031,&quot;alt&quot;:&quot;Figure 7. New model dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. New model dialog" title="Figure 7. New model dialog" srcset="https://substackcdn.com/image/fetch/$s_!ZFEH!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 424w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 848w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 1272w, https://substackcdn.com/image/fetch/$s_!ZFEH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb32785c8-521e-46b0-ab9f-47334cee1127_360x448.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. The New model dialog: (1) project that will hold the model, (2) model type, (3) model name, (4) Save button.</figcaption></figure></div><p><strong>Step 4.</strong> Check that <strong>Project</strong> shows Marketplace X and <strong>Model type</strong> shows PLS-SEM. The default model name is <em>pls-sem</em>; this series names the model <strong>marketplace-x-model</strong>. Click <strong>Save</strong>. The model window opens (Figure 8).</p><h2>6. The model window</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!3Cb2!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!3Cb2!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 424w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 848w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 1272w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!3Cb2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png" width="1000" height="572" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:572,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:76132,&quot;alt&quot;:&quot;Figure 8. Model window of a PLS-SEM model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Model window of a PLS-SEM model" title="Figure 8. Model window of a PLS-SEM model" srcset="https://substackcdn.com/image/fetch/$s_!3Cb2!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 424w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 848w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 1272w, https://substackcdn.com/image/fetch/$s_!3Cb2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcbbc8a5-7122-4df6-a018-cb2d73d61acd_1000x572.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. Model window of a PLS-SEM model without data (captured with a model that kept the default name pls-sem). The numbers are explained right below.</figcaption></figure></div><ol><li><p><strong>Menu bar of the model:</strong> SmartPLS, Edit, Calculate, Accessibility, Themes. The <strong>Calculate</strong> menu holds the algorithms.</p></li><li><p><strong>Management buttons:</strong> <strong>Back</strong> returns to the project screen, <strong>Save</strong> saves the model, <strong>Calculate</strong> runs an algorithm.</p></li><li><p><strong>Drawing buttons:</strong> <strong>Select</strong> (select and move), <strong>Latent variable</strong> (add a construct), <strong>Connect</strong> (draw a path between two constructs), <strong>Quadratic effect</strong>, <strong>Gaussian copula</strong> (endogeneity control), <strong>Comment</strong> (note on the diagram), <strong>Delete</strong>.</p></li><li><p><strong>Indicators pane:</strong> the list of indicators in the data file. Without data, it reads "No indicators available. Select a data file!". The <strong>Filter</strong> and <strong>Sort</strong> links filter and sort the list.</p></li><li><p><strong>Canvas:</strong> where you place constructs and drag indicators onto them. The title bar above shows the model name and the model type (<em>PLS-SEM</em>).</p></li><li><p><strong>Tools pane:</strong> undo and redo, canvas selection, coloring and default settings.</p></li><li><p><strong>Bottom bar:</strong> the <strong>Recalculate automatically</strong> switch (reruns the algorithm whenever the model changes) and the zoom slider.</p></li></ol><p>Post 33 uses the buttons in area (3) and the pane (4) to draw the Marketplace X model.</p><h3>The Calculate menu</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!taq0!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!taq0!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 424w, https://substackcdn.com/image/fetch/$s_!taq0!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 848w, https://substackcdn.com/image/fetch/$s_!taq0!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 1272w, https://substackcdn.com/image/fetch/$s_!taq0!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!taq0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png" width="700" height="266" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:266,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:86461,&quot;alt&quot;:&quot;Figure 9. Calculate menu with the Student license&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Calculate menu with the Student license" title="Figure 9. Calculate menu with the Student license" srcset="https://substackcdn.com/image/fetch/$s_!taq0!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 424w, https://substackcdn.com/image/fetch/$s_!taq0!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 848w, https://substackcdn.com/image/fetch/$s_!taq0!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 1272w, https://substackcdn.com/image/fetch/$s_!taq0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F23ccaeea-aaec-4d42-860f-6dcba4f97313_700x266.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. Calculate menu of a PLS-SEM model with the Student license: (1) PLS-SEM algorithm and Bootstrapping, (2) Consistent PLS-SEM algorithm and Consistent PLS-SEM bootstrapping. The items are grayed out because the model has no data yet.</figcaption></figure></div><p>With the Student license on the computer used for the screenshots, the Calculate menu has exactly four items: the PLS-SEM algorithm, bootstrapping, the consistent PLS algorithm (PLSc, Post 31) and bootstrapping for PLSc. Other tools of this series, such as PLSpredict (Post 38), are available only with the trial and the Professional license (SmartPLS GmbH, n.d.-e). The items become usable once the model has data and is completely drawn.</p><h2>7. Menus and settings worth knowing</h2><h3>The Files menu</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!fcGq!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!fcGq!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 424w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 848w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 1272w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!fcGq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png" width="560" height="677" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:677,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:166465,&quot;alt&quot;:&quot;Figure 10. Files menu&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. Files menu" title="Figure 10. Files menu" srcset="https://substackcdn.com/image/fetch/$s_!fcGq!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 424w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 848w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 1272w, https://substackcdn.com/image/fetch/$s_!fcGq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb1db49ab-e76f-4697-b18b-ed158452e03e_560x677.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 10. Files menu with the trial license: (1) choose a workspace, (2) create a project and the five model types, (3) import projects, data files and AMOS files, (4) export, archive and restore projects.</figcaption></figure></div><ol><li><p><strong>Choose workspace:</strong> switch to another workspace folder.</p></li><li><p><strong>New project</strong> and <strong>New PLS-SEM, REGRESSION, PROCESS, CBSEM, GSCA model:</strong> create projects and models, like the toolbar buttons.</p></li><li><p><strong>Import project from backup file, Import projects from a folder, Import data file, Import SPSS AMOS file:</strong> open a project someone sent you, import a data file into the selected project (Post 33) or bring in a model drawn in AMOS.</p></li><li><p><strong>Export project, Archive project, Restore project from archive:</strong> export a project to back it up or send it to someone, move a project you no longer use into Archive and bring it back. In Figure 10 they are grayed out: they act on a selected project, and a model was selected when the screenshot was taken.</p></li></ol><h3>The SmartPLS and Themes menus</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!cVfz!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!cVfz!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 424w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 848w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 1272w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!cVfz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png" width="820" height="257" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:257,&quot;width&quot;:820,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:84428,&quot;alt&quot;:&quot;Figure 11. SmartPLS menu and Themes menu&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 11. SmartPLS menu and Themes menu" title="Figure 11. SmartPLS menu and Themes menu" srcset="https://substackcdn.com/image/fetch/$s_!cVfz!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 424w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 848w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 1272w, https://substackcdn.com/image/fetch/$s_!cVfz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43b1788b-ace0-429b-b7a6-ed5838453954_820x257.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 11. Left, the SmartPLS menu: (1) Switch license, (2) Preferences, (3) Check for update. Right, the Themes menu (4) with four color themes.</figcaption></figure></div><p>The <strong>SmartPLS</strong> menu holds Switch license (Section 3), Preferences (below), About, Check for update and Exit. The <strong>Themes</strong> menu switches between Light, Dark and two color-blind themes. The screenshots in this series use the Light theme.</p><h3>Preferences</h3><p>Choose <strong>SmartPLS &gt; Preferences</strong> to open the settings dialog (Figure 12).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!gJOe!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!gJOe!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 424w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 848w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 1272w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!gJOe!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png" width="620" height="662" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:662,&quot;width&quot;:620,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:83094,&quot;alt&quot;:&quot;Figure 12. Preferences dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 12. Preferences dialog" title="Figure 12. Preferences dialog" srcset="https://substackcdn.com/image/fetch/$s_!gJOe!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 424w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 848w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 1272w, https://substackcdn.com/image/fetch/$s_!gJOe!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb482310a-21e9-4afd-8e66-ab2bc38aba05_620x662.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 12. Preferences dialog with the default settings on the computer used for the screenshots. The numbers are explained right below.</figcaption></figure></div><ol><li><p><strong>Language:</strong> interface language, English (en-EN) by default. The list offers several other interface languages.</p></li><li><p><strong>Color Scheme:</strong> color theme, the same as the Themes menu.</p></li><li><p><strong>Locale:</strong> number format, by default <em>US (example: 1,000.23)</em>, that is, a decimal point and a comma as the thousands separator.</p></li><li><p><strong>Decimal places:</strong> number of decimals in results tables, 3 by default.</p></li><li><p><strong>Processors for parallel computing:</strong> number of processor cores used for parallel computations such as bootstrapping. On the computer used for the screenshots, the default is 8.</p></li><li><p><strong>Close</strong> and <strong>Apply:</strong> click Apply to save changes; Close without Apply discards them.</p></li></ol><p>This series keeps the <strong>English</strong> interface, for three reasons: the names of the statistics in SmartPLS match textbooks and journal articles (outer loading, HTMT, f-square and so on), they are easy to compare with the references and error messages are easy to search for in forums. The series also keeps the default US locale, so numbers in the text match the SmartPLS output (<em>0.501</em>). If your university or journal uses a decimal comma, change the numbers consistently when you copy results into your thesis.</p><h2>8. Check the result</h2><p>After the steps above, the Workspace pane shows the project <strong>Marketplace X</strong> with the model <strong>marketplace-x-model</strong>. Only the <strong>Import data file</strong> link remains, because the project has no data yet. Figure 13 shows what the same project will look like at the end of this series, after Posts 33 to 42 add the data set and two more models.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!SijU!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!SijU!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 424w, https://substackcdn.com/image/fetch/$s_!SijU!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 848w, https://substackcdn.com/image/fetch/$s_!SijU!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 1272w, https://substackcdn.com/image/fetch/$s_!SijU!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!SijU!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png" width="600" height="333" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:333,&quot;width&quot;:600,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:57092,&quot;alt&quot;:&quot;Figure 13. Project Marketplace X at the end of the series&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 13. Project Marketplace X at the end of the series" title="Figure 13. Project Marketplace X at the end of the series" srcset="https://substackcdn.com/image/fetch/$s_!SijU!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 424w, https://substackcdn.com/image/fetch/$s_!SijU!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 848w, https://substackcdn.com/image/fetch/$s_!SijU!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 1272w, https://substackcdn.com/image/fetch/$s_!SijU!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf34c526-cd0e-47e9-8f9e-d8a4f05fbeb2_600x333.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 13. Project tree at the end of the series: (1) project Marketplace X, (2) its three models, hcm-model (Post 42), marketplace-x-model (Post 33) and moderation-model (Post 41), and the data set marketplace_x_smartpls with 267 rows (Post 33).</figcaption></figure></div><p>Post 33 starts here: import the Marketplace X data file into the project, then draw the measurement model and the structural model.</p><h2>Four common beginner mistakes</h2><ol><li><p><strong>Downloading the wrong Mac installer.</strong> Apple silicon Macs use the <code>_arm.dmg</code> file, Intel Macs the <code>_intel.dmg</code> file. Check the chip in About This Mac before downloading.</p></li><li><p><strong>Clicking Deactivate license to close the license dialog.</strong> The close button is Start using SmartPLS; Deactivate removes the license from the computer.</p></li><li><p><strong>Spending the 30-day trial while just exploring.</strong> Each computer gets one trial. Learn the interface with the Student license and keep the trial for when you need the full set of tools.</p></li><li><p><strong>Not knowing where projects are stored.</strong> The workspace path appears right below the word Workspace. Back up with Files &gt; Export project.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Is the Student license enough for a thesis?</strong></p><p>Usually not. The Student license accepts at most 100 rows of data, whereas a thesis usually has a few hundred respondents; it also cannot export reports and lacks HTMT, mediation analysis and PLSpredict (SmartPLS GmbH, n.d.-e). Use the 30-day trial, a Professional license (from your university or your own; the pricing page lists a 50% academic discount) or estimate the model with JASP and R as in Post 31.</p><p><strong>Does SmartPLS 4 run on Apple silicon Macs?</strong></p><p>Yes. SmartPLS has a separate installer for Apple silicon since version 4.0.9.7 (SmartPLS GmbH, n.d.-f). The screenshots in this post were taken on an Apple silicon Mac.</p><p><strong>Do SmartPLS 4 results change between updates?</strong></p><p>Minor updates mostly fix bugs and improve the interface, but you should still report the exact version (for example 4.1.2.1) in the methods section and in the reference list, following the example at the end of this post.</p><p><strong>How do I cite SmartPLS in my thesis?</strong></p><p>In APA 7, a software citation gives the authors, year, software name, version and publisher, for example: Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><div><hr></div><h2>References</h2><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>SmartPLS GmbH. (n.d.-a). <em>Activating a license key in SmartPLS</em>. Retrieved November 5, 2026, from https://www.smartpls.com/faq/license-and-installation/activation/</p><p>SmartPLS GmbH. (n.d.-b). <em>Download SmartPLS 4</em>. Retrieved November 5, 2026, from https://www.smartpls.com/downloads</p><p>SmartPLS GmbH. (n.d.-c). <em>Free trial for SmartPLS 4</em>. Retrieved November 5, 2026, from https://www.smartpls.com/free-trial/</p><p>SmartPLS GmbH. (n.d.-d). <em>Offline activation</em>. Retrieved November 5, 2026, from https://www.smartpls.com/offline-activation/</p><p>SmartPLS GmbH. (n.d.-e). <em>Pricing and licenses</em>. Retrieved November 6, 2026, from https://smartpls.com/pricing/</p><p>SmartPLS GmbH. (n.d.-f). <em>Release notes</em>. Retrieved November 5, 2026, from https://www.smartpls.com/release_notes</p><p>SmartPLS GmbH. (n.d.-g). <em>SmartPLS license options</em>. Retrieved November 5, 2026, from https://www.smartpls.com/faq/license-and-installation/license-options/</p><p>SmartPLS GmbH. (n.d.-h). <em>System requirements for SmartPLS 4</em>. Retrieved November 5, 2026, from https://www.smartpls.com/faq/smartpls4/system-requirements/</p>]]></content:encoded></item><item><title><![CDATA[Bootstrapping in SmartPLS 4: Testing the Significance of the Paths]]></title><description><![CDATA[Set up bootstrapping, read t values, p values and percentile confidence intervals, and decide which PLS-SEM hypotheses are supported]]></description><link>https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance</link><guid isPermaLink="false">https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:49:36 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!xEVw!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!xEVw!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!xEVw!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:751430,&quot;alt&quot;:&quot;A figure scooping marbles from one glass jar into many small bowls that form a bell-shaped arc, two orange flags marking its edges.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A figure scooping marbles from one glass jar into many small bowls that form a bell-shaped arc, two orange flags marking its edges." title="A figure scooping marbles from one glass jar into many small bowls that form a bell-shaped arc, two orange flags marking its edges." srcset="https://substackcdn.com/image/fetch/$s_!xEVw!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!xEVw!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F20be939a-85b9-4324-96dc-6d5fef85f636_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>PLS-SEM has no formula for standard errors, so it estimates them from bootstrap subsamples drawn with replacement; 5000 subsamples is the usual number.</p></li><li><p>Use percentile confidence intervals, a two-tailed test at 0.05 and Fixed seed, and report these settings with the results.</p></li><li><p>A path is significant when its 95% interval excludes 0; in the Marketplace X model six of seven hypotheses are supported and H4 is not (t = 1.721, p = 0.085).</p></li><li><p>Report the Original sample (O) coefficient, not the bootstrap mean, and write p &lt; 0.001 rather than p = 0.000.</p></li><li><p>Choose one- or two-tailed tests before seeing the results; switching afterwards to rescue H4 is methodologically wrong.</p></li></ul></blockquote><div><hr></div><p>Posts <a href="https://www.phuonglib.com/p/evaluate-structural-model-smartpls">37</a> and <a href="https://www.phuonglib.com/p/plspredict-cvpat-smartpls">38</a> assessed collinearity, R&#178;, f&#178; and the predictive power of the Marketplace X model. The question that remains is the most important one in a thesis: <strong>are hypotheses H1 to H7 supported?</strong> To answer it, you need to know whether each path coefficient is <strong>significantly different from 0</strong>. In PLS-SEM this is done with <strong>bootstrapping</strong>, step 2 of the structural model assessment procedure of Hair et al. (2019).</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch39-bootstrapping-smartpls">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code> from <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>, with casewise deletion on 254 respondents. The numbers in this post come from the SmartPLS report. Section 5 compares them with R (script <code>bootstrap.R</code> in the repository's <code>scripts/</code> folder, package cSEM). Bootstrapping is available with the Student license, but that license accepts at most 100 rows of data. SmartPLS lists paths in alphabetical order (CONFIRM -&gt; INTENT first), so the order of the rows in the report differs from the order of the hypotheses.</p></blockquote><h2>1. What is bootstrapping?</h2><p>CB-SEM (AMOS, lavaan) computes standard errors with a formula that assumes normally distributed data. PLS-SEM makes no such assumption, so it has <strong>no formula for the standard errors</strong>. Instead, the standard errors are estimated empirically:</p><ol><li><p><strong>Draw</strong> 254 respondents at random, with replacement, from the 254 respondents of the original sample. One respondent can be drawn several times, another not at all. This is one <strong>bootstrap subsample</strong>.</p></li><li><p><strong>Estimate</strong> the model again on that subsample and record the seven path coefficients.</p></li><li><p><strong>Repeat</strong> steps 1 and 2 many times, for example 5000 times.</p></li><li><p><strong>Compute</strong> the standard deviation of the 5000 coefficients. This is the <strong>standard error</strong> of the coefficient.</p></li><li><p><strong>Take</strong> the 2.5% and 97.5% cut-offs of the 5000 coefficients as the <strong>95% confidence interval</strong> (percentile confidence interval).</p></li></ol><p>A coefficient is significant at the 5% level when its 95% confidence interval <strong>does not contain 0</strong>. Equivalently, <em>t</em> = coefficient / standard error is larger than 1.96 (two-tailed test), or <em>p</em> is smaller than 0.05.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ry-L!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ry-L!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 424w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 848w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 1272w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ry-L!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png" width="1900" height="760" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:760,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:111845,&quot;alt&quot;:&quot;Figure 1. Bootstrap distribution of H4 and H7&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Bootstrap distribution of H4 and H7" title="Figure 1. Bootstrap distribution of H4 and H7" srcset="https://substackcdn.com/image/fetch/$s_!ry-L!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 424w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 848w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 1272w, https://substackcdn.com/image/fetch/$s_!ry-L!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd35a6264-dd01-479d-846c-dfa1c96718bd_1900x760.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Distribution of the coefficients of H4 and H7 over 5000 bootstrap subsamples (R values, script <code>bootstrap.R</code>). For H4, the range between the two dotted lines contains 0; for H7, it lies entirely to the right of 0.</figcaption></figure></div><p>Figure 1 shows the difference between a non-significant path and a significant one. For H4, 4.8% of the bootstrap subsamples (240 of 5000) give a negative coefficient, so the possibility that the true coefficient is 0 cannot be ruled out. For H7, even the smallest coefficient among the 5000 subsamples is 0.105, far from 0.</p><h2>2. Running bootstrapping in SmartPLS 4</h2><p><strong>Step 1.</strong> Open the model <code>marketplace-x-model</code> in the project Marketplace X. Choose <strong>Calculate &gt; Bootstrapping</strong>.</p><p><strong>Step 2.</strong> On the <strong>BT setup</strong> tab, check five settings (Figure 2):</p><ul><li><p><strong>Subsamples</strong>: 5000, the number of bootstrap subsamples. This is the usual number in PLS-SEM research. You can increase it to 10000 for the final results so that <em>p</em> values and confidence intervals are more stable (Streukens &amp; Leroi-Werelds, 2016), but the calculation takes longer.</p></li><li><p><strong>Confidence interval method</strong>: <strong>Percentile bootstrap</strong>. Percentile confidence intervals are recommended for PLS-SEM (Aguirre-Urreta &amp; R&#246;nkk&#246;, 2018; Hair et al., 2022).</p></li><li><p><strong>Test type</strong>: <strong>Two-tailed</strong> (see section 4).</p></li><li><p><strong>Significance level</strong>: 0.05.</p></li><li><p><strong>Random number generator</strong>: <strong>Fixed seed</strong>, so that running it again gives the same results.</p></li></ul><p>Keep <strong>Amount of results</strong> at <strong>Most important (faster)</strong>: the report still contains all the tables used in this post and in <a href="https://www.phuonglib.com/p/mediation-analysis-smartpls">Post 40</a>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!5kQf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!5kQf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!5kQf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:131642,&quot;alt&quot;:&quot;Figure 2. BT setup tab&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. BT setup tab" title="Figure 2. BT setup tab" srcset="https://substackcdn.com/image/fetch/$s_!5kQf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!5kQf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F695a49ee-1be4-442f-8fe9-963f094f6d8e_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. The BT setup tab: (1) number of bootstrap subsamples, (2) confidence interval method, (3) two-tailed test, (4) significance level, (5) random number generator.</figcaption></figure></div><p><strong>Step 3.</strong> Open the <strong>Data</strong> tab, set <strong>Missing value algorithm</strong> to <strong>Casewise deletion</strong>, then click <strong>Start calculation</strong> (Figure 3). Keep the <strong>PLS setup</strong> tab as for the PLS-SEM algorithm in Post 33 (Path weighting scheme).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!exO2!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!exO2!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!exO2!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!exO2!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!exO2!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!exO2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:89073,&quot;alt&quot;:&quot;Figure 3. Data tab&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Data tab" title="Figure 3. Data tab" srcset="https://substackcdn.com/image/fetch/$s_!exO2!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!exO2!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!exO2!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!exO2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F73032364-e09c-4166-901c-a7fa86c4c5f7_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. The Data tab: (1) the Data tab, (2) Casewise deletion, (3) the Start calculation button.</figcaption></figure></div><p>With 5000 subsamples, SmartPLS estimates the model 5000 times. On a Mac with Apple silicon this takes from a few seconds to a few tens of seconds.</p><h2>3. Reading the results</h2><h3>p values on the diagram</h3><p><strong>Step 4.</strong> The report opens at <strong>Graphical output</strong> (Figure 4). In the <strong>Graphical output</strong> pane at the bottom left, set <strong>Structural model</strong> to <strong>P values</strong> so that each path shows its <em>p</em> value.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!yinT!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!yinT!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!yinT!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!yinT!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!yinT!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!yinT!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png" width="1100" height="479" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:479,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:158309,&quot;alt&quot;:&quot;Figure 4. Diagram with p values&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Diagram with p values" title="Figure 4. Diagram with p values" srcset="https://substackcdn.com/image/fetch/$s_!yinT!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 424w, https://substackcdn.com/image/fetch/$s_!yinT!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 848w, https://substackcdn.com/image/fetch/$s_!yinT!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 1272w, https://substackcdn.com/image/fetch/$s_!yinT!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F06952cd7-a64e-4204-a727-0c0d310dc3cb_1100x479.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Diagram of the bootstrapping results: (1) the item Graphical output, (2) p values shown on the paths of the structural model, (3) H4 CONFIRM to INTENT, p = 0.085, (4) H5 VALUE to INTENT, p = 0.008. The numbers in the circles are R&#178; values.</figcaption></figure></div><p>The diagram gives a quick overview, but some labels overlap (the label 0.085 covers the name VALUE, so that construct appears without a name). When writing up, read the numbers from the tables.</p><h3>Path coefficient table</h3><p><strong>Step 5.</strong> Open <strong>Final results &gt; Path coefficients &gt; Mean, STDEV, T values, p values</strong> (Figure 5).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!j41J!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!j41J!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!j41J!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!j41J!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!j41J!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:122775,&quot;alt&quot;:&quot;Figure 5. Path coefficients table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Path coefficients table" title="Figure 5. Path coefficients table" srcset="https://substackcdn.com/image/fetch/$s_!j41J!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!j41J!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!j41J!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!j41J!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6b42280f-1346-4bd5-a9c9-236de9314dc2_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 5. Significance test of the path coefficients: (1) the item Mean, STDEV, T values, p values, (2) t values, (3) p values, (4) the path CONFIRM -&gt; INTENT, p = 0.085, so SmartPLS shows it in red.</figcaption></figure></div><p>The columns mean:</p><ul><li><p><strong>Original sample (O)</strong>: the coefficient estimated on the original sample; this is the coefficient to report.</p></li><li><p><strong>Sample mean (M)</strong>: the mean of the 5000 bootstrap coefficients. M close to O shows that the bootstrap distribution is not strongly shifted away from the original coefficient.</p></li><li><p><strong>Standard deviation (STDEV)</strong>: the standard deviation of the 5000 coefficients, that is, the standard error.</p></li><li><p><strong>T statistics (|O/STDEV|)</strong>: the absolute value of O divided by STDEV.</p></li><li><p><strong>P values</strong>: the <em>p</em> value that corresponds to <em>t</em>.</p></li></ul><p><strong>Hand calculation of the t value of H7</strong> (TRUST to INTENT): <em>t</em> = 0.2805 / 0.0501 = <strong>5.60</strong>. SmartPLS reports 5.598 because it uses the unrounded coefficient and standard deviation.</p><h3>Confidence intervals</h3><p><strong>Step 6.</strong> Open <strong>Path coefficients &gt; Confidence intervals</strong> (Figure 6). The columns <strong>2.5%</strong> and <strong>97.5%</strong> are the lower and upper limits of the 95% percentile confidence interval.</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!IxQg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!IxQg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png" width="1100" height="246" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:246,&quot;width&quot;:1100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:107411,&quot;alt&quot;:&quot;Figure 6. Confidence intervals table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Confidence intervals table" title="Figure 6. Confidence intervals table" srcset="https://substackcdn.com/image/fetch/$s_!IxQg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 424w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 848w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 1272w, https://substackcdn.com/image/fetch/$s_!IxQg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe2994801-0318-4ba0-a5f5-e0635afc2e07_1100x246.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 6. Percentile confidence intervals: (1) the item Confidence intervals, (2) the 2.5% and 97.5% limits, (3) the interval of CONFIRM -&gt; INTENT contains 0.</figcaption></figure></div><p>The item <strong>Bias-corrected confidence intervals</strong> just below gives intervals corrected for bias. This post uses the percentile intervals chosen in Step 2.</p><h3>Conclusions for the seven hypotheses</h3><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/FN0gM/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ebbe2596-fe50-433a-97be-f72375812730_1220x1014.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/30767ca6-52a9-465f-aed3-48ab5c4a6a9c_1220x1122.png&quot;,&quot;height&quot;:511,&quot;title&quot;:&quot;Table 1. Bootstrapping results for the seven hypotheses&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/FN0gM/1/" width="730" height="511" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!9P8M!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!9P8M!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 424w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 848w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 1272w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!9P8M!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png" width="1900" height="880" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:880,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:116258,&quot;alt&quot;:&quot;Figure 7. Coefficients and confidence intervals&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Coefficients and confidence intervals" title="Figure 7. Coefficients and confidence intervals" srcset="https://substackcdn.com/image/fetch/$s_!9P8M!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 424w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 848w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 1272w, https://substackcdn.com/image/fetch/$s_!9P8M!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfe41651-ba9f-435f-8559-6ac801b23dbf_1900x880.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Path coefficients (dots) and 95% percentile confidence intervals (lines) of the seven hypotheses, values from SmartPLS (Figures 5 and 6).</figcaption></figure></div><p><strong>Reading the results:</strong></p><ul><li><p><strong>Six hypotheses are supported.</strong> The confidence intervals of H1, H2, H3, H5, H6 and H7 lie entirely to the right of 0, and <em>p</em> is below 0.01.</p></li><li><p><strong>H4 is not supported</strong> (<em>t</em> = 1.721, <em>p</em> = 0.085). The confidence interval [&#8722;0.017, 0.214] contains 0. This agrees with f&#178; = 0.009 in Post 37 and with CB-SEM in <a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31</a> (<em>p</em> = 0.443).</p></li><li><p><strong>H5 is significant but its confidence interval is wide</strong> [0.048, 0.322]: the effect of perceived value on intention is clearly positive, but it may be very small.</p></li><li><p><strong>Statistical significance is not size.</strong> H5 has <em>p</em> = 0.008 but f&#178; = 0.032 (small; Post 37). Report both.</p></li></ul><h2>4. One-tailed or two-tailed?</h2><p>Hypotheses H1 to H7 are all directional (positive effects), so in theory a one-tailed test could be used. For H4, the one-tailed <em>p</em> value is half the two-tailed value, about 0.043, which is <strong>below 0.05</strong>.</p><p>This post still concludes that H4 is <strong>not supported</strong>, for two reasons:</p><ol><li><p><strong>The test must be chosen before looking at the results.</strong> Switching to a one-tailed test after seeing <em>p</em> = 0.085 to "rescue" a hypothesis is methodologically wrong.</p></li><li><p><strong>The two-tailed test is the common default</strong> in PLS-SEM studies and is more conservative. If a thesis uses one-tailed tests, state this in the methods section, before the results, and apply it to every hypothesis.</p></li></ol><p>Add a note in the text: with a one-tailed test, H4 is close to the 0.05 threshold. This is a borderline result that should be reported honestly, neither ignored nor exaggerated.</p><h2>5. Checking with R</h2><p>The script <code>bootstrap.R</code> runs the same model with the cSEM package: ordinary PLS-SEM, 5000 subsamples, percentile confidence intervals, 254 respondents. Because the random bootstrap subsamples differ from those of SmartPLS, the standard errors and <em>p</em> values agree only approximately:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/k1aiJ/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e2848997-0e8d-42f1-8415-00a28eeb3664_1220x530.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/13cd6c85-773c-46d5-b5e0-c68ee3ea27f6_1220x600.png&quot;,&quot;height&quot;:295,&quot;title&quot;:&quot;Table 2. Bootstrap results from SmartPLS and R compared&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/k1aiJ/1/" width="730" height="295" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The original coefficients (O) are identical, because they do not depend on the bootstrap subsamples. The conclusions for the seven hypotheses are the same in both programs. Two runs with different seeds in the same program also differ in the third decimal place in this way.</p><h2>6. Reporting the results</h2><blockquote><p>The significance of the path coefficients was tested with bootstrapping (5000 subsamples, 95% percentile confidence intervals, two-tailed test). The results support six of the seven hypotheses. Expectation confirmation has a positive effect on perceived value (coefficient = 0.501; <em>t</em> = 9.603; <em>p</em> &lt; 0.001; 95% CI [0.390, 0.599]) and on satisfaction (coefficient = 0.345; <em>t</em> = 5.782; <em>p</em> &lt; 0.001). Perceived value has a positive effect on satisfaction (coefficient = 0.383; <em>p</em> &lt; 0.001) and on repurchase intention (coefficient = 0.185; <em>t</em> = 2.662; <em>p</em> = 0.008). Satisfaction (coefficient = 0.221; <em>p</em> &lt; 0.001) and trust (coefficient = 0.280; <em>t</em> = 5.598; <em>p</em> &lt; 0.001) have positive effects on repurchase intention. The direct effect of expectation confirmation on repurchase intention is not statistically significant (coefficient = 0.101; <em>t</em> = 1.721; <em>p</em> = 0.085; 95% CI [&#8722;0.017, 0.214]), so H4 is not supported.</p></blockquote><p>The hypothesis table in a thesis should include: the coefficient (O), <em>t</em>, <em>p</em>, the 95% confidence interval, f&#178; (Post 37) and the conclusion. In a note below the table, give the number of bootstrap subsamples, the type of confidence interval and whether the test is one-tailed or two-tailed.</p><h2>Five common mistakes</h2><ol><li><p><strong>Reporting the Sample mean (M) instead of the Original sample (O).</strong> The coefficient to report is O.</p></li><li><p><strong>Writing p = 0.000.</strong> A <em>p</em> value is never 0; write <em>p</em> &lt; 0.001.</p></li><li><p><strong>Switching to a one-tailed test after seeing the results</strong> to make a hypothesis significant (section 4).</p></li><li><p><strong>Using too few bootstrap subsamples</strong>, for example 500. The <em>p</em> values and confidence intervals will vary a lot between runs.</p></li><li><p><strong>Not reporting the settings.</strong> Without the number of subsamples, the type of confidence interval and the number of tails, readers cannot check the results.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Why does the p value change slightly when I run it again?</strong></p><p>With Random seed, every run draws different bootstrap subsamples. Choose Fixed seed to get the same results. With 5000 subsamples, runs differ only in the third decimal place.</p><p><strong>Is t larger than 1.96 significant?</strong></p><p>Yes, for a two-tailed test at the 5% level. For a one-tailed test the threshold is 1.645. At the 1% level (two-tailed) it is 2.576.</p><p><strong>If H4 is not significant, does CONFIRM have no influence on INTENT?</strong></p><p>You can only conclude that the <strong>direct effect</strong> is not significant. Confirmation can still affect intention <strong>indirectly</strong> through perceived value and satisfaction. Post 40 tests this indirect effect with the tables Specific indirect effects and Total effects of the same bootstrapping report.</p><p><strong>Do I need to bootstrap the loadings and weights?</strong></p><p>For formative measurement, bootstrap the outer weights, as in <a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">Post 36</a>. For reflective measurement, the loadings are usually large and therefore significant; the Outer loadings table of the same report gives the <em>p</em> values if your supervisor asks for them.</p><div><hr></div><h2>References</h2><p>Aguirre-Urreta, M. I., &amp; R&#246;nkk&#246;, M. (2018). Statistical inference with PLSc using bootstrap confidence intervals. <em>MIS Quarterly, 42</em>(3), 1001&#8211;1020. https://doi.org/10.25300/MISQ/2018/13587</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>Streukens, S., &amp; Leroi-Werelds, S. (2016). Bootstrapping and PLS-SEM: A step-by-step guide to get more out of your bootstrap results. <em>European Management Journal, 34</em>(6), 618&#8211;632. https://doi.org/10.1016/j.emj.2016.06.003</p>]]></content:encoded></item><item><title><![CDATA[Multigroup Analysis and Moderation in JASP: Invariance and Path Tests]]></title><description><![CDATA[Run a multigroup SEM by gender in free JASP, test measurement invariance, compare each path and test a continuous moderator with product indicators]]></description><link>https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:49:20 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!l95C!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!l95C!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!l95C!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!l95C!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!l95C!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!l95C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:890362,&quot;alt&quot;:&quot;A long level balance beam on one pivot with two identical small glasshouses at its ends, an orange plumb line hanging from the pivot.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A long level balance beam on one pivot with two identical small glasshouses at its ends, an orange plumb line hanging from the pivot." title="A long level balance beam on one pivot with two identical small glasshouses at its ends, an orange plumb line hanging from the pivot." srcset="https://substackcdn.com/image/fetch/$s_!l95C!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!l95C!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!l95C!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!l95C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F85dcda5c-552c-4d52-aa65-33a51223faf2_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Test configural, metric and scalar invariance before comparing paths or latent means; here all three hold.</p></li><li><p>In JASP 0.99.0, equality constraints on paths take effect only when the syntax contains at least one <code>:=</code> defined parameter.</p></li><li><p>Constraining the seven paths to be equal does not worsen the fit (<em>p</em> = 0.484): gender does not moderate the model.</p></li><li><p>A path that is significant for women and not for men is not evidence of moderation; test the difference itself.</p></li><li><p>Test a continuous moderator with a latent interaction, not a median split; trust does not moderate satisfaction (<em>p</em> = 0.615).</p></li></ul></blockquote><div><hr></div><p>Posts <a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">26</a> to <a href="https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp">28</a> estimated <strong>one</strong> model for all respondents. The natural next question is whether that model holds equally for every group of customers. For example, does trust affect the repurchase intention of women more strongly than that of men? That is a question about a <strong>moderator</strong>.</p><p>This post does two things:</p><ul><li><p><strong>Multigroup analysis</strong> by gender: test measurement invariance, then compare the seven paths between men and women.</p></li><li><p><strong>A continuous moderator</strong>: test whether trust moderates the effect of satisfaction, using a latent interaction variable.</p></li></ul><p>Both results are <strong>not significant</strong>. The post keeps them as they are and uses them to show a very common error of interpretation: seeing a coefficient that is significant in one group and not in the other, and concluding that there is moderation.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch29-multigroup-analysis-moderation-sem">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The screenshots come from JASP 0.99.0 (JASP Team, 2026) on macOS, and the result tables are taken from JASP's <em>Export Results</em>. The numbers were checked with the lavaan package for R (Rosseel, 2012), using the scripts <code>multigroup.R</code> and <code>continuous_moderation.R</code> in the repository: both give the same results to the third decimal. Where JASP does not print a number (CFI, RMSEA, SRMR, the two-decimal chi-square differences, R&#178;, standardized coefficients of the interaction model), the number comes from lavaan and the text says so.</p></blockquote><h2>1. Moderators and two ways to test them</h2><p>A moderator <strong>changes the strength or the direction</strong> of the effect between two other variables (Baron &amp; Kenny, 1986). It differs from the mediator of Post 28: a mediator is the route the effect travels along, while a moderator is the condition that makes the effect stronger or weaker.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/SEiPb/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e8f9c9fe-beb5-40ea-890f-aae38b7d1b37_1220x456.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0e79c7f1-62a5-4f67-9a5b-9c47bedf32be_1220x576.png&quot;,&quot;height&quot;:206,&quot;title&quot;:&quot;Table 1. Two types of moderator and how to test them in CB-SEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/SEiPb/1/" width="730" height="206" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Do not turn a continuous variable into a grouping variable by splitting it at the median: this loses information, reduces power and can produce spurious results (MacCallum et al., 2002). Section 7 shows this on the real data.</p><h2>2. The multigroup procedure</h2><p>Comparing paths between two groups only makes sense when both groups <strong>understand the scales in the same way</strong>. If women read the item "Marketplace X is trustworthy" differently from men, the path coefficients of the two groups measure two different things. That is why <strong>measurement invariance</strong> must be tested first (Vandenberg &amp; Lance, 2000; Putnick &amp; Bornstein, 2016).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!jBef!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!jBef!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!jBef!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!jBef!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!jBef!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!jBef!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png" width="2100" height="1000" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1000,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:186581,&quot;alt&quot;:&quot;Figure 1. Four nested models&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Four nested models" title="Figure 1. Four nested models" srcset="https://substackcdn.com/image/fetch/$s_!jBef!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!jBef!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!jBef!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!jBef!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcda5bf7a-ecc7-456a-87b1-538a472d81a0_2100x1000.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. The four nested models of the multigroup analysis by gender and the results of the comparisons.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/2DmVX/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7a6b850b-aa89-462f-a97d-83b2d4a2b9fb_1220x804.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f5c3ac8d-befa-48f3-b8e1-5422cdba1665_1220x874.png&quot;,&quot;height&quot;:400,&quot;title&quot;:&quot;Table 2. The four nested models of the multigroup analysis&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/2DmVX/1/" width="730" height="400" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Each step is accepted when the more constrained model <strong>does not fit significantly worse</strong>. There are two criteria, and it is good practice to report both:</p><ul><li><p><strong>Chi-square difference test</strong>: <em>p</em> above 0.05 means the constraints are accepted. This test is sensitive to sample size (<a href="https://www.phuonglib.com/p/model-fit-indices-sem-jasp">Post 23, "SEM Model Fit Indices in JASP"</a>).</p></li><li><p><strong>Change in CFI</strong> of no more than &#8722;0.01 (Cheung &amp; Rensvold, 2002). Chen (2007) adds: RMSEA should not increase by more than 0.015, and SRMR by no more than 0.030 at the metric step and 0.010 at the scalar step.</p></li></ul><p><strong>The size of each group</strong> also matters, because each group has its own set of parameters. The Marketplace X model has 68 free parameters per group in M1; with 125 men and 137 women, that is fewer than 2 respondents per parameter. The model can still be estimated, but the <strong>power</strong> to detect small differences between the groups is <strong>low</strong>. Mention this in the limitations.</p><h2>3. Preparing the data</h2><p>Five respondents chose "Other" for gender. Five people are not enough to estimate a separate model, so they are left out of the multigroup analysis (not out of the other analyses). The syntax file <code>prepare_post29_data.sps</code> (it runs in SPSS and PSPP; checked in PSPP 2.1.2, Free Software Foundation, 2026) creates these files:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/m08a1/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8b4d46cd-1666-49c0-8119-75ed4a2aca58_1220x792.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1a77a846-f737-403c-950e-d6f09d98c7ee_1220x862.png&quot;,&quot;height&quot;:396,&quot;title&quot;:&quot;Table 3. Data files created by the syntax file&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/m08a1/1/" width="730" height="396" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The main part of the syntax:</p><pre><code>GET FILE='../data/marketplace_x_sem.sav'.
SELECT IF (GENDER = 1 OR GENDER = 2).
VALUE LABELS GENDER 1 'Male' 2 'Female'.
VARIABLE LEVEL GENDER (NOMINAL).
SAVE OUTFILE='../data/marketplace_x_multigroup.sav'.</code></pre><p>The PSPP frequency tables confirm the group sizes: 125 men and 137 women (262) in the first file, and 119 men and 131 women (250) among the respondents who answered all 20 items.</p><h2>4. Running the multigroup analysis in JASP</h2><h3>Step 1. Open the data and check the grouping variable</h3><ol><li><p><strong>Open</strong> <code>marketplace_x_multigroup.sav</code> with the three-line menu button at the top left, then <em>Open &gt; Computer</em>.</p></li><li><p><strong>Check</strong> the GENDER column: the icon with three circles (<em>Nominal</em>) and the labels <strong>Male</strong> and <strong>Female</strong> (Figure 2, boxes 1 and 2).</p></li><li><p><strong>Check</strong> that the 20 items are <em>Scale</em> (ruler icon, box 3). If they are <em>Ordinal</em>, JASP estimates the model with DWLS instead of ML (<a href="https://www.phuonglib.com/p/first-model-jasp-free-amos-alternative">Post 20, "Your First Model in JASP"</a>).</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!1Cxq!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!1Cxq!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 424w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 848w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 1272w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!1Cxq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png" width="1000" height="312" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:312,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:120343,&quot;alt&quot;:&quot;Figure 2. Multigroup data in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Multigroup data in JASP" title="Figure 2. Multigroup data in JASP" srcset="https://substackcdn.com/image/fetch/$s_!1Cxq!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 424w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 848w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 1272w, https://substackcdn.com/image/fetch/$s_!1Cxq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F88ae32e3-d39d-4230-a93b-049ec12d45b9_1000x312.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) The grouping variable GENDER is Nominal, (2) the labels Male and Female, (3) the items are Scale.</figcaption></figure></div><p>JASP orders the groups by <strong>their order of appearance in the data</strong>. The first respondent in the file is a woman, so JASP makes <strong>Female group 1 and Male group 2</strong>. This matters in Steps 4 and 5, when you write labels by group.</p><h3>Step 2. The configural model (M1)</h3><ol><li><p><strong>Choose</strong> <em>SEM &gt; Structural Equation Modeling</em>.</p></li><li><p><strong>Enter</strong> the model syntax of Post 26, with a last line that fixes the latent means to 0 (needed with FIML, see Post 26):</p></li></ol><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE   =~ PV1 + PV2 + PV3 + PV4
SATISF  =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST   =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT  =~ RI1 + RI2 + RI3 + RI4
VALUE   ~ CONFIRM
SATISF  ~ CONFIRM + VALUE
INTENT  ~ CONFIRM + VALUE + SATISF + TRUST
CONFIRM ~~ TRUST
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1</code></pre><ol><li><p><strong>Check</strong> <em>Estimation Options</em>: <em>Missing data handling</em> = <em>(FI)ML</em> (the default).</p></li><li><p><strong>Open</strong> the <em>Multigroup SEM</em> section and set <em>Grouping Variable</em> = <strong>GENDER</strong>.</p></li><li><p><strong>Press</strong> Cmd + Enter to run the model.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!PDMk!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!PDMk!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 424w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 848w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 1272w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!PDMk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png" width="1400" height="794" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:794,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:345726,&quot;alt&quot;:&quot;Figure 3. The configural model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. The configural model" title="Figure 3. The configural model" srcset="https://substackcdn.com/image/fetch/$s_!PDMk!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 424w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 848w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 1272w, https://substackcdn.com/image/fetch/$s_!PDMk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2f398edd-5341-47ea-b927-e48f0e0e7022_1400x794.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. (1) Latent means fixed to 0, (2) the grouping variable, (3) chi-square of the configural model: 347.0 with 324 degrees of freedom.</figcaption></figure></div><p>The <em>Model Fit</em> table has three rows: <em>all</em> is the overall result to report; the Female and Male rows are each group's share of the chi-square. Do not use the <em>p</em> column of these two rows for conclusions. The results pane in the screenshots is scrolled slightly to the right, so the first letters of the <em>Group</em> column are cut off.</p><h3>Step 3. Metric (M2) and scalar (M3) invariance</h3><ol><li><p><strong>Tick</strong> <em>Equality Constraints &gt; Loadings</em>. JASP reruns the model at once (Figure 4).</p></li><li><p><strong>Write down</strong> the chi-square and degrees of freedom: 371.1 and 339.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!-7vG!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!-7vG!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 424w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 848w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 1272w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!-7vG!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png" width="1400" height="794" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:794,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:357510,&quot;alt&quot;:&quot;Figure 4. Metric invariance&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Metric invariance" title="Figure 4. Metric invariance" srcset="https://substackcdn.com/image/fetch/$s_!-7vG!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 424w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 848w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 1272w, https://substackcdn.com/image/fetch/$s_!-7vG!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe8c00aab-9133-46d4-9525-6799ab9a2681_1400x794.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. (1) Loadings constrained equal, (2) fit of M2, (3) shared labels .p2., .p3. &#8230; for the loadings of both groups, (4) the Release constraints box, for freeing single constraints when only partial invariance holds.</figcaption></figure></div><p>To test the intercepts:</p><ol><li><p><strong>Tick</strong> <em>Intercepts</em> as well.</p></li><li><p><strong>Change</strong> the last line of the syntax to <code>CONFIRM + VALUE + SATISF + TRUST + INTENT ~ c(0, NA)*1</code>. The 0 fixes the latent means of group 1 (Female, the reference group); <code>NA</code> frees the means of group 2 (Male). If you keep <code>0*1</code> for both groups, any difference in mean level is pushed into the intercepts and the test is wrong.</p></li><li><p><strong>Press</strong> Cmd + Enter.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!VFnR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!VFnR!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 424w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 848w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 1272w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!VFnR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png" width="1400" height="785" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:785,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:366815,&quot;alt&quot;:&quot;Figure 5. Scalar invariance&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Scalar invariance" title="Figure 5. Scalar invariance" srcset="https://substackcdn.com/image/fetch/$s_!VFnR!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 424w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 848w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 1272w, https://substackcdn.com/image/fetch/$s_!VFnR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F12f5f180-8afa-4ca0-8d48-5ff9815c1c37_1400x785.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. (1) Loadings and intercepts constrained equal, (2) latent means of the Male group freed, (3) fit of M3: 381.3 with 354 degrees of freedom.</figcaption></figure></div><h3>In AMOS</h3><p>AMOS fits the same nested models on the 250 complete respondents (<code>marketplace_x_multigroup_complete.sav</code>). Define the groups Male and Female in <strong>Analyze &gt; Manage Groups</strong>, then link each group to the file in <strong>File &gt; Data Files</strong> with <em>Grouping Variable</em> GENDER and <em>Group Value</em> 1 (Male) or 2 (Female). Tick <em>Estimate means and intercepts</em> (<em>Estimation</em> tab) and <em>Critical ratios for differences</em> (<em>Output</em> tab) in <strong>View &gt; Analysis Properties</strong>. Then choose <strong>Analyze &gt; Multiple-Group Analysis</strong>: AMOS labels every parameter by group and creates the <em>Unconstrained</em>, <em>Measurement weights</em>, <em>Measurement intercepts</em> and <em>Structural weights</em> models in one run. The <em>Model Comparison</em> section of the Text Output prints the chi-square differences. In AMOS, Male is group 1 and the reference group whose latent means are fixed to 0. With complete cases the conclusions are the same as below: metric 23.82 (15 df), <em>p</em> = 0.068; scalar 12.05 (15 df), <em>p</em> = 0.675; seven equal paths (run again without means) 7.13 (7 df), <em>p</em> = 0.415.</p><h2>5. Results of the measurement invariance tests</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/ALK62/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3b74c38c-7417-444f-9964-0415efbfe44a_1220x666.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b90ca81f-1826-46a4-89ac-464880bcf60e_1220x802.png&quot;,&quot;height&quot;:365,&quot;title&quot;:&quot;Table 4. Measurement invariance tests by gender&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/ALK62/1/" width="730" height="365" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The JASP <em>Model Fit</em> table shows only the chi-square; CFI, RMSEA and SRMR are added by ticking <em>Additional fit measures</em> in <em>Output Options</em> (Post 20). The difference test can be computed by hand: 371.13 &#8722; 346.97 = 24.17 with 339 &#8722; 324 = 15 degrees of freedom; the <em>p</em>-value comes from <code>=CHISQ.DIST.RT(24.17, 15)</code> in Excel.</p><p>Reading the results:</p><ul><li><p><strong>The configural model fits well</strong> (CFI 0.989, RMSEA 0.023): both groups have the same five-factor structure.</p></li><li><p><strong>Metric invariance holds</strong>, but it is borderline by the chi-square test: <em>p</em> = 0.062 (FIML) and 0.068 (complete cases). The CFI change is only &#8722;0.004, well inside the 0.01 threshold. SRMR increases by 0.010, below Chen's (2007) 0.030 threshold.</p></li><li><p><strong>Scalar invariance holds</strong> clearly (<em>p</em> = 0.809; CFI does not drop).</p></li></ul><p>A score test on M2 (in lavaan) shows which item deviates most: the loading of <strong>PV2</strong> (1-df chi-square = 5.78, <em>p</em> = 0.016). With 15 loadings tested at the same time, a <em>p</em>-value of about 0.016 can easily occur by chance (the Bonferroni threshold is 0.05/15 = 0.003), and the overall test does not reject invariance. So this post does <strong>not</strong> free the constraint on PV2. If the overall test had been rejected, the remedy would be <strong>partial invariance</strong>: free the constraints of the one or two items that deviate most, as long as each factor keeps at least two constrained items (Byrne, 2016).</p><p>Once scalar invariance holds, the <strong>latent means</strong> of the two groups can be compared. The <em>Intercepts</em> table of model M3 (Figure 6) gives the means of the Male group relative to the Female group (fixed to 0):</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!kRLo!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!kRLo!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 424w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 848w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 1272w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!kRLo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png" width="760" height="309" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:309,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:89298,&quot;alt&quot;:&quot;Figure 6. Latent means of the two groups&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Latent means of the two groups" title="Figure 6. Latent means of the two groups" srcset="https://substackcdn.com/image/fetch/$s_!kRLo!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 424w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 848w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 1272w, https://substackcdn.com/image/fetch/$s_!kRLo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fceb89aae-7372-4f81-843b-0c22d219ba85_760x309.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. The Intercepts table of M3 (the rows with the item intercepts are left out): (1) latent means of the Female group fixed to 0, (2) latent means of the Male group relative to the Female group.</figcaption></figure></div><p><strong>Latent means.</strong> No construct has a different mean for men and women: the largest difference is perceived value, where men score 0.11 lower than women (<em>p</em> = 0.234); repurchase intention is almost identical (difference 0.01; <em>p</em> = 0.914). This conclusion is <strong>valid only because scalar invariance holds</strong>. Comparing mean scores between two groups without testing invariance (for example, a t-test on the average scores, <a href="https://www.phuonglib.com/p/t-test-independent-paired-one-sample-pspp-spss">Post 17, "T-Tests in PSPP (SPSS)"</a>) silently assumes it.</p><h2>6. Comparing the paths between men and women</h2><h3>A JASP 0.99.0 pitfall to know about</h3><p>The obvious way is to untick <em>Intercepts</em>, restore the <code>~ 0*1</code> line and tick <em>Equality Constraints &gt; Regressions</em>. When you try this, <strong>the chi-square does not change</strong> (still 371.1 with 339 degrees of freedom), which means the constraint was not applied. Putting shared labels in the syntax, such as <code>c(h1, h1)*CONFIRM</code>, gives exactly the same result.</p><p>The cause: to compute the <em>Indirect Effects</em> and <em>Total Effects</em> tables, JASP gives each path in each group its own label, which overrides the user's constraints. JASP <strong>skips this step only when the syntax contains at least one defined parameter</strong> with <code>:=</code>. So the correct way is: put the labels in the syntax <strong>and</strong> add at least one <code>:=</code> line. The <em>Regression Coefficients</em> table of model M2 (Figure 7) shows what JASP prints without such a line: no labels, and a separate coefficient for each group.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!7fpx!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!7fpx!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 424w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 848w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 1272w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!7fpx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png" width="760" height="359" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:359,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:134686,&quot;alt&quot;:&quot;Figure 7. Path coefficients of each group in model M2&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Path coefficients of each group in model M2" title="Figure 7. Path coefficients of each group in model M2" srcset="https://substackcdn.com/image/fetch/$s_!7fpx!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 424w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 848w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 1272w, https://substackcdn.com/image/fetch/$s_!7fpx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F72a474e2-f95a-47ea-9942-65694c6693ae_760x359.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Path coefficients of each group in the equal-loadings model: (1) VALUE to INTENT in the Female group, (2) the same path in the Male group.</figcaption></figure></div><h3>Way 1. Constrain all seven paths at once (M4)</h3><ol><li><p><strong>Untick</strong> <em>Intercepts</em> and keep <em>Loadings</em>.</p></li><li><p><strong>Edit</strong> the syntax: give the same label to both groups and add one defined parameter. The indirect effect <code>CONF_PV_RI</code> (CONFIRM to VALUE to INTENT) was chosen because it is meaningful for the study (Post 28), not only to work around the pitfall:</p></li></ol><pre><code>VALUE   ~ c(h1, h1)*CONFIRM
SATISF  ~ c(h2, h2)*CONFIRM + c(h3, h3)*VALUE
INTENT  ~ c(h4, h4)*CONFIRM + c(h5, h5)*VALUE + c(h6, h6)*SATISF + c(h7, h7)*TRUST
CONFIRM ~~ TRUST
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1
CONF_PV_RI := h1*h5</code></pre><ol><li><p><strong>Press</strong> Cmd + Enter. The number of free parameters drops from 121 to 114 and the degrees of freedom rise from 339 to 346: the constraints are now in effect.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!6-5C!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!6-5C!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 424w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 848w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 1272w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!6-5C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png" width="1400" height="383" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:383,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:176620,&quot;alt&quot;:&quot;Figure 8. The model with seven equal paths&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. The model with seven equal paths" title="Figure 8. The model with seven equal paths" srcset="https://substackcdn.com/image/fetch/$s_!6-5C!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 424w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 848w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 1272w, https://substackcdn.com/image/fetch/$s_!6-5C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd9016e35-7aac-4ea5-952a-f7bb2b0a6450_1400x383.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. (1) Seven paths with labels shared by both groups, (2) the defined parameter that makes JASP keep the user's labels, (3) fit of M4: 114 free parameters, chi-square 377.6 with 346 degrees of freedom.</figcaption></figure></div><p>In M4, both groups now show the same coefficient for each path, and the <em>Defined Parameters</em> table gives CONF_PV_RI = 0.132 (<em>p</em> = .017).</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/iPiv1/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fbab85b7-91e1-4745-9b0f-1c8f5ff1949b_1220x428.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/40efba6b-fde1-459b-9e96-e723cc160c1a_1220x614.png&quot;,&quot;height&quot;:288,&quot;title&quot;:&quot;Table 5. Comparing the model with seven equal paths with the metric model&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/iPiv1/1/" width="730" height="288" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Constraining the seven paths to be equal <strong>does not make the model worse</strong>. Overall, gender <strong>does not moderate</strong> the model.</p><h3>Way 2. Test each path separately</h3><p>An overall test can hide a difference in one path. To test each path, give each group <strong>its own</strong> label and define the differences. Remember that group 1 is Female and group 2 is Male:</p><pre><code>VALUE   ~ c(h1f, h1m)*CONFIRM
SATISF  ~ c(h2f, h2m)*CONFIRM + c(h3f, h3m)*VALUE
INTENT  ~ c(h4f, h4m)*CONFIRM + c(h5f, h5m)*VALUE + c(h6f, h6m)*SATISF + c(h7f, h7m)*TRUST
CONFIRM ~~ TRUST
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1
diff_H1 := h1m - h1f
diff_H2 := h2m - h2f
diff_H3 := h3m - h3f
diff_H4 := h4m - h4f
diff_H5 := h5m - h5f
diff_H6 := h6m - h6f
diff_H7 := h7m - h7f</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!FUTv!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!FUTv!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 424w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 848w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 1272w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!FUTv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png" width="700" height="282" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:282,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:106644,&quot;alt&quot;:&quot;Figure 9. Syntax with separate labels per group&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Syntax with separate labels per group" title="Figure 9. Syntax with separate labels per group" srcset="https://substackcdn.com/image/fetch/$s_!FUTv!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 424w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 848w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 1272w, https://substackcdn.com/image/fetch/$s_!FUTv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F713c5b6f-d9a3-4d41-8b02-c5693661df59_700x282.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. (1) Separate labels for the Female group (first position) and the Male group (second position), (2) seven differences, Male minus Female. Line 8 is longer than the editor, so the end of it is hidden.</figcaption></figure></div><p>The chi-square is still 371.1 (model M2), because no constraint was added. The <em>Defined Parameters</em> table gives a z test for each difference (Figure 10). This is the equivalent of AMOS's <strong>Critical ratios for differences</strong>.</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!HkLn!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 424w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 848w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 1272w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!HkLn!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png" width="600" height="220" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:220,&quot;width&quot;:600,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:67937,&quot;alt&quot;:&quot;Figure 10. Tests of the path differences&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. Tests of the path differences" title="Figure 10. Tests of the path differences" srcset="https://substackcdn.com/image/fetch/$s_!HkLn!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 424w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 848w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 1272w, https://substackcdn.com/image/fetch/$s_!HkLn!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ee64277-f3f7-4690-86a4-269494e1afa5_600x220.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 10. Differences in the paths between men and women; (1) perceived value to intention: men are 0.214 lower than women, z = &#8722;1.28, p = .201.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!o-hf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!o-hf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 424w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 848w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 1272w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!o-hf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png" width="2100" height="1080" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1080,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:132026,&quot;alt&quot;:&quot;Figure 11. Path coefficients of men and women&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 11. Path coefficients of men and women" title="Figure 11. Path coefficients of men and women" srcset="https://substackcdn.com/image/fetch/$s_!o-hf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 424w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 848w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 1272w, https://substackcdn.com/image/fetch/$s_!o-hf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d4a2951-0744-4089-a6bd-cff347839504_2100x1080.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 11. Unstandardized path coefficients and 95% confidence intervals of the two groups (model M2).</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/RYB61/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d6d7c76a-1138-4c6e-9ab9-a4b923312bf3_1220x990.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dd036f39-4902-4ea9-b28b-e9087b269fe5_1220x1232.png&quot;,&quot;height&quot;:586,&quot;title&quot;:&quot;Table 6. Path coefficients of men and women and tests of the differences&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/RYB61/1/" width="730" height="586" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><h3>An interpretation error to avoid</h3><p>Look at H5: the coefficient is <strong>significant for women</strong> (<em>p</em> &lt; 0.01) but <strong>not significant for men</strong> (<em>p</em> close to 0.4). It is very tempting to write "perceived value affects the repurchase intention of women only; gender moderates this relationship". That conclusion is <strong>wrong</strong>. The direct test of the difference gives a <em>p</em> of about 0.2: the two coefficients are <strong>not significantly different</strong>. The confidence intervals of the two groups in Figure 11 overlap almost completely.</p><p>The difference between "significant" and "not significant" is not itself statistically significant (Gelman &amp; Stern, 2006). A coefficient of 0.32 with <em>p</em> = 0.005 and one of 0.11 with <em>p</em> = 0.37 can easily come from the same true value and differ only through sampling error. <strong>A conclusion about moderation must rest on a test of the difference</strong>, not on a comparison of two <em>p</em>-values.</p><p>A note on power: with about 125 people per group, the standard error of the H5 difference is about 0.17. The true difference would have to be about 0.47 for the test to have 80% power. The "no difference" result here should be written as <strong>no difference was detected</strong>, not "the two groups are proved to be the same".</p><h2>7. A continuous moderator: the latent interaction</h2><p>An additional hypothesis: <strong>the higher the trust, the weaker the effect of satisfaction on repurchase intention</strong>, because customers who trust the marketplace already have a reason to come back even when this purchase was not fully satisfying. It is tested with the product SATISF &#215; TRUST.</p><p>In CB-SEM, the product of two latent variables cannot be observed. There are three main approaches (Marsh et al., 2004):</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/NN1wM/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2a5496ca-6f34-4d46-830c-45ff7e61db1a_1220x558.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2c0ed469-612e-4391-a878-4b69347ca529_1220x628.png&quot;,&quot;height&quot;:293,&quot;title&quot;:&quot;Table 7. Three approaches to a latent interaction&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/NN1wM/1/" width="730" height="293" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>This post uses <strong>matched-pair product indicators with double-mean centering</strong> (Marsh et al., 2004; Lin et al., 2010), because it works in both AMOS and JASP:</p><ol><li><p><strong>Pair</strong> the items: SAT1 with TRU1, SAT2 with TRU2, SAT3 with TRU3, SAT4 with TRU4R. Each item is used only once, so the 4 product indicators have no correlated errors.</p></li><li><p><strong>Center</strong> each item: subtract the item's mean.</p></li><li><p><strong>Multiply</strong> the pairs.</p></li><li><p><strong>Center again</strong>: subtract the mean of each product. This avoids having to estimate an extra, complicated mean structure (Lin et al., 2010).</p></li></ol><p>Part B of <code>prepare_post29_data.sps</code> performs these four steps with <code>AGGREGATE</code> and <code>COMPUTE</code> and saves <code>marketplace_x_product.sav</code> with the variables SATxTRU1 to SATxTRU4. The product indicators equal those of the <code>indProd()</code> function in the semTools package (Jorgensen et al., 2022) to the 15th decimal.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!wUb9!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!wUb9!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!wUb9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png" width="2100" height="1000" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1000,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:195132,&quot;alt&quot;:&quot;Figure 12. The latent interaction model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 12. The latent interaction model" title="Figure 12. The latent interaction model" srcset="https://substackcdn.com/image/fetch/$s_!wUb9!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!wUb9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd834deda-b849-4eb8-b0eb-b9a73cd66fa7_2100x1000.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 12. The interaction SATxTRU is measured by 4 product indicators and affects repurchase intention through path h8.</figcaption></figure></div><h3>Steps in JASP</h3><ol><li><p><strong>Open</strong> <code>marketplace_x_product.sav</code>. Check that SATxTRU1 to SATxTRU4 are <em>Scale</em>.</p></li><li><p><strong>Open</strong> a new <em>Structural Equation Modeling</em> analysis (without a grouping variable).</p></li><li><p><strong>Enter</strong> the syntax:</p></li></ol><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE   =~ PV1 + PV2 + PV3 + PV4
SATISF  =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST   =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT  =~ RI1 + RI2 + RI3 + RI4
SATxTRU =~ SATxTRU1 + SATxTRU2 + SATxTRU3 + SATxTRU4
VALUE   ~ CONFIRM
SATISF  ~ CONFIRM + VALUE
INTENT  ~ CONFIRM + VALUE + h6*SATISF + h7*TRUST + h8*SATxTRU
CONFIRM ~~ TRUST + SATxTRU
TRUST   ~~ vTRU*TRUST + SATxTRU
CONFIRM + VALUE + SATISF + TRUST + INTENT + SATxTRU ~ 0*1
slope_TRU_low  := h6 - h8*sqrt(vTRU)
slope_TRU_mean := h6
slope_TRU_high := h6 + h8*sqrt(vTRU)</code></pre><p>The label <code>vTRU</code> is attached to the variance of TRUST so that the <strong>simple slopes</strong> of SATISF can be computed at low trust (mean minus 1 standard deviation), average trust and high trust (plus 1 standard deviation). The three <code>:=</code> lines also make JASP keep the labels h6, h7 and h8 (Section 6).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!aemw!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!aemw!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 424w, https://substackcdn.com/image/fetch/$s_!aemw!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 848w, https://substackcdn.com/image/fetch/$s_!aemw!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 1272w, https://substackcdn.com/image/fetch/$s_!aemw!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!aemw!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png" width="1400" height="383" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:383,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:199185,&quot;alt&quot;:&quot;Figure 13. The interaction model in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 13. The interaction model in JASP" title="Figure 13. The interaction model in JASP" srcset="https://substackcdn.com/image/fetch/$s_!aemw!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 424w, https://substackcdn.com/image/fetch/$s_!aemw!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 848w, https://substackcdn.com/image/fetch/$s_!aemw!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 1272w, https://substackcdn.com/image/fetch/$s_!aemw!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8936505d-c4ed-40de-baa1-9d8e83428ffb_1400x383.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 13. (1) The interaction measured by 4 product indicators, (2) path h8, (3) simple slopes, (4) fit: chi-square 229.5 with 241 degrees of freedom.</figcaption></figure></div><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!hvnd!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 424w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 848w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 1272w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!hvnd!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png" width="700" height="227" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:227,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:82499,&quot;alt&quot;:&quot;Figure 14. The interaction coefficient&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 14. The interaction coefficient" title="Figure 14. The interaction coefficient" srcset="https://substackcdn.com/image/fetch/$s_!hvnd!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 424w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 848w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 1272w, https://substackcdn.com/image/fetch/$s_!hvnd!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcba9b15c-4c8f-41b8-a1a8-bc0c3974a8f7_700x227.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 14. (1) The effect of SATISF when TRUST is at its mean, (2) the interaction coefficient h8 = &#8722;0.041; p = .615.</figcaption></figure></div><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!LMUu!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 424w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 848w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 1272w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!LMUu!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png" width="600" height="131" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:131,&quot;width&quot;:600,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:36615,&quot;alt&quot;:&quot;Figure 15. Simple slopes&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 15. Simple slopes" title="Figure 15. Simple slopes" srcset="https://substackcdn.com/image/fetch/$s_!LMUu!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 424w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 848w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 1272w, https://substackcdn.com/image/fetch/$s_!LMUu!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb877649d-166f-4f94-8d2a-4394dd36a15d_600x131.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 15. (1) The slope of SATISF at three levels of TRUST.</figcaption></figure></div><h3>Results (FIML, N = 267)</h3><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/cX79C/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6c815b5f-8538-4028-ab6a-c503d9cbc848_1220x742.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b937664b-c6aa-40e4-9b1a-5be38d18ae97_1220x928.png&quot;,&quot;height&quot;:353,&quot;title&quot;:&quot;Table 8. Paths to repurchase intention in the latent interaction model&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/cX79C/1/" width="730" height="353" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>In AMOS, the same model is drawn on the 254 complete respondents (<code>marketplace_x_product_complete.sav</code>), with SATxTRU1 to SATxTRU4 as the indicators of a latent variable SATxTRU. The interaction coefficient is h8 = &#8722;0.055 (SE 0.082, <em>p</em> = 0.504; beta = &#8722;0.050), so the conclusion is the same.</p><p>Reading the results:</p><ul><li><p><strong>The interaction coefficient h8 is negative</strong> (in the hypothesized direction) <strong>but very small and not significant</strong> (<em>p</em> = 0.615; the confidence interval contains 0). The R&#178; of repurchase intention rises only from 0.444 to 0.445 when the interaction is added (f&#178; = 0.0003; lavaan). Trust <strong>does not moderate</strong> the effect of satisfaction.</p></li><li><p><strong>h6 is now the effect of satisfaction when trust is at its mean</strong>, no longer the "overall effect" of Post 26. This is the correct way to read the main coefficient when an interaction is in the model.</p></li><li><p>Ignore the fit indices of a model with product indicators: the product of two variables is not normally distributed, so the ML chi-square is no longer accurate. The purpose of this model is to test h8.</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!pYu5!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!pYu5!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 424w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 848w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 1272w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!pYu5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png" width="2100" height="960" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:960,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:179414,&quot;alt&quot;:&quot;Figure 16. Simple slopes&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 16. Simple slopes" title="Figure 16. Simple slopes" srcset="https://substackcdn.com/image/fetch/$s_!pYu5!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 424w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 848w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 1272w, https://substackcdn.com/image/fetch/$s_!pYu5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fae8cd9a6-b6df-48ba-8a3d-0311cbe06249_2100x960.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 16. The effect of satisfaction on repurchase intention at three levels of trust (FIML, N = 267).</figcaption></figure></div><p>Figure 16 repeats the lesson of Section 6 in another form: the slope of satisfaction is significant at low trust (<em>p</em> = 0.034) and average trust (<em>p</em> = 0.024), and <strong>not</strong> significant at high trust (<em>p</em> = 0.088). But the three slopes (0.245, 0.215, 0.185) are almost equal and the three lines are nearly parallel. The test of moderation is the test of h8, not a count of which slopes pass the 0.05 threshold.</p><p><strong>Trying a median split.</strong> If customers are split into a low-trust group (mean TRU score of 4 or less, 151 people) and a high-trust group (above 4, 116 people) and a multigroup model is run, the effect of satisfaction is 0.279 (<em>p</em> = 0.035) in the low group and 0.205 (<em>p</em> = 0.234) in the high group. Again a "significant, not significant" pair, but the difference is not significant (chi-square difference 0.12; <em>p</em> = 0.733; lavaan). The split also creates groups of unequal size, because many people score exactly 4 (the median), and it puts a person scoring 3.75 in the same group as a person scoring 1 (MacCallum et al., 2002).</p><h2>8. Reporting the results</h2><blockquote><p>A multigroup analysis by gender was run with FIML estimation on 262 respondents (125 men, 137 women; 5 respondents who chose another gender were excluded because the group was too small). The configural model fitted well (chi-square = 346.97; df = 324; CFI = 0.989; RMSEA = 0.023). Constraining the loadings to be equal did not significantly worsen the fit (chi-square difference = 24.17; df = 15; <em>p</em> = 0.062; CFI change = &#8722;0.004), nor did additionally constraining the intercepts (chi-square difference = 10.16; df = 15; <em>p</em> = 0.809; CFI change = +0.002), indicating metric and scalar invariance across genders (Cheung &amp; Rensvold, 2002). Constraining the seven path coefficients to be equal did not worsen the fit (chi-square difference = 6.49; df = 7; <em>p</em> = 0.484), and no coefficient differed significantly between the groups when tested separately (<em>p</em> from 0.164 to 0.814). Gender did not moderate the relationships in the model. The latent interaction between satisfaction and trust, measured by four matched-pair, double-mean-centered product indicators (Lin et al., 2010), had no significant effect on repurchase intention (B = &#8722;0.041; 95% CI [&#8722;0.202, 0.119]). With about 130 respondents per group, the power to detect small differences between the groups was limited.</p></blockquote><h2>Five common mistakes</h2><ol><li><p><strong>Comparing the p-values of two groups</strong> instead of testing the difference: concluding "there is moderation" because a coefficient is significant in one group and not in the other.</p></li><li><p><strong>Skipping measurement invariance</strong> and comparing paths or mean scores straight away.</p></li><li><p><strong>Using only the chi-square difference</strong> with a large sample: even tiny differences become significant. Report the changes in CFI, RMSEA and SRMR as well.</p></li><li><p><strong>Splitting a continuous variable</strong> at the median to run a multigroup analysis.</p></li><li><p><strong>Writing "the two groups are the same"</strong> when the test is not significant with small groups. Write "no difference was detected" and mention the limited power.</p></li></ol><h2>Frequently asked questions</h2><p><strong>How many respondents does each group need?</strong></p><p>There is no fixed number. Each group must be able to estimate the model on its own (the configural model), so the sample size rules of <a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a> apply to <strong>each group</strong>. About 100 to 150 respondents per group is a common minimum for a model the size of the Marketplace X model; detecting a path difference of about 0.2 needs considerably more.</p><p><strong>What if metric invariance does not hold?</strong></p><p>Use a score test (JASP: <em>Modification indices</em> in model M2; AMOS: <em>Modification Indices</em> with a low threshold) to find the item that deviates most, free its constraint (<em>Release constraints</em> in JASP, a separate label in AMOS) and test again. Report clearly which item was freed and why.</p><p><strong>Can more than two groups be compared?</strong></p><p>Yes, for example four income groups. The procedure is the same; the overall test has more degrees of freedom. When the overall test is significant, compare the groups in pairs and correct for multiple comparisons.</p><p><strong>How does SmartPLS handle multigroup analysis?</strong></p><p>SmartPLS uses MICOM to test invariance and PLS-MGA or a permutation test to compare paths, and it offers more ways to build interaction terms (<a href="https://www.phuonglib.com/p/moderation-analysis-smartpls">Post 41, "Moderation Analysis in SmartPLS 4"</a>).</p><div><hr></div><h2>References</h2><p>Baron, R. M., &amp; Kenny, D. A. (1986). The moderator&#8211;mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. <em>Journal of Personality and Social Psychology, 51</em>(6), 1173&#8211;1182. https://doi.org/10.1037/0022-3514.51.6.1173</p><p>Byrne, B. M. (2016). <em>Structural equation modeling with AMOS: Basic concepts, applications, and programming</em> (3rd ed.). Routledge.</p><p>Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. <em>Structural Equation Modeling, 14</em>(3), 464&#8211;504. https://doi.org/10.1080/10705510701301834</p><p>Cheung, G. W., &amp; Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. <em>Structural Equation Modeling, 9</em>(2), 233&#8211;255. https://doi.org/10.1207/S15328007SEM0902_5</p><p>Free Software Foundation. (2026). <em>GNU PSPP</em> (Version 2.1.2) [Computer software]. https://www.gnu.org/software/pspp/</p><p>Gelman, A., &amp; Stern, H. (2006). The difference between "significant" and "not significant" is not itself statistically significant. <em>The American Statistician, 60</em>(4), 328&#8211;331. https://doi.org/10.1198/000313006X152649</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Jorgensen, T. D., Pornprasertmanit, S., Schoemann, A. M., &amp; Rosseel, Y. (2022). <em>semTools: Useful tools for structural equation modeling</em> (R package version 0.5-6) [Computer software]. https://CRAN.R-project.org/package=semTools</p><p>Lin, G.-C., Wen, Z., Marsh, H. W., &amp; Lin, H.-S. (2010). Structural equation models of latent interactions: Clarification of orthogonalizing and double-mean-centering strategies. <em>Structural Equation Modeling, 17</em>(3), 374&#8211;391. https://doi.org/10.1080/10705511.2010.488999</p><p>MacCallum, R. C., Zhang, S., Preacher, K. J., &amp; Rucker, D. D. (2002). On the practice of dichotomization of quantitative variables. <em>Psychological Methods, 7</em>(1), 19&#8211;40. https://doi.org/10.1037/1082-989X.7.1.19</p><p>Marsh, H. W., Wen, Z., &amp; Hau, K.-T. (2004). Structural equation models of latent interactions: Evaluation of alternative estimation strategies and indicator construction. <em>Psychological Methods, 9</em>(3), 275&#8211;300. https://doi.org/10.1037/1082-989X.9.3.275</p><p>Putnick, D. L., &amp; Bornstein, M. H. (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. <em>Developmental Review, 41</em>, 71&#8211;90. https://doi.org/10.1016/j.dr.2016.06.004</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p><p>Vandenberg, R. J., &amp; Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. <em>Organizational Research Methods, 3</em>(1), 4&#8211;70. https://doi.org/10.1177/109442810031002</p>]]></content:encoded></item><item><title><![CDATA[Predictive Power in SmartPLS 4: Q²predict, PLSpredict and CVPAT]]></title><description><![CDATA[Set up PLSpredict, read Q&#178;predict, compare RMSE with the linear model and run CVPAT to test out-of-sample prediction of a PLS-SEM model]]></description><link>https://www.phuonglib.com/p/plspredict-cvpat-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/plspredict-cvpat-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:49:19 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/c06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!V2hQ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!V2hQ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!V2hQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:955526,&quot;alt&quot;:&quot;An archer on a near practice field shooting at a new target half hidden in mist across a fence, an orange arrow close to its center.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="An archer on a near practice field shooting at a new target half hidden in mist across a fence, an orange arrow close to its center." title="An archer on a near practice field shooting at a new target half hidden in mist across a fence, an orange arrow close to its center." srcset="https://substackcdn.com/image/fetch/$s_!V2hQ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!V2hQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc06c0107-1499-4dd8-baec-47cbd53a3575_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>R&#178; is in-sample explanatory power; PLSpredict measures how well the model predicts respondents who were not used to estimate it.</p></li><li><p>Read PLSpredict in two steps: Q&#178;predict above 0 for every indicator, then compare PLS-SEM RMSE with the linear model (LM) benchmark.</p></li><li><p>The Marketplace X model has Q&#178;predict from 0.085 to 0.213 and a lower RMSE than LM for all 12 indicators: high predictive power, though the RMSE gaps are small.</p></li><li><p>CVPAT tests whether the advantage is significant: overall yes (p &lt; 0.001), but not for repurchase intention against LM (p = 0.137), and that result must be reported.</p></li><li><p>Use Fixed seed and report the number of folds, repetitions and prediction type, because the results depend on how the folds are split.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/evaluate-structural-model-smartpls">Post 37</a> showed that the Marketplace X model explains 31.8% of the variance of repurchase intention (R&#178; = 0.318). R&#178; measures <strong>in-sample</strong> explanatory power: the model is estimated and evaluated on the same data. This post asks a different question: if you apply the model to <strong>people who were not used to estimate it</strong>, how well does it predict their answers? This is step 5 of the structural model assessment of Hair et al. (2019): <strong>out-of-sample predictive power</strong>.</p><p>This post uses two tools built into SmartPLS 4:</p><ul><li><p><strong>PLSpredict</strong> (Shmueli et al., 2016, 2019): computes Q&#178;predict and the prediction errors RMSE and MAE for each indicator.</p></li><li><p><strong>CVPAT</strong> (Liengaard et al., 2021; Sharma et al., 2023): tests whether the prediction error of PLS-SEM is <strong>significantly lower</strong> than that of a benchmark model.</p></li></ul><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch38-plspredict-cvpat">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code> from <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>, with casewise deletion on 254 respondents. The Student license does not include PLSpredict and CVPAT. The numbers in this post come from the SmartPLS report. SmartPLS lists constructs and indicators in alphabetical order (INTENT, SATISF, VALUE; PV, RI, SAT). Section 6 checks the results in R (script <code>plspredict.R</code> in the repository's <code>scripts/</code> folder, cSEM package).</p></blockquote><h2>1. Why not use the blindfolding Q&#178;?</h2><p>Predictive power used to be assessed with the <strong>Stone-Geisser Q&#178;</strong> (Geisser, 1974; Stone, 1974), computed by the blindfolding procedure: part of the data points is omitted in turn, the model is re-estimated and the omitted points are predicted. Shmueli et al. (2016) showed that this Q&#178; still uses information from the very observations being predicted when the model is estimated, so it is <strong>closer to in-sample explanatory power than to out-of-sample prediction</strong>. For this reason, Hair et al. (2019) recommend PLSpredict for step 5, as does the primer by Hair et al. (2022).</p><p>If earlier studies or your supervisor ask for the blindfolding Q&#178;, you can report it as well, but state clearly that it is the blindfolding Q&#178; and that it does not replace PLSpredict.</p><h2>2. How does PLSpredict work?</h2><p>PLSpredict uses <strong>k-fold cross-validation</strong>:</p><ol><li><p><strong>Split</strong> the 254 respondents at random into 10 folds of about 25 respondents each.</p></li><li><p><strong>Estimate</strong> the model on 9 folds (the training sample, about 229 respondents).</p></li><li><p><strong>Predict</strong> the answers of the remaining fold (the holdout sample) from the exogenous constructs of the model.</p></li><li><p><strong>Repeat</strong> steps 2 and 3 until every fold has served as the holdout sample once.</p></li><li><p><strong>Repeat</strong> the whole procedure 10 times with different splits, then average.</p></li></ol><p>The result is the out-of-sample prediction error for each indicator of the <strong>endogenous constructs</strong> (the constructs being predicted: VALUE, SATISF and INTENT). The indicators of CONFIRM and TRUST do not appear in the results, because these exogenous constructs are the predictors.</p><p>The error of PLS-SEM is compared with two benchmarks:</p><ul><li><p><strong>IA</strong> (indicator average): predicts every respondent with the <strong>mean</strong> of the indicator in the training sample. This is the simplest possible prediction. <strong>Q&#178;predict = 1 - (sum of squared prediction errors of PLS-SEM) / (sum of squared prediction errors of IA)</strong>. A Q&#178;predict above 0 means that PLS-SEM predicts better than the mean.</p></li><li><p><strong>LM</strong> (linear model): regresses each indicator of an endogenous construct on <strong>all indicators</strong> of the exogenous constructs, without the structure of the model. This is a harder benchmark: PLS-SEM must predict at least as well as LM for the theoretical structure to add predictive value.</p></li></ul><p>Two error measures:</p><ul><li><p><strong>RMSE</strong> (root mean squared error): penalizes large errors more heavily. Used by default.</p></li><li><p><strong>MAE</strong> (mean absolute error): used when the distribution of the prediction errors is highly skewed (Shmueli et al., 2019). You can view this distribution under <strong>Histograms &gt; PLS-SEM MV error histogram</strong> in the report.</p></li></ul><h2>3. Running PLSpredict in SmartPLS 4</h2><p><strong>Step 1.</strong> Open the model <code>marketplace-x-model</code> in the project Marketplace X. Choose the menu <strong>Calculate &gt; PLSpredict / CVPAT</strong>.</p><p><strong>Step 2.</strong> On the <strong>PLSpredict / CVPAT</strong> tab of the dialog, check four settings (Figure 1):</p><ul><li><p><strong>Number of folds</strong>: 10.</p></li><li><p><strong>No. of repetitions</strong>: 10.</p></li><li><p><strong>Prediction type</strong>: Early antecedents, that is, predict from the first exogenous constructs of the path chain.</p></li><li><p><strong>Random number generator</strong>: Fixed seed, so that a rerun gives the same results. With Random seed, the numbers vary slightly from run to run.</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!mJkJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!mJkJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!mJkJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:139212,&quot;alt&quot;:&quot;Figure 1. PLSpredict settings&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. PLSpredict settings" title="Figure 1. PLSpredict settings" srcset="https://substackcdn.com/image/fetch/$s_!mJkJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!mJkJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91bdd07f-fa3c-49b1-b6b6-63f92cf67270_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. The PLSpredict / CVPAT tab: (1) number of folds and repetitions, (2) prediction type, (3) random number generator, (4) the Start calculation button.</figcaption></figure></div><p><strong>Step 3.</strong> Open the <strong>Data</strong> tab and set <strong>Missing value algorithm</strong> to <strong>Casewise deletion</strong>, as in the previous posts (Figure 2).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Ee__!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Ee__!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Ee__!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png" width="700" height="791" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:791,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:118672,&quot;alt&quot;:&quot;Figure 2. Data tab of PLSpredict&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Data tab of PLSpredict" title="Figure 2. Data tab of PLSpredict" srcset="https://substackcdn.com/image/fetch/$s_!Ee__!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 424w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 848w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 1272w, https://substackcdn.com/image/fetch/$s_!Ee__!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc53bac35-b10f-4290-9de4-82eb16c29d84_700x791.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. The Data tab: (1) the Data tab, (2) missing values handled by casewise deletion.</figcaption></figure></div><p><strong>Step 4.</strong> Click <strong>Start calculation</strong>. With 10 folds and 10 repetitions, SmartPLS estimates the model 100 times, so it takes a few seconds.</p><blockquote><p><strong>Sample size in each fold.</strong> Each training sample contains only about 90% of the respondents. With small samples, use fewer folds (for example 5) so that the training sample does not become too small (Shmueli et al., 2019). With 254 respondents, 10 folds are appropriate.</p></blockquote><h2>4. Reading the PLSpredict results</h2><h3>Q&#178;predict and RMSE of each indicator</h3><p><strong>Step 5.</strong> In the report, open <strong>Final results &gt; PLSpredict MV summary &gt; Overview</strong> (Figure 3).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!hJ_A!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!hJ_A!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 424w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 848w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 1272w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!hJ_A!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png" width="1000" height="373" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:373,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:184395,&quot;alt&quot;:&quot;Figure 3. PLSpredict MV summary&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. PLSpredict MV summary" title="Figure 3. PLSpredict MV summary" srcset="https://substackcdn.com/image/fetch/$s_!hJ_A!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 424w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 848w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 1272w, https://substackcdn.com/image/fetch/$s_!hJ_A!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2d9e1cc9-ccf3-4658-a939-a7253bc1427d_1000x373.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. PLSpredict results for the 12 indicators: (1) the item Overview, (2) Q&#178;predict, (3) PLS-SEM RMSE, (4) LM RMSE.</figcaption></figure></div><p>Shmueli et al. (2019) propose reading the results in two steps:</p><ol><li><p><strong>Check Q&#178;predict.</strong> If the Q&#178;predict of an indicator is above 0, PLS-SEM predicts better than IA. If Q&#178;predict is 0 or below, the model has no predictive power for that indicator, and the assessment stops there.</p></li><li><p><strong>Compare the RMSE of PLS-SEM with LM</strong> for the indicators of the key endogenous construct (here all 12 indicators are considered):</p><ul><li><p>All indicators have a lower PLS-SEM RMSE than LM: <strong>high</strong> predictive power.</p></li><li><p>The majority are lower: <strong>medium</strong> predictive power.</p></li><li><p>A minority are lower: <strong>low</strong> predictive power.</p></li><li><p>None are lower: the model <strong>lacks</strong> predictive power.</p></li></ul></li></ol><p><strong>Step 1 for the Marketplace X model:</strong> the Q&#178;predict of the 12 indicators ranges from 0.085 (RI3) to 0.213 (SAT3), <strong>all above 0</strong> (Figure 4).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!A141!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!A141!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!A141!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!A141!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!A141!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!A141!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png" width="1900" height="840" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:840,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:79401,&quot;alt&quot;:&quot;Figure 4. Q&#178;predict of the 12 indicators&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Q&#178;predict of the 12 indicators" title="Figure 4. Q&#178;predict of the 12 indicators" srcset="https://substackcdn.com/image/fetch/$s_!A141!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!A141!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!A141!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!A141!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7ec52f1f-963c-4a84-bdec-8e62c00003bc_1900x840.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Q&#178;predict of the 12 indicators of the three endogenous constructs, values from SmartPLS (Figure 3).</figcaption></figure></div><p><strong>Step 2:</strong> the PLS-SEM RMSE is <strong>lower than LM for all 12 indicators</strong> (Figure 5). MAE is also lower than LM for all 12 indicators (columns PLS-SEM_MAE and LM_MAE in Figure 3), so RMSE and MAE lead to the same conclusion.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!BKAS!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!BKAS!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!BKAS!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png" width="1900" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:127162,&quot;alt&quot;:&quot;Figure 5. RMSE of PLS-SEM against LM&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. RMSE of PLS-SEM against LM" title="Figure 5. RMSE of PLS-SEM against LM" srcset="https://substackcdn.com/image/fetch/$s_!BKAS!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!BKAS!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4b890d64-23db-4aec-9084-4bcfed275262_1900x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. RMSE of PLS-SEM (green) and LM (orange) for each indicator. The longer the connecting segment, the larger the advantage of PLS-SEM over LM.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Kkfpj/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/451c0e04-c391-47ee-ad1d-951d419f8f41_1220x1100.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bf8e8bfc-e97b-4e48-b51a-32bad393ba3d_1220x1170.png&quot;,&quot;height&quot;:583,&quot;title&quot;:&quot;Table 1. PLSpredict results for the 12 indicators&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Kkfpj/1/" width="730" height="583" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Conclusion following Shmueli et al. (2019):</strong> the Marketplace X model has <strong>high</strong> out-of-sample predictive power. Two further points should be read alongside this:</p><ul><li><p><strong>The RMSE differences are very small</strong>, from 0.005 (RI1) to 0.022 (SAT3). PLS-SEM beats LM for every indicator, but not by much.</p></li><li><p><strong>Q&#178;predict is low to moderate.</strong> The Q&#178;predict of 0.085 for RI3 means that, compared with predicting by the mean, the squared error drops by only about 8.5%. Repurchase intention remains hard to predict from CONFIRM and TRUST, in line with the weak to moderate R&#178; of Post 37.</p></li></ul><h3>Q&#178;predict of the latent variables</h3><p><strong>Step 6.</strong> Open <strong>PLSpredict LV summary &gt; PLS-SEM</strong> (Figure 6).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!eii6!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!eii6!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 424w, https://substackcdn.com/image/fetch/$s_!eii6!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 848w, https://substackcdn.com/image/fetch/$s_!eii6!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 1272w, https://substackcdn.com/image/fetch/$s_!eii6!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!eii6!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png" width="1000" height="338" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:338,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:93876,&quot;alt&quot;:&quot;Figure 6. PLSpredict LV summary&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. PLSpredict LV summary" title="Figure 6. PLSpredict LV summary" srcset="https://substackcdn.com/image/fetch/$s_!eii6!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 424w, https://substackcdn.com/image/fetch/$s_!eii6!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 848w, https://substackcdn.com/image/fetch/$s_!eii6!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 1272w, https://substackcdn.com/image/fetch/$s_!eii6!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2b59ce16-58e1-4fa5-bade-5b859f882c44_1000x338.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. Q&#178;predict, RMSE and MAE of the three endogenous latent variables: (1) the item PLS-SEM, (2) the results table.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/jZtFm/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d6803200-2c59-4508-8241-1e55d4236f07_1220x434.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/46061758-f633-42a3-9380-d4be4e65db10_1220x554.png&quot;,&quot;height&quot;:253,&quot;title&quot;:&quot;Table 2. Q&#178;predict of the three endogenous latent variables compared with R&#178;&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/jZtFm/1/" width="730" height="253" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The Q&#178;predict of the latent variables is <strong>lower than R&#178;</strong>, most clearly for SATISF (0.280 versus 0.399) and INTENT (0.217 versus 0.318). There are two reasons: Q&#178;predict is computed on respondents outside the estimation sample, and with Early antecedents each construct is predicted only from the exogenous constructs (CONFIRM; plus TRUST for INTENT), not from the actual scores of VALUE and SATISF. This drop is normal, not an error.</p><p>Shmueli et al. (2019) recommend assessing predictive power on the <strong>indicators</strong> (the MV table), because users of the model predict actual answers, not latent variable scores. Use the LV table for reference.</p><h2>5. The CVPAT test</h2><p>PLSpredict tells you whether PLS-SEM has a lower error, but not whether the difference is <strong>statistically significant</strong> or just due to chance. CVPAT (cross-validated predictive ability test) answers this question (Liengaard et al., 2021). For each endogenous construct, CVPAT computes the <strong>average loss</strong> (the average squared prediction error) of PLS-SEM and of the benchmark, then runs a <em>t</em> test on the difference. <strong>A negative and significant difference</strong> means that PLS-SEM predicts better than the benchmark (Sharma et al., 2023).</p><h3>Against the mean (IA)</h3><p><strong>Step 7.</strong> Open <strong>CVPAT LV summary &gt; PLS-SEM vs. Indicator average (IA)</strong> (Figure 7).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!aLYV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!aLYV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 424w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 848w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 1272w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!aLYV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png" width="1000" height="446" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:446,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:139563,&quot;alt&quot;:&quot;Figure 7. CVPAT against IA&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. CVPAT against IA" title="Figure 7. CVPAT against IA" srcset="https://substackcdn.com/image/fetch/$s_!aLYV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 424w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 848w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 1272w, https://substackcdn.com/image/fetch/$s_!aLYV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F584c3c42-72be-4def-9475-8390f3c691bd_1000x446.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. CVPAT against IA: (1) the item PLS-SEM vs. Indicator average (IA), (2) p values of the three constructs, (3) the Overall row for the whole model.</figcaption></figure></div><p>The loss difference is negative for all three constructs (&#8722;0.116 to &#8722;0.166), with <em>p</em> from 0.000 to 0.001. For the whole model: PLS-SEM loss 0.762, IA loss 0.907, difference &#8722;0.145, <em>t</em> = 4.688, <em>p</em> &lt; 0.001. <strong>PLS-SEM predicts significantly better than the mean</strong>, consistent with Q&#178;predict above 0 in Section 4.</p><h3>Against the linear model (LM)</h3><p><strong>Step 8.</strong> Open <strong>CVPAT LV summary &gt; PLS-SEM vs. Linear model (LM)</strong> (Figure 8).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!zNTM!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!zNTM!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 424w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 848w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 1272w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!zNTM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png" width="1000" height="446" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:446,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:139874,&quot;alt&quot;:&quot;Figure 8. CVPAT against LM&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. CVPAT against LM" title="Figure 8. CVPAT against LM" srcset="https://substackcdn.com/image/fetch/$s_!zNTM!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 424w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 848w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 1272w, https://substackcdn.com/image/fetch/$s_!zNTM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69dbb252-4184-45e1-ba3d-bc478697f3c2_1000x446.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. CVPAT against LM: (1) the item PLS-SEM vs. Linear model (LM), (2) the INTENT row, p value 0.137, so SmartPLS shows it in red, (3) the Overall row.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/irYKK/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fd164710-1af4-4dc0-b448-04856603a05c_1220x508.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2855a90a-bcf6-499f-b702-5954e98f36cb_1220x578.png&quot;,&quot;height&quot;:283,&quot;title&quot;:&quot;Table 3. CVPAT of PLS-SEM against the linear model (LM)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/irYKK/1/" width="730" height="283" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Reading the results:</strong></p><ul><li><p><strong>Whole model:</strong> PLS-SEM predicts significantly better than LM (<em>t</em> = 4.193, <em>p</em> &lt; 0.001).</p></li><li><p><strong>VALUE and SATISF:</strong> PLS-SEM is significantly better than LM (<em>p</em> = 0.002 and <em>p</em> &lt; 0.001).</p></li><li><p><strong>Repurchase intention (INTENT):</strong> the PLS-SEM loss is lower than the LM loss (0.774 versus 0.792), but the difference is <strong>not statistically significant</strong> (<em>t</em> = 1.491, <em>p</em> = 0.137). This <em>p</em> value is two-tailed; the one-tailed <em>p</em> would be about 0.069, still above 0.05.</p></li></ul><p>This result <strong>complements, rather than contradicts,</strong> PLSpredict. PLSpredict shows a lower PLS-SEM RMSE than LM for all four indicators RI1 to RI4; CVPAT shows that, for INTENT, this advantage is so small that chance cannot be ruled out. Report it exactly that way: for the most important construct of the model, the theoretical structure <strong>does not yet predict clearly better</strong> than a direct linear regression on the indicators.</p><h2>6. Checking in R</h2><p>The script <code>plspredict.R</code> runs the same model with the <code>predict()</code> function of the cSEM package: 10 folds, 10 repetitions, LM as the benchmark, regular PLS-SEM (no PLSc correction), 254 respondents. Because the random fold assignment differs from SmartPLS, the results agree only approximately:</p><ul><li><p><strong>Q&#178;predict</strong> agrees to the second decimal place, for example PV1 0.177 (R) versus 0.178 (SmartPLS), RI3 0.085 versus 0.085, SAT3 0.211 versus 0.213.</p></li><li><p><strong>The PLS-SEM RMSE</strong> is almost identical, for example PV1 0.929 versus 0.929.</p></li><li><p><strong>The LM RMSE</strong> in cSEM is larger than in SmartPLS (for example PV1 0.986 versus 0.948), so the gap between PLS-SEM and LM is larger in cSEM. We have not identified which setting of the two programs causes this difference. Still, the conclusion is the same: the PLS-SEM RMSE is lower than LM for all 12 indicators.</p></li></ul><p>The R values above come from the run saved in <code>plspredict.csv</code>. In cSEM 0.7.1, <code>set.seed()</code> does not fix the folds of <code>predict()</code> with more than one repetition, so rerunning the script gives values that differ in the third decimal place; the conclusions do not change.</p><p>This post uses the SmartPLS numbers. If you use other software, report the software, the number of folds, the number of repetitions and how random numbers were generated.</p><h2>7. Reporting the results</h2><blockquote><p>Out-of-sample predictive power was assessed with PLSpredict (Shmueli et al., 2019) using 10 folds and 10 repetitions. The Q&#178;predict values of the 12 indicators of the three endogenous constructs were all above 0 (0.085 to 0.213), and the PLS-SEM RMSE was lower than that of the linear model for all 12 indicators, indicating high predictive power according to the criteria of Shmueli et al. (2019). However, the RMSE differences were small (0.005 to 0.022). The CVPAT test (Sharma et al., 2023) showed that PLS-SEM predicted better than the indicator average (loss difference &#8722;0.145, <em>t</em> = 4.688, <em>p</em> &lt; 0.001) and better than the linear model (&#8722;0.026, <em>t</em> = 4.193, <em>p</em> &lt; 0.001) at the overall model level. For repurchase intention, however, PLS-SEM did not predict significantly better than the linear model (<em>t</em> = 1.491, <em>p</em> = 0.137).</p></blockquote><p>In a thesis table, present Q&#178;predict and the RMSE (or MAE) of PLS-SEM and of LM for each indicator, as in the table in Section 4, together with the CVPAT table, and note the number of folds, the number of repetitions and the prediction type.</p><h2>Five common mistakes</h2><ol><li><p><strong>Reporting the blindfolding Q&#178; as evidence of out-of-sample prediction.</strong> This Q&#178; is closer to in-sample explanatory power (Shmueli et al., 2016).</p></li><li><p><strong>Reporting only that Q&#178;predict is above 0 and concluding "good prediction".</strong> Q&#178;predict above 0 is only the minimum requirement; you still need the comparison with LM.</p></li><li><p><strong>Comparing the PLS-SEM RMSE with the IA RMSE.</strong> The benchmark in step 2 is LM. IA is already used in Q&#178;predict.</p></li><li><p><strong>Ignoring a non-significant CVPAT result.</strong> The INTENT result in this post must be reported and discussed, not just the Overall row.</p></li><li><p><strong>Not reporting the settings.</strong> The results vary with the fold assignment; without the number of folds, the number of repetitions and the seed, readers cannot check the results.</p></li></ol><h2>Frequently asked questions</h2><p><strong>What Q&#178;predict counts as high?</strong></p><p>Shmueli et al. (2019) give no threshold for Q&#178;predict other than being above 0; the main assessment rests on the comparison with LM. To describe its size, you can use the rule of thumb in Hair et al. (2019): Q&#178; values above 0, 0.25 and 0.50 indicate small, medium and large predictive relevance. By this rule, the Q&#178;predict values of the indicators in the Marketplace X model are small.</p><p><strong>Why do only the indicators of VALUE, SATISF and INTENT appear in the results?</strong></p><p>PLSpredict predicts the endogenous constructs from the exogenous constructs. CONFIRM and TRUST are exogenous, so their indicators are inputs, not prediction targets.</p><p><strong>My results differ by a few thousandths from my previous run. Why?</strong></p><p>Check the Random number generator setting. With Random seed, each run splits the folds differently, so the numbers vary slightly. Choose Fixed seed to get the same results when you rerun.</p><p><strong>Does a non-significant CVPAT result mean the model is rejected?</strong></p><p>No. CVPAT assesses predictive power; it does not test causal hypotheses. Hypotheses H1 to H7 are tested with bootstrapping in <a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a>. Put the CVPAT result in the discussion and limitations: the model explains the relationships between the constructs, but its ability to predict repurchase intention does not yet clearly exceed that of a simple regression model.</p><div><hr></div><h2>References</h2><p>Geisser, S. (1974). A predictive approach to the random effect model. <em>Biometrika, 61</em>(1), 101&#8211;107. https://doi.org/10.1093/biomet/61.1.101</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>Liengaard, B. D., Sharma, P. N., Hult, G. T. M., Jensen, M. B., Sarstedt, M., Hair, J. F., &amp; Ringle, C. M. (2021). Prediction: Coveted, yet forsaken? Introducing a cross-validated predictive ability test in partial least squares path modeling. <em>Decision Sciences, 52</em>(2), 362&#8211;392. https://doi.org/10.1111/deci.12445</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>Sharma, P. N., Liengaard, B. D., Hair, J. F., Sarstedt, M., &amp; Ringle, C. M. (2023). Predictive model assessment and selection in composite-based modeling using PLS-SEM: Extensions and guidelines for using CVPAT. <em>European Journal of Marketing, 57</em>(6), 1662&#8211;1677. https://doi.org/10.1108/EJM-08-2020-0636</p><p>Shmueli, G., Ray, S., Velasquez Estrada, J. M., &amp; Chatla, S. B. (2016). The elephant in the room: Predictive performance of PLS models. <em>Journal of Business Research, 69</em>(10), 4552&#8211;4564. https://doi.org/10.1016/j.jbusres.2016.03.049</p><p>Shmueli, G., Sarstedt, M., Hair, J. F., Cheah, J.-H., Ting, H., Vaithilingam, S., &amp; Ringle, C. M. (2019). Predictive model assessment in PLS-SEM: Guidelines for using PLSpredict. <em>European Journal of Marketing, 53</em>(11), 2322&#8211;2347. https://doi.org/10.1108/EJM-02-2019-0189</p><p>Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions. <em>Journal of the Royal Statistical Society: Series B (Methodological), 36</em>(2), 111&#8211;147. https://doi.org/10.1111/j.2517-6161.1974.tb00994.x</p>]]></content:encoded></item><item><title><![CDATA[Reflective Measurement Model in SmartPLS 4: Loadings, Reliability and AVE]]></title><description><![CDATA[Check outer loadings, Cronbach's alpha, rho_a, rho_c and AVE step by step, and see why PLS-SEM reports a higher AVE than CB-SEM]]></description><link>https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:49:00 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!aQDP!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!aQDP!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!aQDP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:776592,&quot;alt&quot;:&quot;A large orange bell ringing above four glass bells that echo its vibration in concentric rings, a small figure listening with an ear trumpet.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A large orange bell ringing above four glass bells that echo its vibration in concentric rings, a small figure listening with an ear trumpet." title="A large orange bell ringing above four glass bells that echo its vibration in concentric rings, a small figure listening with an ear trumpet." srcset="https://substackcdn.com/image/fetch/$s_!aQDP!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!aQDP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f1c1084-499c-4b19-bdb6-a96b5bf4eac6_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>A reflective measurement model is evaluated in order: outer loadings, internal consistency reliability, convergent validity (AVE), then discriminant validity.</p></li><li><p>Loadings should reach 0.708; an indicator below that is not removed automatically, and only one below 0.40 is always removed.</p></li><li><p>Report Cronbach's alpha, rho_a and rho_c together: alpha tends to be low, rho_c tends to be high and rho_a lies in between; values above 0.95 signal redundant indicators.</p></li><li><p>On the Marketplace X data all 20 loadings range from 0.754 to 0.876 and AVE from 0.612 to 0.682, so every criterion is met.</p></li><li><p>PLS-SEM inflates loadings: the AVE of Repurchase intention is 0.612 here, 0.482 with consistent PLS and 0.477 in CB-SEM. Choose the method for its rationale, never for its results.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a> drew the Marketplace X model and ran the PLS-SEM algorithm. Before reading the path coefficients, you need to make sure that <strong>the scales measure their constructs correctly and consistently</strong>. For a <strong>reflective</strong> measurement model, this takes three steps: check the outer loadings, the internal consistency reliability and the convergent validity (Hair et al., 2019). The fourth step, discriminant validity, is the topic of <a href="https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls">Post 35</a>.</p><p>The results on the Marketplace X data raise one point worth discussing: <strong>the AVE of the Repurchase intention scale is 0.612</strong>, above the threshold, whereas <a href="https://www.phuonglib.com/p/composite-reliability-ave-cfa-jasp">Post 24</a> (CB-SEM) gave 0.477, below it. Section 5 explains why and how to report it.</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch34-evaluate-reflective-measurement-model">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of the SmartPLS 4.1.2.1 report (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code> from Post 33, run with casewise deletion on 254 respondents, taken in November 2026. Every number was checked with the cSEM 0.7.1 package (Rademaker &amp; Schuberth, 2026), script <code>evaluate_measurement.R</code> in the repository's <code>scripts/</code> folder, and matches SmartPLS to the third decimal place.</p></blockquote><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!l8Jr!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!l8Jr!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 424w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 848w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 1272w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!l8Jr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png" width="1675" height="540" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:540,&quot;width&quot;:1675,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:101353,&quot;alt&quot;:&quot;Figure 1. Four steps to evaluate a reflective measurement model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Four steps to evaluate a reflective measurement model" title="Figure 1. Four steps to evaluate a reflective measurement model" srcset="https://substackcdn.com/image/fetch/$s_!l8Jr!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 424w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 848w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 1272w, https://substackcdn.com/image/fetch/$s_!l8Jr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc46db323-04ce-43ed-91d4-6c74c4fc9483_1675x540.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Four steps to evaluate a reflective measurement model in PLS-SEM, following Hair et al. (2019). This post covers the first three.</figcaption></figure></div><h2>1. Open the results in SmartPLS</h2><p>Open the report of the PLS-SEM algorithm (Post 33, Section 6). The two tables you need are in the <strong>Report</strong> pane on the left:</p><p><strong>Step 1.</strong> In the <strong>Final results</strong> group, click the arrow next to <strong>Outer loadings</strong>, then choose <strong>Matrix</strong> (Figure 2).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!1J4m!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!1J4m!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!1J4m!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/eb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:137337,&quot;alt&quot;:&quot;Figure 2. The Outer loadings table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. The Outer loadings table" title="Figure 2. The Outer loadings table" srcset="https://substackcdn.com/image/fetch/$s_!1J4m!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!1J4m!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Feb33b80d-9c86-426f-b146-bc2bdd7b5833_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. Outer loadings: (1) the Outer loadings &gt; Matrix item, (2) the table of loadings, one column per latent variable, (3) the two buttons that copy the table to Excel, Word or R.</figcaption></figure></div><p><strong>Step 2.</strong> In the <strong>Quality criteria</strong> group, click the arrow next to <strong>Construct reliability and validity</strong>, then choose <strong>Overview</strong> (Figure 3).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!cfKH!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!cfKH!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!cfKH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:144647,&quot;alt&quot;:&quot;Figure 3. The Construct reliability and validity table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. The Construct reliability and validity table" title="Figure 3. The Construct reliability and validity table" srcset="https://substackcdn.com/image/fetch/$s_!cfKH!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!cfKH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a86828f-1b2b-48f2-bff9-ebc25ac93a30_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. Reliability and convergent validity: (1) the Overview item, (2) the table with four criteria for the five scales, (3) the AVE of Repurchase intention (INTENT), (4) the bar charts of each criterion.</figcaption></figure></div><p>Three things to note when reading the report:</p><ul><li><p><strong>The line Calculated for Complete data set</strong> at the top of each table tells you that the results are based on the complete responses (casewise deletion, Post 33).</p></li><li><p><strong>SmartPLS sorts constructs and indicators alphabetically</strong> in these tables (CONFIRM, INTENT, SATISF, TRUST, VALUE), not in the order in which you drew them. The tables in this post follow the same order.</p></li><li><p><strong>The Copy to Excel/Word and Copy to R buttons</strong> copy the whole table so you can paste it into your thesis without retyping. According to the license comparison, the Student license cannot copy or export reports (SmartPLS GmbH, n.d.).</p></li></ul><h2>2. Step 1: Outer loadings</h2><p>The <strong>outer loading</strong> is the correlation between an indicator and its latent variable. The squared loading is the <strong>indicator reliability</strong>: the share of the indicator's variance explained by the construct.</p><p><strong>Threshold:</strong> loadings of <strong>0.708</strong> or higher, because the squared loading is then 0.5: the construct explains at least half of the indicator's variance (Hair et al., 2019). An indicator with a loading below 0.708 is not removed automatically (Hair et al., 2022):</p><ul><li><p><strong>Below 0.40:</strong> remove the indicator.</p></li><li><p><strong>From 0.40 to below 0.708:</strong> remove it only if removal brings rho_c or AVE above the threshold, and does not remove an important aspect of the construct's content.</p></li></ul><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/ZeqLy/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4b3bf38c-a8cd-4721-bc07-77c2ce85b566_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/40c26483-d0b2-4b68-ba00-7b8bf08414ad_1220x704.png&quot;,&quot;height&quot;:347,&quot;title&quot;:&quot;Table 1. Outer loadings and indicator reliability of the five reflective scales&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/ZeqLy/1/" width="730" height="347" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> all 20 indicators have loadings from 0.754 (CONF3) to 0.876 (PV3), <strong>all above 0.708</strong>. No indicator needs to be considered for removal. TRU4R (the reverse-worded item recoded in <a href="https://www.phuonglib.com/p/recode-compute-composite-scores-pspp-spss">Post 10</a>) has a loading of 0.817: the recoding went in the right direction; had it not been recoded, its loading would be negative.</p><p>Testing the significance of the loadings requires bootstrapping, covered in <a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a>.</p><h2>3. Step 2: Internal consistency reliability</h2><p>SmartPLS reports three reliability criteria (Figure 3):</p><ul><li><p><strong>Cronbach's alpha</strong> (Cronbach, 1951): assumes that every indicator contributes equally (<a href="https://www.phuonglib.com/p/cronbach-alpha-spss-pspp">Post 11</a>). When loadings differ, alpha is usually <strong>lower</strong> than the true reliability.</p></li><li><p><strong>Composite reliability (rho_c)</strong> (J&#246;reskog, 1971): uses the estimated loadings, so each indicator can contribute differently. This is the CR from Post 24, computed with the same formula. In PLS-SEM, rho_c is usually <strong>higher</strong> than the true reliability, because the loadings are inflated (<a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31</a>).</p></li><li><p><strong>Composite reliability (rho_a)</strong> (Dijkstra &amp; Henseler, 2015): computed from the weights of the composite. Its value usually lies <strong>between</strong> alpha and rho_c, so it is regarded as a good compromise and the most reliable measure for PLS-SEM (Hair et al., 2019).</p></li></ul><p><strong>Thresholds</strong> (Hair et al., 2019):</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/UPGUT/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/729093b8-cb27-4447-919b-bcb8b2738309_1220x536.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dcfc7363-5345-493d-b0bd-09065e4af886_1220x656.png&quot;,&quot;height&quot;:281,&quot;title&quot;:&quot;Table 2. Thresholds for internal consistency reliability (Hair et al., 2019)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/UPGUT/1/" width="730" height="281" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/oN8E7/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7e74f608-ce67-40b5-8e64-143597a4ede4_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6adc6fcd-30b5-47a9-bf54-e5d68d6423c6_1220x616.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 3. Internal consistency reliability of the five scales&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/oN8E7/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> all three criteria of the five scales lie between 0.70 and 0.90: <strong>satisfactory</strong>. The order alpha below rho_a, rho_a below rho_c is as theory predicts. Cronbach's alpha here is close to Post 11 (for example CONFIRM 0.825 versus 0.824); the small difference arises because Post 11 used a different sample.</p><h3>Computing rho_c by hand for the Repurchase intention scale</h3><p>The formula is the same as in Post 24, with L the loading:</p><ul><li><p><strong>rho_c</strong> = (Sum of L)&#178; / [(Sum of L)&#178; + Sum of (1 - L&#178;)]</p></li><li><p><strong>Add</strong> the four loadings: 0.805 + 0.783 + 0.761 + 0.778 = 3.128 (from the unrounded loadings, 3.1277; the rounded values add to 3.127); squaring the unrounded sum 3.1277 gives 9.782 (squaring 3.128 itself would give 9.784).</p></li><li><p><strong>Add</strong> the four error terms (1 - L&#178;): 4 - (0.649 + 0.613 + 0.579 + 0.605) = 4 - 2.446 = 1.554.</p></li><li><p><strong>Compute</strong> rho_c = 9.782 / (9.782 + 1.554) = <strong>0.863</strong>, exactly as in SmartPLS.</p></li></ul><h2>4. Step 3: Convergent validity (AVE)</h2><p><strong>AVE</strong> (average variance extracted) is the mean of the squared loadings of the indicators in a scale (Fornell &amp; Larcker, 1981). AVE tells you how much of the indicators' variance the construct explains on average.</p><p><strong>Threshold:</strong> AVE of <strong>0.50</strong> or higher (Fornell &amp; Larcker, 1981; Hair et al., 2019).</p><p>By hand for the Repurchase intention scale: AVE = (0.649 + 0.613 + 0.579 + 0.605) / 4 = 2.446 / 4 = <strong>0.612</strong>.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/wXa8u/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/eb5c79f1-a4ab-4503-bda7-16579971c6d5_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/25c81553-ca9a-4010-9962-578bad1e5a4c_1220x616.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 4. Average variance extracted (AVE) of the five scales&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/wXa8u/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> the five scales have AVE from 0.612 to 0.682: <strong>convergent validity is established</strong>.</p><h2>5. Why does the AVE of Repurchase intention differ from Post 24?</h2><p>In Post 24, the CB-SEM CFA gave the Repurchase intention scale an AVE of 0.477, below the threshold. Here PLS-SEM gives 0.612. Same data, same four indicators, yet different conclusions. The reason lies in how the construct is understood (Post 31):</p><ul><li><p><strong>PLS-SEM</strong> treats the construct as a <strong>composite</strong> of its own indicators, which includes measurement error. Each indicator is part of the composite, so it correlates more strongly with it: <strong>the loadings are inflated</strong>, and so are AVE and rho_c.</p></li><li><p><strong>Consistent PLS (PLSc)</strong> corrects this bias. With PLSc, the AVE of Repurchase intention is <strong>0.482</strong>, close to the 0.477 from CB-SEM (Figure 4).</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!B8mf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!B8mf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!B8mf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png" width="1900" height="840" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c1002789-891e-47e1-9900-41db6c9a4396_1900x840.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:840,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:93077,&quot;alt&quot;:&quot;Figure 4. AVE by three estimation methods&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. AVE by three estimation methods" title="Figure 4. AVE by three estimation methods" srcset="https://substackcdn.com/image/fetch/$s_!B8mf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!B8mf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc1002789-891e-47e1-9900-41db6c9a4396_1900x840.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. AVE of the five scales by PLS-SEM, PLSc-SEM (cSEM, 254 respondents) and CB-SEM (Post 24, FIML, 267 respondents). The dashed line is the 0.50 threshold.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/TrnAg/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/59e1c5fd-f4aa-4788-a4f9-e1c930298670_1220x582.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/36034700-e837-4025-8942-5319ff5e263c_1220x652.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 5. AVE by three estimation methods&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/TrnAg/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/bJalH/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c841ea12-339b-42f2-950a-e27a620be00a_1220x582.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f69b8274-a041-48b7-8289-2008c0834077_1220x768.png&quot;,&quot;height&quot;:339,&quot;title&quot;:&quot;Table 6. Composite reliability (rho_c) by three estimation methods&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/bJalH/1/" width="730" height="339" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>How should you report this?</p><ul><li><p><strong>If your thesis uses PLS-SEM:</strong> report the PLS-SEM results as SmartPLS gives them, because the thresholds in Sections 2 to 4 were proposed for these very criteria (Hair et al., 2019). The Repurchase intention scale meets every criterion.</p></li><li><p><strong>Do not choose the method by its results.</strong> Switching from CB-SEM to PLS-SEM only to push AVE above 0.5 is not honest. The choice of PLS-SEM must rest on how the constructs are understood and on the research goal (Post 31).</p></li><li><p><strong>If the constructs are understood as common factors</strong> (like the five constructs of the Marketplace X study) and you want less biased estimates, use PLSc, or state in the limitations that the AVE of Repurchase intention is close to the threshold once measurement error is corrected (0.482). This is the same honest reporting as in Post 24.</p></li></ul><h2>6. Reporting the results</h2><blockquote><p>The measurement model was evaluated following Hair et al. (2019). The loadings of the 20 indicators ranged from 0.754 to 0.876, all above the 0.708 threshold. The internal consistency reliability of the five scales was satisfactory: Cronbach's alpha ranged from 0.789 to 0.844, rho_a from 0.791 to 0.846 and rho_c from 0.863 to 0.896, all between 0.70 and 0.90. The average variance extracted (AVE) ranged from 0.612 to 0.682, above the 0.50 threshold, indicating convergent validity. When measurement error was corrected with consistent PLS, the AVE of the Repurchase intention scale was 0.482, close to the threshold; this point is noted in the limitations.</p></blockquote><p>The table in your thesis should combine the loadings, alpha, rho_a, rho_c and AVE of each scale. Use the <strong>Copy to Excel/Word</strong> button to copy the tables from SmartPLS.</p><h2>Five common mistakes</h2><ol><li><p><strong>Removing an indicator only because its loading is below 0.708</strong>, without checking whether removal brings rho_c or AVE above the threshold, and without considering its content.</p></li><li><p><strong>Reporting only Cronbach's alpha.</strong> In PLS-SEM, report rho_a and rho_c as well.</p></li><li><p><strong>Assuming that a higher rho_c is always better.</strong> Above 0.95 signals redundant indicators or straight-lining.</p></li><li><p><strong>Comparing PLS-SEM AVE directly with studies that used CB-SEM</strong> without noting the bias of PLS-SEM (Section 5).</p></li><li><p><strong>Reading the results of an old run</strong> after changing the model. Every time you change the model, run the algorithm again before reading the tables.</p></li></ol><h2>Frequently asked questions</h2><p><strong>CONF3 has the lowest loading (0.754). Should I remove it to raise AVE?</strong></p><p>No. CONF3 is above the 0.708 threshold, and the AVE of CONFIRM (0.655) is already met. Removing an indicator that meets the criteria only loses content from the scale.</p><p><strong>How do rho_a and rho_c differ, and which should I report?</strong></p><p>Report both, together with Cronbach's alpha, as in the table in Section 3. Alpha tends to be low and rho_c tends to be high in PLS-SEM, with rho_a in between (Hair et al., 2019). If all three meet the threshold, as in the Marketplace X data, the conclusion is robust.</p><p><strong>Do I need to test the significance of the loadings?</strong></p><p>Yes, with bootstrapping (Post 39). With loadings of 0.75 or higher on 254 respondents, they are almost certainly significant, but you should still report the test.</p><p><strong>Are formative scales evaluated this way?</strong></p><p>No. For formative scales, loadings, alpha and AVE do not have the meaning described above. Their evaluation (VIF, weights) is covered in <a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">Post 36</a>.</p><div><hr></div><h2>References</h2><p>Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. <em>Psychometrika, 16</em>(3), 297&#8211;334. https://doi.org/10.1007/BF02310555</p><p>Dijkstra, T. K., &amp; Henseler, J. (2015). Consistent partial least squares path modeling. <em>MIS Quarterly, 39</em>(2), 297&#8211;316. https://doi.org/10.25300/MISQ/2015/39.2.02</p><p>Fornell, C., &amp; Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. <em>Journal of Marketing Research, 18</em>(1), 39&#8211;50. https://doi.org/10.2307/3151312</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>J&#246;reskog, K. G. (1971). Statistical analysis of sets of congeneric tests. <em>Psychometrika, 36</em>(2), 109&#8211;133. https://doi.org/10.1007/BF02291393</p><p>Rademaker, M. E., &amp; Schuberth, F. (2026). <em>cSEM: Composite-based structural equation modeling</em> (R package version 0.7.1) [Computer software]. https://CRAN.R-project.org/package=cSEM</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>SmartPLS GmbH. (n.d.). <em>Pricing and licenses</em>. Retrieved November 6, 2026, from https://smartpls.com/pricing/</p>]]></content:encoded></item><item><title><![CDATA[Evaluating the Structural Model in SmartPLS 4: Collinearity, R² and f²]]></title><description><![CDATA[Check inner VIF, read R&#178; and adjusted R&#178;, compute the f&#178; effect size by hand and report weak to moderate explanatory power honestly]]></description><link>https://www.phuonglib.com/p/evaluate-structural-model-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/evaluate-structural-model-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:56 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!2thC!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2thC!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!2thC!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!2thC!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!2thC!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!2thC!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:776748,&quot;alt&quot;:&quot;Four streams of different widths feeding a stone reservoir filled about one third, an orange stream the widest, a figure reading a level gauge.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Four streams of different widths feeding a stone reservoir filled about one third, an orange stream the widest, a figure reading a level gauge." title="Four streams of different widths feeding a stone reservoir filled about one third, an orange stream the widest, a figure reading a level gauge." srcset="https://substackcdn.com/image/fetch/$s_!2thC!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!2thC!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!2thC!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!2thC!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7a3e32a9-a228-437e-b79c-0d225fb6a133_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>A PLS-SEM structural model is evaluated in five steps: collinearity, path significance, R&#178;, f&#178; and predictive power; this post covers collinearity, R&#178; and f&#178;.</p></li><li><p>Collinearity between latent variables is read from Inner model VIF, not Outer model VIF; on the Marketplace X model the highest is 1.681, below 3.</p></li><li><p>The model explains 25.1% of perceived value, 39.9% of satisfaction and 31.8% of repurchase intention: weak to moderate, and it should be reported that way.</p></li><li><p>f&#178; is the unique contribution of one predictor; trust contributes most to repurchase intention (0.102), while confirmation has almost no direct effect (0.009).</p></li><li><p>PLS-SEM gives lower R&#178; than CB-SEM on the same data (0.318 versus 0.444 for intention), so check the method before comparing with earlier studies.</p></li></ul></blockquote><div><hr></div><p>Posts <a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">34</a> and <a href="https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls">35</a> showed that the measurement scales of the Marketplace X model meet the requirements. From this post on, the series turns to the <strong>structural model</strong>: the relationships between the five constructs and hypotheses H1 to H7. Hair et al. (2019) recommend evaluating the structural model in five steps:</p><ol><li><p><strong>Collinearity</strong> between the latent variables (this post).</p></li><li><p><strong>Significance and size</strong> of the path coefficients (<a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a>, with bootstrapping).</p></li><li><p><strong>Explanatory power</strong>: R&#178; (this post).</p></li><li><p><strong>Effect size</strong> of each predictor: f&#178; (this post).</p></li><li><p><strong>Predictive power</strong>: Q&#178; and PLSpredict (<a href="https://www.phuonglib.com/p/plspredict-cvpat-smartpls">Post 38</a>).</p></li></ol><p>The posts in this series follow the editorial schedule; in a thesis, present the results in the order of the five steps above.</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch37-evaluate-structural-model">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The figures are real screenshots of the SmartPLS 4.1.2.1 report (Ringle et al., 2026; 30-day trial) for the model <code>marketplace-x-model</code> from <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>, running the PLS-SEM algorithm again with casewise deletion on 254 respondents. Every number was checked in R with the script <code>structural_model.R</code> (in the repository's <code>scripts/</code> folder) and matches SmartPLS to the third decimal place. SmartPLS lists constructs and paths in alphabetical order (CONFIRM, INTENT, SATISF, TRUST, VALUE).</p></blockquote><h2>1. Step 1: Collinearity between the latent variables</h2><p>The path coefficients of an endogenous construct are estimated like a multiple regression on the latent variables that predict it. If these predictors are too highly correlated with each other, the coefficients become biased and unstable, just as with multicollinearity in the regression of <a href="https://www.phuonglib.com/p/regression-assumptions-spss-pspp">Post 15</a>.</p><p><strong>Step 1.</strong> In the report of the PLS-SEM algorithm, open <strong>Quality criteria &gt; Collinearity statistics (VIF) &gt; Inner model - List</strong> (Figure 1). The item <strong>Outer model - List</strong> in the same group gives the VIF of the indicators, which is used for formative measurement (<a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">Post 36</a>).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ZUFO!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ZUFO!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ZUFO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png" width="1000" height="859" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:859,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:184310,&quot;alt&quot;:&quot;Figure 1. VIF of the structural model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. VIF of the structural model" title="Figure 1. VIF of the structural model" srcset="https://substackcdn.com/image/fetch/$s_!ZUFO!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!ZUFO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9eec7233-e824-49ee-b3bb-b0e82543fcf3_1000x859.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. VIF between the latent variables: (1) the item Inner model - List, (2) the VIF of the seven paths.</figcaption></figure></div><p><strong>Threshold</strong> (Hair et al., 2019): a VIF of 5 or more indicates a critical problem; ideally the VIF is below 3.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/BdvTb/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bce3a452-40db-44aa-bf9a-2c0039beb29c_1220x398.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e393914f-b96a-428c-ae30-58a9d29853f7_1220x468.png&quot;,&quot;height&quot;:227,&quot;title&quot;:&quot;Table 1. VIF between the latent variables&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/BdvTb/1/" width="730" height="227" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> the highest VIF is 1.681, <strong>below 3</strong>: there is no collinearity. VALUE has only one predictor, so its VIF equals 1. For SATISF, the two predictors have the same VIF, because the VIF of each one is computed from its correlation with the other.</p><h2>2. Step 3: Explanatory power (R&#178;)</h2><p><strong>R&#178;</strong> is the share of the variance of an endogenous construct that its predictors explain. <strong>Adjusted R&#178;</strong> removes the part of the increase that comes only from the number of predictors; use it when comparing models with different numbers of predictors:</p><ul><li><p>Adjusted R&#178; = 1 - (1 - R&#178;) x (n - 1) / (n - k - 1), where n is the sample size and k the number of predictors.</p></li></ul><p><strong>Step 2.</strong> Open <strong>Quality criteria &gt; R-square &gt; Overview</strong> (Figure 2).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ytYJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ytYJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ytYJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png" width="1000" height="859" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:859,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:164388,&quot;alt&quot;:&quot;Figure 2. R-square table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. R-square table" title="Figure 2. R-square table" srcset="https://substackcdn.com/image/fetch/$s_!ytYJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!ytYJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1e599c14-e580-4734-bf40-957e143fa0fd_1000x859.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. R&#178; and adjusted R&#178;: (1) the item R-square &gt; Overview, (2) the R&#178; table of the three endogenous constructs.</figcaption></figure></div><p><strong>Rules of thumb</strong> (Hair et al., 2019; Hair et al., 2022): an R&#178; of 0.75 is substantial, 0.50 moderate and 0.25 weak. These thresholds depend on the field: in consumer behavior research, lower R&#178; values can still be acceptable, because behavior is influenced by many factors outside the model.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/wcDDF/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c68bcbc1-db4c-4006-b758-a6807110947b_1220x498.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e510188d-809d-4949-8d8a-abdb578e9ad0_1220x618.png&quot;,&quot;height&quot;:278,&quot;title&quot;:&quot;Table 2. R&#178; and adjusted R&#178; of the three endogenous constructs&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/wcDDF/1/" width="730" height="278" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Hand calculation of the adjusted R&#178; of INTENT (n = 254, k = 4): 1 - (1 - 0.318) x 253 / 249 = 1 - 0.682 x 1.016 = <strong>0.307</strong>.</p><p><strong>Result:</strong> the four constructs of the model explain 31.8% of the variance of repurchase intention, and two constructs explain 39.9% of the variance of satisfaction. By the thresholds of Hair et al., explanatory power is <strong>between weak and moderate</strong>. This is an honest result of the data and should be reported as such, not described as "good explanatory power".</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!UcBf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!UcBf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!UcBf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png" width="1900" height="840" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:840,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:74141,&quot;alt&quot;:&quot;Figure 3. R&#178; from PLS-SEM and CB-SEM&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. R&#178; from PLS-SEM and CB-SEM" title="Figure 3. R&#178; from PLS-SEM and CB-SEM" srcset="https://substackcdn.com/image/fetch/$s_!UcBf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 424w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 848w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 1272w, https://substackcdn.com/image/fetch/$s_!UcBf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcfd0485-7d26-4a59-b05f-201a7da0ff64_1900x840.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. R&#178; and adjusted R&#178; from PLS-SEM compared with the R&#178; from CB-SEM (<a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31</a>) on the same 254 respondents. The shaded bands mark the thresholds 0.25, 0.50 and 0.75.</figcaption></figure></div><p>The PLS-SEM R&#178; is <strong>lower</strong> than the CB-SEM R&#178; for all three constructs (for example, INTENT: 0.318 versus 0.444; Post 31). This is a known bias of PLS-SEM when the constructs are common factors: the composites contain measurement error, which dilutes the relationships between constructs. When you compare R&#178; values with earlier studies, check which method those studies used.</p><h2>3. Step 4: Effect size (f&#178;)</h2><p><strong>f&#178;</strong> shows <strong>how much</strong> the R&#178; of an endogenous construct drops when one predictor is removed from the model, that is, the unique contribution of that predictor:</p><ul><li><p>f&#178; = (R&#178; included - R&#178; excluded) / (1 - R&#178; included)</p></li></ul><p><strong>Thresholds</strong> (Cohen, 1988): 0.02 is a small effect, 0.15 medium and 0.35 large. Values below 0.02 mean no effect.</p><p><strong>Step 3.</strong> Open <strong>Quality criteria &gt; f-square &gt; Matrix</strong> (Figure 4). Read it with <strong>predictors in the rows and endogenous constructs in the columns</strong>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!cJwD!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!cJwD!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!cJwD!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png" width="1000" height="859" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:859,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:168984,&quot;alt&quot;:&quot;Figure 4. f-square matrix&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. f-square matrix" title="Figure 4. f-square matrix" srcset="https://substackcdn.com/image/fetch/$s_!cJwD!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 424w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 848w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 1272w, https://substackcdn.com/image/fetch/$s_!cJwD!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F17b97363-5fc0-4844-9463-f41e60b2fc60_1000x859.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. The f&#178; matrix: (1) the item f-square &gt; Matrix, (2) the f&#178; matrix, (3) CONFIRM to SATISF, just below the 0.15 threshold, (4) CONFIRM to INTENT, below 0.02, so SmartPLS shows it in red (0.334 and 0.183 appear in green).</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/9uR0f/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8692957d-7015-43e3-96d9-790bd1db4c11_1220x822.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b70fe72f-e166-4cbe-90a7-110ac411d71d_1220x942.png&quot;,&quot;height&quot;:410,&quot;title&quot;:&quot;Table 3. Path coefficients and f&#178; effect sizes of the seven hypotheses&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/9uR0f/1/" width="730" height="410" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!k_ee!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!k_ee!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 424w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 848w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 1272w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!k_ee!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png" width="1900" height="800" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:800,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:79023,&quot;alt&quot;:&quot;Figure 5. Effect size f&#178;&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Effect size f&#178;" title="Figure 5. Effect size f&#178;" srcset="https://substackcdn.com/image/fetch/$s_!k_ee!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 424w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 848w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 1272w, https://substackcdn.com/image/fetch/$s_!k_ee!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2067a012-1cdd-4e48-b53d-b8933792c24e_1900x800.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. f&#178; of the seven paths. Green is 0.15 or more, dark blue is from 0.02 to below 0.15, orange is below 0.02. The dashed lines are the thresholds 0.02, 0.15 and 0.35.</figcaption></figure></div><h3>Hand calculation of f&#178; for TRUST to INTENT</h3><ol><li><p><strong>Take</strong> the R&#178; of INTENT with all four predictors: 0.31755.</p></li><li><p><strong>Run</strong> the model again without TRUST; the R&#178; of INTENT drops to 0.24765 (script <code>structural_model.R</code>).</p></li><li><p><strong>Compute</strong> f&#178; = (0.31755 - 0.24765) / (1 - 0.31755) = 0.06990 / 0.68245 = <strong>0.102</strong>, the same as SmartPLS.</p></li></ol><p>For an endogenous construct with only one predictor (VALUE), removing that predictor makes R&#178; equal to 0, so f&#178; = R&#178; / (1 - R&#178;) = 0.251 / 0.749 = 0.334.</p><p><strong>Reading the results:</strong></p><ul><li><p><strong>Expectation confirmation has the largest effect</strong>, on perceived value (f&#178; = 0.334), in line with expectation-confirmation theory.</p></li><li><p><strong>Perceived value has a medium effect on satisfaction</strong> (0.183); the effect of confirmation on satisfaction (0.149) is <strong>just below the medium threshold</strong>. The correct report is "small, just below medium", not rounded up to "medium".</p></li><li><p><strong>For repurchase intention, trust makes the largest unique contribution</strong> (0.102), although all four effects are small.</p></li><li><p><strong>Confirmation has almost no direct effect on intention</strong> (f&#178; = 0.009). This agrees with Post 31: H4 was not statistically significant (<em>p</em> = 0.086). The effect of confirmation on intention, if any, runs through perceived value and satisfaction; <a href="https://www.phuonglib.com/p/mediation-analysis-smartpls">Post 40</a> tests this indirect effect.</p></li></ul><h2>4. Reporting the results</h2><blockquote><p>There is no collinearity between the latent variables: VIF values range from 1.000 to 1.681, below 3 (Hair et al., 2019). The model explains 25.1% of the variance of perceived value, 39.9% of satisfaction and 31.8% of repurchase intention (adjusted R&#178; of 0.248, 0.394 and 0.307 respectively), so explanatory power is weak to moderate. In terms of effect size (Cohen, 1988), expectation confirmation has a medium effect, close to large, on perceived value (f&#178; = 0.334); perceived value has a medium effect on satisfaction (f&#178; = 0.183); the effects on repurchase intention are all small, with trust the largest (f&#178; = 0.102). The direct effect of expectation confirmation on repurchase intention is negligible (f&#178; = 0.009).</p></blockquote><p>The table in a thesis should combine, for each hypothesis: the path coefficient, <em>t</em>, <em>p</em>, the confidence interval (Post 39), f&#178; and VIF; together with the R&#178; and adjusted R&#178; of each endogenous construct.</p><h2>Four common mistakes</h2><ol><li><p><strong>Reading the VIF of the indicators</strong> (Outer model) to judge collinearity in the structural model. Read the Inner model instead.</p></li><li><p><strong>Reading the f&#178; matrix the wrong way round</strong>: rows are predictors, columns are endogenous constructs.</p></li><li><p><strong>Concluding "the model is good" only because R&#178; is statistically significant.</strong> An R&#178; of 0.318 means that almost 70% of the variance of repurchase intention is still unexplained; mention this in the limitations and directions for future research.</p></li><li><p><strong>Rounding effect sizes</strong> (0.149 reported as "medium"). Report against the exact thresholds; when a value is close to a threshold, say so.</p></li></ol><h2>Frequently asked questions</h2><p><strong>How high does R&#178; need to be for a thesis?</strong></p><p>There is no common threshold for every field. Compare with studies on the same topic that used the same method. The PLS-SEM R&#178; is usually lower than the CB-SEM R&#178; on the same data (Figure 3).</p><p><strong>Does f&#178; need a significance test?</strong></p><p>f&#178; is a measure of size; it does not replace the significance test of the path coefficient (Post 39). Report both: whether the coefficient is significant and, if it is, how large the effect is.</p><p><strong>Should I report R&#178; or adjusted R&#178;?</strong></p><p>Report both. Adjusted R&#178; is useful when you compare models with different numbers of predictors, for example a model with and without a moderator (<a href="https://www.phuonglib.com/p/moderation-analysis-smartpls">Post 41</a>).</p><div><hr></div><h2>References</h2><p>Cohen, J. (1988). <em>Statistical power analysis for the behavioral sciences</em> (2nd ed.). Lawrence Erlbaum Associates.</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p>]]></content:encoded></item><item><title><![CDATA[Formative Measurement in SmartPLS 4: Redundancy Analysis, VIF and Weights]]></title><description><![CDATA[Evaluate formative constructs in three steps, redundancy analysis, collinearity and bootstrapped outer weights, and decide which indicators to keep]]></description><link>https://www.phuonglib.com/p/formative-measurement-model-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/formative-measurement-model-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:39 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/ffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!HOXx!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!HOXx!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!HOXx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:729758,&quot;alt&quot;:&quot;A stone arch built from distinct stones of different shapes, an orange keystone at the top, a small mason checking each stone.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A stone arch built from distinct stones of different shapes, an orange keystone at the top, a small mason checking each stone." title="A stone arch built from distinct stones of different shapes, an orange keystone at the top, a small mason checking each stone." srcset="https://substackcdn.com/image/fetch/$s_!HOXx!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!HOXx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0a3c2-7cf2-4786-87e1-574fe8c1ef35_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Formative indicators form their construct, so loadings above 0.708, Cronbach's alpha, AVE and HTMT no longer apply.</p></li><li><p>Convergent validity comes from a redundancy analysis against a global single item, which must be planned when you design the questionnaire; the path should reach 0.708.</p></li><li><p>Collinearity is checked with VIF: 5 or higher is critical and below 3 is ideal. In the Corporate reputation example the highest VIF is 2.269.</p></li><li><p>A non-significant outer weight is not a reason to remove an indicator: keep it if its loading is 0.50 or higher or significant, as for all five such indicators here.</p></li><li><p>The measurement mode must follow the definition of the construct, never the results.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">Post 34</a> and <a href="https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls">Post 35</a> evaluated the <strong>reflective</strong> measures of the Marketplace X model. Not every construct is measured that way, though. Some constructs are <strong>formed</strong> by their components; their indicators are called <strong>formative</strong>. For this type of measure, the 0.708 loading threshold, Cronbach's alpha, AVE and HTMT <strong>no longer apply</strong>. This post presents the three evaluation steps specific to formative measurement models (Hair et al., 2019).</p><p>All five constructs of the Marketplace X study are reflective, so this post uses the <strong>sample project "Corporate reputation (primer)"</strong> that comes with SmartPLS. It is the classic example of the book by Hair et al. (2022) and has four carefully designed formative constructs.</p><blockquote><p><strong>Sample scripts.</strong> The R script and output tables for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch36-formative-measurement-models">research-data-analysis repository on GitHub</a> (folder <code>scripts/</code>). The data belong to SmartPLS GmbH and are not redistributed; the file <code>data/NOTICE.txt</code> explains how to open the sample project and point the script to it. Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> All figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026; 30-day trial) with the sample project Corporate reputation (primer): 344 respondents, missing values handled by Mean replacement as set in the project. The citation note with a book-cover image shown on the canvas of the sample project has been painted white or cropped out of the screenshots. Every number was cross-checked with the R package cSEM 0.7.1 (Rademaker &amp; Schuberth, 2026) in the script <code>evaluate_formative.R</code> (in the repository's <code>scripts/</code> folder); bootstrap <em>p</em> values may differ in the third decimal between the two programs because the random bootstrap samples differ. The sample data belong to SmartPLS and are not included in the repository; install the sample project yourself (<a href="https://www.phuonglib.com/p/install-smartpls-4-interface">Post 32</a>, Section 4).</p></blockquote><h2>1. How do formative measures differ from reflective ones?</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/23jMS/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a0f9adfa-c0bd-4972-a98c-c4d7d2cb0877_1220x738.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0ce361a5-54b5-422c-8201-bac28bea6f09_1220x858.png&quot;,&quot;height&quot;:384,&quot;title&quot;:&quot;Table 1. Reflective and formative measurement models compared&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:false}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/23jMS/1/" width="730" height="384" frameborder="0" scrolling="no"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Example from the sample project: <strong>QUAL</strong> (quality of the company) is formed by eight indicators, qual_1 to qual_8, each about a different aspect of quality. A company can be very good in one aspect but weak in another: the indicators need not move in the same direction, and each one contributes part of the overall construct (Diamantopoulos &amp; Winklhofer, 2001).</p><p>The extended model of the sample project (Figure 1) has four formative constructs: <strong>ATTR</strong> (attractiveness), <strong>CSOR</strong> (corporate social responsibility), <strong>PERF</strong> (performance) and <strong>QUAL</strong> (quality). They affect two dimensions of corporate reputation, <strong>COMP</strong> (competence) and <strong>LIKE</strong> (likeability), which are reflective; these two in turn affect customer <strong>CUSA</strong> (satisfaction) and <strong>CUSL</strong> (loyalty).</p><p><strong>Step 1.</strong> Open the project <strong>Example &#8211; Corporate reputation (primer)</strong> in the Workspace pane and double-click the model <strong>2 Extended model</strong>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!s-xy!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!s-xy!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 424w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 848w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 1272w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!s-xy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png" width="1000" height="626" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:626,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:130048,&quot;alt&quot;:&quot;Figure 1. Extended model of the sample project&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Extended model of the sample project" title="Figure 1. Extended model of the sample project" srcset="https://substackcdn.com/image/fetch/$s_!s-xy!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 424w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 848w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 1272w, https://substackcdn.com/image/fetch/$s_!s-xy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fea54dfbb-575e-4465-917c-5f59f8a83bc6_1000x626.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. The model 2 Extended model: (1) the formative construct ATTR, with arrows from the indicators to the latent variable, (2) the reflective construct COMP, with arrows from the latent variable to the indicators.</figcaption></figure></div><p>To switch a measure from reflective to formative or back, right-click the latent variable and choose <strong>Invert measurement model</strong> (<a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>). The choice of measurement mode must rest on the <strong>theoretical content</strong> of the construct, not on the results.</p><h2>2. Step 1: Convergent validity with redundancy analysis</h2><p>Does the formative measure capture the <strong>right</strong> construct? The check is a <strong>redundancy analysis</strong> (Chin, 1998): you relate the formative construct to another measure of the <strong>same</strong> construct, usually a global single item. For ATTR, for example, the global item attr_global asks about the overall attractiveness of the company.</p><p><strong>Threshold:</strong> a path coefficient between the two measures of <strong>0.708 or higher</strong>, that is, an R&#178; of at least 0.50 (Hair et al., 2019).</p><p>Redundancy analysis has to be <strong>planned when you design the questionnaire</strong>: you need a global item for each formative construct. It cannot be added after the data are collected.</p><p><strong>Step 2.</strong> Open the model <strong>3a Redundancy analysis ATTR</strong> (already drawn in the sample project), choose <strong>Calculate &gt; PLS-SEM algorithm</strong> and click <strong>Start calculation</strong> (Figure 2).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ymCv!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 424w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 848w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 1272w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ymCv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png" width="1000" height="214" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:214,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:59670,&quot;alt&quot;:&quot;Figure 2. Redundancy analysis of ATTR&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Redundancy analysis of ATTR" title="Figure 2. Redundancy analysis of ATTR" srcset="https://substackcdn.com/image/fetch/$s_!ymCv!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 424w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 848w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 1272w, https://substackcdn.com/image/fetch/$s_!ymCv!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0c6fe9bb-cef4-4ae9-89e0-f14fa2a2873f_1000x214.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 2. Redundancy analysis: ATTR_F is ATTR measured by the three formative indicators, ATTR_G is ATTR measured by the global item attr_global. (1) Path coefficient 0.874, (2) R&#178; = 0.764.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Gsozq/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/188988d9-a8e4-44c4-9d5c-8777a4d3990a_1220x472.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/46977376-f257-4d03-9d41-774de17a4a72_1220x630.png&quot;,&quot;height&quot;:283,&quot;title&quot;:&quot;Table 2. Redundancy analysis of the four formative constructs&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Gsozq/1/" width="730" height="283" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> all four formative measures have path coefficients above 0.708: <strong>convergent validity is established</strong>. Note that the weights of the three ATTR indicators in the redundancy model (0.573, 0.176 and 0.523) differ from those in the full model, because weights in PLS-SEM depend on the latent variables the construct is connected to. The weights evaluated in Step 3 are those of the full model.</p><h2>3. Step 2: Collinearity (VIF)</h2><p>The outer weights of formative indicators are estimated like coefficients of a multiple regression. When the indicators correlate too strongly with each other, the weights become unstable and may change sign or lose significance. The <strong>VIF</strong> (variance inflation factor) measures this.</p><p><strong>Thresholds</strong> (Hair et al., 2019): a VIF of <strong>5 or higher</strong> signals a critical problem; ideally the VIF is <strong>below 3</strong>. VIF values between 3 and 5 may already have an effect.</p><p><strong>Step 3.</strong> Run the <strong>PLS-SEM algorithm</strong> on the model 2 Extended model, then open <strong>Quality criteria &gt; Collinearity statistics (VIF) &gt; Outer model &#8211; List</strong> (Figure 3).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!DPW9!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!DPW9!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 424w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 848w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 1272w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!DPW9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png" width="700" height="700" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:700,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:171554,&quot;alt&quot;:&quot;Figure 3. VIF of the indicators&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. VIF of the indicators" title="Figure 3. VIF of the indicators" srcset="https://substackcdn.com/image/fetch/$s_!DPW9!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 424w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 848w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 1272w, https://substackcdn.com/image/fetch/$s_!DPW9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07ab3ccf-5911-432d-9a7d-931a61b9a14a_700x700.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. VIF of the indicators: (1) the entry Outer model &#8211; List, (2) the highest VIF of a formative indicator, qual_3. Some cells are truncated (2.2...) because the table is zoomed out; the full values are in the table below.</figcaption></figure></div><p>This table lists the VIF of <strong>every</strong> indicator. Read only the VIF of the <strong>formative</strong> indicators; the VIF of reflective indicators (comp, like, cusl) is not used in this evaluation.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/lq4dy/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a3cb9925-d776-4ffc-b81c-2397300cccf2_1220x472.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/58186d16-f1e8-4d23-9006-1a9512513849_1220x542.png&quot;,&quot;height&quot;:264,&quot;title&quot;:&quot;Table 3. VIF of the formative indicators&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/lq4dy/1/" width="730" height="264" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Result:</strong> the highest VIF is 2.269, <strong>below 3</strong>: there is no collinearity problem.</p><h2>4. Step 3: Significance and relevance of the outer weights</h2><p>The <strong>outer weight</strong> shows the <strong>relative</strong> contribution of each indicator to the construct, controlling for the other indicators. The <strong>outer loading</strong> shows the <strong>absolute</strong> contribution: the correlation of the indicator with the construct on its own (Cenfetelli &amp; Bassellier, 2009). Testing significance requires <strong>bootstrapping</strong>.</p><p><strong>Step 4.</strong> Choose <strong>Calculate &gt; Bootstrapping</strong>, keep the settings of the sample project (5000 subsamples) and click <strong>Start calculation</strong>. In the report, open <strong>Final results &gt; Outer weights &gt; Mean, STDEV, T values, p values</strong> (Figure 4). <a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a> explains the bootstrap settings in detail.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Q5Hb!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 424w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 848w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 1272w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png" width="1000" height="597" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:597,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:291964,&quot;alt&quot;:&quot;Figure 4. Bootstrapping of the outer weights&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Bootstrapping of the outer weights" title="Figure 4. Bootstrapping of the outer weights" srcset="https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 424w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 848w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 1272w, https://substackcdn.com/image/fetch/$s_!Q5Hb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0dd1de09-26d6-4771-9972-766aa2577205_1000x597.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Outer weights and bootstrap tests: (1) the entry Outer weights &gt; Mean, STDEV, T values, p values, (2) csor_2 and csor_4, (3) qual_2, qual_3 and qual_4: the five indicators with p values above 0.05, shown in red by SmartPLS.</figcaption></figure></div><p>Five indicators have <strong>non-significant</strong> weights:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/ciKDY/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/01cafb14-445a-4f3f-b240-a56949629fa1_1220x582.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f072a291-a415-47e1-8e2b-2afb71754a99_1220x740.png&quot;,&quot;height&quot;:339,&quot;title&quot;:&quot;Table 4. Formative indicators with non-significant outer weights&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/ciKDY/1/" width="730" height="339" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Do not remove these indicators straight away. The decision rule (Hair et al., 2019) is summarized in Figure 5:</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!NL4i!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!NL4i!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 424w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 848w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 1272w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!NL4i!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png" width="1675" height="756" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:756,&quot;width&quot;:1675,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:107451,&quot;alt&quot;:&quot;Figure 5. Keeping or removing a formative indicator&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Keeping or removing a formative indicator" title="Figure 5. Keeping or removing a formative indicator" srcset="https://substackcdn.com/image/fetch/$s_!NL4i!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 424w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 848w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 1272w, https://substackcdn.com/image/fetch/$s_!NL4i!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F471b39c4-b241-4679-9eb3-1945bcc3defc_1675x756.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. Decision process for keeping or removing a formative indicator.</figcaption></figure></div><ul><li><p><strong>Significant weight:</strong> keep the indicator.</p></li><li><p><strong>Non-significant weight, but a loading of 0.50 or higher, or a significant loading:</strong> usually keep it, because the indicator makes a substantial absolute contribution even though its relative contribution next to the other indicators is small.</p></li><li><p><strong>Both weight and loading low and non-significant:</strong> consider removing it, but check the content first: removing a formative indicator removes part of the definition of the construct.</p></li></ul><p><strong>Step 5.</strong> In the same bootstrap report, open <strong>Final results &gt; Outer loadings &gt; Mean, STDEV, T values, p values</strong> (Figure 6) to see the loadings.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!w4ft!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!w4ft!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 424w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 848w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 1272w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!w4ft!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png" width="1000" height="597" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:597,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:294781,&quot;alt&quot;:&quot;Figure 6. Bootstrapping of the outer loadings&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Bootstrapping of the outer loadings" title="Figure 6. Bootstrapping of the outer loadings" srcset="https://substackcdn.com/image/fetch/$s_!w4ft!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 424w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 848w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 1272w, https://substackcdn.com/image/fetch/$s_!w4ft!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F77510b1c-4f33-460e-9a5e-7e4eef47ba7e_1000x597.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. Outer loadings and bootstrap tests: (1) the entry Outer loadings, (2) csor_2 and csor_4, (3) qual_2 to qual_4: all loadings are 0.57 or higher and significant.</figcaption></figure></div><p><strong>Result:</strong> the five indicators have loadings from 0.570 to 0.749, all significant (<em>p</em> &lt; 0.001): <strong>keep all five</strong>. Figure 7 places the weights next to the loadings for all 21 formative indicators.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!CEIf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!CEIf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 424w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 848w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 1272w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!CEIf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png" width="1900" height="1240" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1240,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:173318,&quot;alt&quot;:&quot;Figure 7. Outer weights and outer loadings&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Outer weights and outer loadings" title="Figure 7. Outer weights and outer loadings" srcset="https://substackcdn.com/image/fetch/$s_!CEIf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 424w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 848w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 1272w, https://substackcdn.com/image/fetch/$s_!CEIf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc7258457-ecbc-4d2b-ba32-f2d606cc0582_1900x1240.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Outer weights (dark bars) and outer loadings (light bars) of the 21 formative indicators, computed with cSEM. Orange marks the five indicators with non-significant weights; all five have loadings above 0.50.</figcaption></figure></div><p>Two points stand out in Figure 7:</p><ul><li><p><strong>The negative weight of qual_4 (&#8722;0.005)</strong> does not mean that quality in this aspect lowers QUAL. The weight is close to zero and not significant; the positive loading (0.664) shows that qual_4 still moves in the same direction as the construct.</p></li><li><p><strong>The indicator with the largest weight</strong> in each construct (attr_3, csor_5, perf_1, qual_6) is the aspect the company should prioritize when it wants to improve that construct. This is the practical reading of formative weights.</p></li></ul><h2>5. Reporting the results</h2><blockquote><p>The four formative measures were evaluated following Hair et al. (2019). The redundancy analysis yielded path coefficients from 0.805 to 0.874, above the 0.708 threshold, establishing convergent validity. The VIF values of the formative indicators ranged from 1.129 to 2.269, below 3, indicating no collinearity. Bootstrapping (5000 subsamples) showed that 16 of the 21 outer weights were significant at the 5% level. The five indicators with non-significant weights (csor_2, csor_4, qual_2, qual_3, qual_4) all had loadings of 0.570 or higher that were significant (<em>p</em> &lt; 0.001), so they were retained for their absolute contribution to the construct.</p></blockquote><p>The table in your thesis should give, for each formative indicator: the outer weight, the <em>p</em> value of the weight, the outer loading, the <em>p</em> value of the loading and the VIF; add a small table for the redundancy analysis.</p><h2>Five common mistakes</h2><ol><li><p><strong>Evaluating formative measures with alpha, AVE or HTMT.</strong> These criteria apply to reflective measures only.</p></li><li><p><strong>Removing an indicator only because its weight is not significant</strong>, without looking at its loading.</p></li><li><p><strong>Choosing formative or reflective based on the results</strong> (flipping the arrows to make the criteria look better). The measurement mode must follow from the definition of the construct.</p></li><li><p><strong>No global item</strong> in the questionnaire, so redundancy analysis is impossible. Add one global item for each formative construct when you design the questionnaire.</p></li><li><p><strong>Reading the VIF of reflective indicators</strong> to draw conclusions about collinearity in the formative measurement model.</p></li></ol><h2>Frequently asked questions</h2><p><strong>How do I know whether a construct is formative or reflective?</strong></p><p>Ask: if the construct increases, do all indicators increase together? If yes (for example, when satisfaction rises, all four SAT items rise), it is reflective. If the indicators are separate components, and one rising does not pull the others up (for example, product quality and service quality), it is formative (Diamantopoulos &amp; Winklhofer, 2001).</p><p><strong>Does a formative measure need many indicators?</strong></p><p>It needs enough indicators to cover every aspect of the construct according to its definition. A missing aspect means a missing part of the construct, unlike reflective measures, where four items are interchangeable.</p><p><strong>Can I use formative measures in AMOS?</strong></p><p>CB-SEM has difficulty estimating formative measures: a formative construct must be connected to at least two other variables to be identified. This is one reason to choose PLS-SEM when the model has formative constructs (<a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31</a>).</p><p><strong>Can the Student license run this example?</strong></p><p>The Student license includes the PLS-SEM algorithm and bootstrapping, but it accepts at most 100 rows of data (Post 32); the sample project has 344 rows, so you need the trial or the Professional license.</p><div><hr></div><h2>References</h2><p>Cenfetelli, R. T., &amp; Bassellier, G. (2009). Interpretation of formative measurement in information systems research. <em>MIS Quarterly, 33</em>(4), 689&#8211;707. https://doi.org/10.2307/20650323</p><p>Chin, W. W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (Ed.), <em>Modern methods for business research</em> (pp. 295&#8211;336). Lawrence Erlbaum Associates.</p><p>Diamantopoulos, A., &amp; Winklhofer, H. M. (2001). Index construction with formative indicators: An alternative to scale development. <em>Journal of Marketing Research, 38</em>(2), 269&#8211;277. https://doi.org/10.1509/jmkr.38.2.269.18845</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>Rademaker, M. E., &amp; Schuberth, F. (2026). <em>cSEM: Composite-based structural equation modeling</em> (R package version 0.7.1) [Computer software]. https://CRAN.R-project.org/package=cSEM</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p>]]></content:encoded></item><item><title><![CDATA[Importing Data and Drawing a PLS-SEM Model in SmartPLS 4]]></title><description><![CDATA[Import a CSV file into SmartPLS 4 with the missing value marker -99, draw the constructs and paths, then check the model with one PLS-SEM run]]></description><link>https://www.phuonglib.com/p/import-data-draw-model-smartpls-4</link><guid isPermaLink="false">https://www.phuonglib.com/p/import-data-draw-model-smartpls-4</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:36 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Uz80!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Uz80!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!Uz80!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:787160,&quot;alt&quot;:&quot;Small paper cards flowing from an open filing drawer onto a drafting table, arranged into five clusters joined by arrows, one cluster orange.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Small paper cards flowing from an open filing drawer onto a drafting table, arranged into five clusters joined by arrows, one cluster orange." title="Small paper cards flowing from an open filing drawer onto a drafting table, arranged into five clusters joined by arrows, one cluster orange." srcset="https://substackcdn.com/image/fetch/$s_!Uz80!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Uz80!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a461ef-574f-4720-9d28-d20a25d6ed74_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>SmartPLS reads CSV, TXT, SPSS and Excel files, but the Student license accepts only CSV files of at most 100 rows.</p></li><li><p>Declare the missing value marker (-99) in the import dialog; otherwise SmartPLS treats -99 as a real answer.</p></li><li><p>Drag a group of items onto the canvas and name the construct; connect the seven paths with the Connect tool.</p></li><li><p>One run of the PLS-SEM algorithm matches cSEM to the third decimal (H1 = 0.501, R&#178; of intention = 0.318).</p></li><li><p>Casewise deletion keeps 254 of 267 respondents; report the number of observations actually used.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/install-smartpls-4-interface">Post 32, "How to Install SmartPLS 4 and Find Your Way Around the Interface"</a> created the project <strong>Marketplace X</strong> and introduced the interface. This post brings the Marketplace X data into the project, draws the five constructs with their 20 items, connects the seven paths H1 to H7 as in <a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a> and then runs the PLS-SEM algorithm once to make sure the model is drawn correctly. Evaluating the results starts in <a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">Post 34, "Evaluating the Reflective Measurement Model in SmartPLS 4"</a>.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch33-import-data-draw-model-smartpls">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the screenshots and numbers.</strong> All figures are real screenshots of SmartPLS 4.1.2.1 (Ringle et al., 2026) on macOS (Apple silicon Mac). Most were taken in November 2026 with the 30-day trial, which has all features of the Professional license (Post 32). Figures 1 and 3 were taken on 6 November 2026 with the Student license; Figure 11 and the left half of Figure 13 also date from 6 November 2026 and were taken with the trial. Because the Student license does not accept data with more than 100 rows (Section 1), the screens that only appear with the Student license (the SPSS import attempt and the <em>not licensed</em> data file) are described in words. The numbers were cross-checked with the R package cSEM 0.7.1 (Rademaker &amp; Schuberth, 2026), script <code>compare_pls.R</code>.</p></blockquote><h2>1. Prepare the data file</h2><p>SmartPLS 4 reads four file types: CSV, TXT (tab-delimited), SPSS (.sav) and Excel (.xls, .xlsx). The first row of the file is used as the variable names (SmartPLS GmbH, n.d.-a). Check three requirements before importing:</p><ul><li><p><strong>One delimiter and one decimal separator throughout the file,</strong> with no empty or extra columns (SmartPLS GmbH, n.d.-e).</p></li><li><p><strong>One convention for missing values:</strong> leave the cell empty, or write a code that cannot occur in the data, for example -99 (SmartPLS GmbH, n.d.-b).</p></li><li><p><strong>A suitable license:</strong> the Student license imports only CSV files and at most 100 rows of data; the trial and the Professional license have no such limit (SmartPLS GmbH, n.d.-d).</p></li></ul><p>The Marketplace X data set used in the AMOS posts (<code>marketplace_x_sem.sav</code>) has 267 respondents. Across the 20 items there are 21 blank cells, spread over 13 respondents (0.39% of the cells); the item with the most blanks is CONF3 with 3 cells, that is, 1.12%. So that the file works with both the Student license and the full versions, this post converts it to CSV with the script <code>make_smartpls_csv.R</code>. The script <strong>only changes the format</strong>: comma as delimiter, decimal point, blank cells written as -99. No other value is changed. The result is the file <code>marketplace_x_smartpls.csv</code> with 267 rows and 26 columns: the 20 items, the demographic variables and the column N_MISSING (the number of blank items of each respondent).</p><h2>2. Import the data into the project</h2><p><strong>Step 1.</strong> Select the project <strong>Marketplace X</strong> in the Workspace pane, then choose <strong>Files &gt; Import data file</strong> (Figure 1). You can also click the <strong>Import data file</strong> link right below the project name.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!2Kuj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2Kuj!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 424w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 848w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 1272w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!2Kuj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png" width="360" height="651" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:651,&quot;width&quot;:360,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:96503,&quot;alt&quot;:&quot;Figure 1. Files > Import data file&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Files > Import data file" title="Figure 1. Files > Import data file" srcset="https://substackcdn.com/image/fetch/$s_!2Kuj!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 424w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 848w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 1272w, https://substackcdn.com/image/fetch/$s_!2Kuj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9bfed239-f8a3-45e0-9aa2-6e4455614acd_360x651.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. The Files menu: (1) the Import data file command. This menu was captured with the Student license, so it lists three model types; with the trial, the menu also lists PROCESS and GSCA models (Post 32, Figure 10).</figcaption></figure></div><p><strong>Step 2.</strong> In the file dialog, go to the folder that holds the data, select <code>marketplace_x_smartpls.csv</code> and click <strong>Open</strong> (Figure 2).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!JX9v!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!JX9v!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 424w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 848w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 1272w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!JX9v!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png" width="760" height="387" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:387,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:99929,&quot;alt&quot;:&quot;Figure 2. Choosing the CSV file&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Choosing the CSV file" title="Figure 2. Choosing the CSV file" srcset="https://substackcdn.com/image/fetch/$s_!JX9v!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 424w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 848w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 1272w, https://substackcdn.com/image/fetch/$s_!JX9v!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa26b6a6c-758d-4c7a-b887-78182948b78a_760x387.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. The file dialog on macOS with the trial license: (1) the selected CSV file. The SPSS files (.sav) are listed normally, because the trial can import them.</figcaption></figure></div><blockquote><p><strong>Importing an SPSS file with the Student license.</strong> In an earlier test, importing the .sav version of the same data with the Student license was tried. SmartPLS still opened the <strong>SPSS import</strong> dialog and read the data correctly: 267 respondents, 22 missing values and the number of blank cells per item matched the original file (CONF1 2 blanks, CONF3 3 blanks). A .sav file marks its own missing values, so no code needs to be declared. But when <strong>Import</strong> was clicked, SmartPLS showed <em>Feature locked</em> (Figure 3): the Student license cannot import SPSS files. From the next attempt on, .sav files were grayed out in the file dialog. With the Student license, use the CSV file.</p></blockquote><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!5NNx!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 424w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 848w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 1272w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!5NNx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png" width="420" height="211" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:211,&quot;width&quot;:420,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:43979,&quot;alt&quot;:&quot;Figure 3. Feature locked message&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Feature locked message" title="Figure 3. Feature locked message" srcset="https://substackcdn.com/image/fetch/$s_!5NNx!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 424w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 848w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 1272w, https://substackcdn.com/image/fetch/$s_!5NNx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F76b91d37-6189-453c-9d71-22ea2d4942e6_420x211.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 3. The Feature locked message shown when importing an SPSS file with the Student license (1).</figcaption></figure></div><p><strong>Step 3.</strong> The <strong>CSV import</strong> dialog opens (Figure 4). Check the four settings in the second row: <strong>Delimiter character</strong> is <em>Comma</em>, <strong>Escape character</strong> is <em>None</em>, <strong>Locale</strong> is <em>US (example: 1,000.23)</em> (a decimal point) and <strong>Encoding</strong> is <em>UTF-8</em>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!M6fx!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!M6fx!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 424w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 848w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 1272w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!M6fx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png" width="700" height="517" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:517,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:105124,&quot;alt&quot;:&quot;Figure 4. CSV import before declaring the missing value marker&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. CSV import before declaring the missing value marker" title="Figure 4. CSV import before declaring the missing value marker" srcset="https://substackcdn.com/image/fetch/$s_!M6fx!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 424w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 848w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 1272w, https://substackcdn.com/image/fetch/$s_!M6fx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0bd6dd52-cecb-4024-bd27-58e4e9eddaf7_700x517.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. The CSV import dialog before the missing value marker is declared: (1) settings for reading the file, (2) Missing: 0, (3) the Min column with values of -99, (4) the missing value marker link.</figcaption></figure></div><p>Look closely at Figure 4: the top line reads <strong>Missing: 0</strong>, and the <strong>Min</strong> column shows <strong>-99</strong> for AGEGROUP, CONF1, CONF3, CONF4 and PV1. SmartPLS treats -99 as a real answer. If you click Import now, the -99 values enter the calculations and distort every result. This is the easiest mistake to make when importing data.</p><p><strong>Step 4.</strong> Click the <strong>missing value marker</strong> link at the bottom of the dialog. Type <strong>-99</strong> in the <strong>Missing</strong> box, then click <strong>Apply</strong> (Figure 5).</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!-71C!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 424w, https://substackcdn.com/image/fetch/$s_!-71C!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 848w, https://substackcdn.com/image/fetch/$s_!-71C!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 1272w, https://substackcdn.com/image/fetch/$s_!-71C!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!-71C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png" width="300" height="206" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:206,&quot;width&quot;:300,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:16417,&quot;alt&quot;:&quot;Figure 5. Declaring the missing value marker&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Declaring the missing value marker" title="Figure 5. Declaring the missing value marker" srcset="https://substackcdn.com/image/fetch/$s_!-71C!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 424w, https://substackcdn.com/image/fetch/$s_!-71C!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 848w, https://substackcdn.com/image/fetch/$s_!-71C!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 1272w, https://substackcdn.com/image/fetch/$s_!-71C!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4331c807-c4b9-4372-95a2-2da29c60ce10_300x206.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 5. The Missing dialog: (1) the missing value code -99, (2) the Apply button.</figcaption></figure></div><p><strong>Step 5.</strong> Check the dialog again (Figure 6). The top line must read <strong>Missing: 22</strong>, the Missing column shows the number of blank cells of each variable, the Min column is back to 1 and the bottom of the dialog reads <em>Your marker -99 matches 22 fields</em>. The 22 consists of 21 cells in the items and 1 cell in AGEGROUP. Click <strong>Import</strong>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Gvfr!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Gvfr!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 424w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 848w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 1272w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Gvfr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png" width="700" height="517" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:517,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:98001,&quot;alt&quot;:&quot;Figure 6. CSV import after declaring -99&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. CSV import after declaring -99" title="Figure 6. CSV import after declaring -99" srcset="https://substackcdn.com/image/fetch/$s_!Gvfr!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 424w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 848w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 1272w, https://substackcdn.com/image/fetch/$s_!Gvfr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6f8846eb-e07f-4178-8333-51c5e0fda637_700x517.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. After the declaration: (1) Missing: 22, (2) the number of blank cells of each variable, (3) confirmation that the marker -99 matches 22 cells.</figcaption></figure></div><h2>3. Read the data view</h2><p>After the import, SmartPLS opens the <strong>data view</strong> (Figure 7).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!4w4l!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!4w4l!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!4w4l!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:229565,&quot;alt&quot;:&quot;Figure 7. The data view&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. The data view" title="Figure 7. The data view" srcset="https://substackcdn.com/image/fetch/$s_!4w4l!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!4w4l!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65ef7632-a6d4-41a6-b852-7742455266fd_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Data view of the file marketplace_x_smartpls. The numbers are explained right below.</figcaption></figure></div><ol><li><p><strong>Toolbar:</strong> Back (return to the project), Setup (reopen the import settings), Derive indicators (create new variables), Add group and Generate groups (create data groups for multigroup analysis), Clear groups, Export to Excel / CSV.</p></li><li><p><strong>Tabs:</strong> Indicators (statistics of each variable), Correlations (correlation matrix), Data groups, Raw data.</p></li><li><p><strong>Summary line:</strong> <em>26 indicators, 267 samples, 22 missing values</em>, that is, 26 variables, 267 respondents and 22 blank cells.</p></li><li><p><strong>Missing values column:</strong> the number of blank cells of each variable. Compare it with the original file before you continue, as the SmartPLS documentation advises (SmartPLS GmbH, n.d.-b).</p></li><li><p><strong>Standard deviation column.</strong></p></li></ol><p>The table also shows the mean, median, minimum and maximum of the scale (Scale min, Scale max) and of the data (Observed min, Observed max). The table has more columns than the window shows; drag the horizontal scroll bar to see the rest.</p><p>Two points when comparing with SPSS (the SPSS posts):</p><ul><li><p><strong>The means match:</strong> CONF1 has a mean of 3.875 in both SmartPLS and R.</p></li><li><p><strong>The standard deviation is slightly smaller:</strong> CONF1 has 0.961 in SmartPLS but 0.963 in SPSS and R. The reason is that SmartPLS divides by n, whereas SPSS and R divide by n - 1. The difference has no effect on the PLS-SEM results, but when you report descriptive statistics in your thesis, take all numbers from one program for consistency.</p></li></ul><blockquote><p><strong>If the data file shows "not licensed".</strong> With the Student license, the 267-row file can still be imported, but in the Workspace pane it is grayed out with the label <em>not licensed</em> next to its name, and it cannot be used in a model. This was observed in an earlier test with the Student license; after the trial was activated, the file returned to normal. The screenshots of the data in this post were taken with the trial, so the data file is shown as usable.</p></blockquote><h2>4. Draw the measurement model</h2><p><strong>Step 1.</strong> Select the project <strong>Marketplace X</strong>, click the <strong>PLS-SEM</strong> button on the toolbar, name the model (here <strong>marketplace-x-model</strong>) and click <strong>Save</strong>, as in Post 32. Because the project already holds data, the <strong>Indicators</strong> pane on the left lists the 26 variables (Figure 8).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!2vxg!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2vxg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!2vxg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/377342f8-2a26-4922-8f46-9efb81062850_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:102881,&quot;alt&quot;:&quot;Figure 8. Model window with the list of indicators&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Model window with the list of indicators" title="Figure 8. Model window with the list of indicators" srcset="https://substackcdn.com/image/fetch/$s_!2vxg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!2vxg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F377342f8-2a26-4922-8f46-9efb81062850_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. The model window: (1) the name of the data file in use, (2) the list of indicators, (3) the empty canvas.</figcaption></figure></div><p><strong>Step 2.</strong> In the Indicators pane, click <strong>CONF1</strong>, hold <strong>Shift</strong> and click <strong>CONF4</strong> to select all four items. Drag the group and drop it on the canvas. A naming box appears: type <strong>CONFIRM</strong> and press <strong>Enter</strong> (Figure 9). SmartPLS creates the construct CONFIRM measured by CONF1 to CONF4 (SmartPLS GmbH, n.d.-f). While several items are selected, the bottom of the Indicators pane shows the highest correlation among them; in the screenshot it reads <em>The best correlation between CONF2 and CONF4 is 0.580</em>.</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!b_Kg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 424w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 848w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 1272w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!b_Kg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png" width="800" height="230" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:230,&quot;width&quot;:800,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:40417,&quot;alt&quot;:&quot;Figure 9. Dragging four items onto the canvas and naming the construct&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Dragging four items onto the canvas and naming the construct" title="Figure 9. Dragging four items onto the canvas and naming the construct" srcset="https://substackcdn.com/image/fetch/$s_!b_Kg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 424w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 848w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 1272w, https://substackcdn.com/image/fetch/$s_!b_Kg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbbe3e613-ee41-42d7-9b6b-3f925fe61396_800x230.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 9. (1) The four items CONF1 to CONF4 selected, (2) the naming box for the construct; press Enter to create it. The naming box is a separate window; it was captured separately and placed onto the main window at its screen position.</figcaption></figure></div><p><strong>Naming tip:</strong> use a construct name that differs from the item names, in capital letters without spaces or accents, following the convention of the AMOS posts: CONFIRM, VALUE, SATISF, TRUST, INTENT. The later posts use the same names.</p><p><strong>Step 3.</strong> Repeat Step 2 for the other four constructs:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Nh6aO/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dd846bc1-522e-48a9-bb0c-beeb3f96dca3_1220x504.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/af18507a-50ca-4598-a67f-1d8e9a44cd5f_1220x624.png&quot;,&quot;height&quot;:264,&quot;title&quot;:&quot;Table 1. Constructs, items and suggested positions on the canvas&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Nh6aO/1/" width="730" height="264" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!kuDe!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!kuDe!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 424w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 848w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 1272w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!kuDe!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png" width="1000" height="499" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:499,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:81399,&quot;alt&quot;:&quot;Figure 10. Five constructs before connecting the paths&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. Five constructs before connecting the paths" title="Figure 10. Five constructs before connecting the paths" srcset="https://substackcdn.com/image/fetch/$s_!kuDe!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 424w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 848w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 1272w, https://substackcdn.com/image/fetch/$s_!kuDe!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F862c288c-286e-4ee9-9cb4-6a001a46df84_1000x499.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 10. The five new constructs, shown in red because they have no paths yet: (1) the construct CONFIRM with its four items.</figcaption></figure></div><p>Figure 10 shows two things:</p><ul><li><p><strong>The arrows point from the construct to the items:</strong> by default SmartPLS uses a <strong>reflective</strong> measurement model, which fits the five concepts of the Marketplace X study (<a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31, "What Is SmartPLS? PLS-SEM vs CB-SEM"</a>). If a concept is <strong>formative</strong>, right-click the construct and choose <strong>Invert measurement model</strong> to reverse the arrows (SmartPLS GmbH, n.d.-f). <a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">Post 36, "Evaluating Formative Measurement Models in SmartPLS 4"</a> discusses formative models.</p></li><li><p><strong>The red circles:</strong> on the computer used for the screenshots, a construct is red while it is not connected to any other construct, and turns blue once it is connected by a path (Section 5).</p></li></ul><p><strong>Step 4 (optional).</strong> To make the diagram easier to read, move the items outwards: right-click the construct, choose <strong>Align the indicators</strong>, then choose a direction. Or hold <strong>Alt + Shift</strong>, click the construct and drag in the direction where the items should go (SmartPLS GmbH, n.d.-f). The complete model in Section 5 places the items of CONFIRM and TRUST on the left, VALUE at the top, SATISF at the bottom and INTENT on the right.</p><blockquote><p><strong>Drawing without a mouse.</strong> The <strong>Accessibility</strong> menu of the model window has two commands, <strong>New construct</strong> and <strong>New connection</strong> (Figure 11). The New construct dialog has a name box, filled in from the selected items, and two boxes for the x and y position on the canvas. This is useful with a screen reader or when drag and drop is difficult.</p></blockquote><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!bHRh!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 424w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 848w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 1272w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!bHRh!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png" width="560" height="125" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:125,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:45926,&quot;alt&quot;:&quot;Figure 11. The Accessibility menu&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 11. The Accessibility menu" title="Figure 11. The Accessibility menu" srcset="https://substackcdn.com/image/fetch/$s_!bHRh!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 424w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 848w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 1272w, https://substackcdn.com/image/fetch/$s_!bHRh!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91880312-eddb-4acf-a72a-7b8f56c77237_560x125.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 11. The Accessibility menu: (1) the two commands New construct and New connection.</figcaption></figure></div><h2>5. Draw the structural model</h2><p><strong>Step 1.</strong> Click the <strong>Connect</strong> button on the toolbar.</p><p><strong>Step 2.</strong> Click the source construct, hold the mouse button, drag to the target construct and release (SmartPLS GmbH, n.d.-f). Repeat for the seven paths:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/34lW7/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4b9168cf-0a72-448b-912b-3181d7b15390_1220x694.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ef98f862-8f9a-4538-a3ef-f135b4df537a_1220x764.png&quot;,&quot;height&quot;:377,&quot;title&quot;:&quot;Table 2. The seven paths of the structural model&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/34lW7/1/" width="730" height="377" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><blockquote><p><strong>Note on hypothesis numbering.</strong> Posts <a href="https://www.phuonglib.com/p/quantitative-research-process-7-steps">2</a> and <a href="https://www.phuonglib.com/p/multiple-linear-regression-pspp-spss">14</a> use six hypotheses. From Post 26 and throughout the PLS-SEM part, the model has an extra direct path from Confirmation to Repurchase intention, so the series uses seven hypotheses and renumbers from H4: H4 = Confirmation to Repurchase intention (the new path, not in Post 2); H5 = Perceived value to Repurchase intention (H4 in Post 2); H6 = Satisfaction to Repurchase intention (H5 in Post 2); H7 = Trust to Repurchase intention (H6 in Post 2). H1 to H3 stay the same.</p></blockquote><p><strong>Step 3.</strong> Click <strong>Select</strong> to leave the connect mode, then click <strong>Save</strong> to save the model. Once the model is saved, the Save button turns pale.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!iNPR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!iNPR!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 424w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 848w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 1272w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!iNPR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png" width="1000" height="584" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/aed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:584,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:104649,&quot;alt&quot;:&quot;Figure 12. The complete Marketplace X model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 12. The complete Marketplace X model" title="Figure 12. The complete Marketplace X model" srcset="https://substackcdn.com/image/fetch/$s_!iNPR!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 424w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 848w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 1272w, https://substackcdn.com/image/fetch/$s_!iNPR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faed28ab0-85a7-4ed3-bff3-9d8b83b57e56_1000x584.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 12. The complete Marketplace X model in SmartPLS: five constructs, 20 items, seven paths. The constructs have turned blue.</figcaption></figure></div><p>Check Figure 12 before running anything:</p><ul><li><p><strong>All seven arrows between the constructs are present and point the right way.</strong> Connecting in the wrong direction (for example from VALUE to CONFIRM) is a common mistake; delete that path with the <strong>Delete</strong> button and connect it again.</p></li><li><p><strong>No red circles remain.</strong></p></li><li><p><strong>TRUST is connected only to INTENT</strong>, as in the model of Post 26.</p></li></ul><h2>6. Run the algorithm once to check the model</h2><p>Run the algorithm once to make sure the model is drawn correctly. This is not yet the evaluation step.</p><p><strong>Step 1.</strong> Choose <strong>Calculate &gt; PLS-SEM algorithm</strong>. The dialog has two tabs (Figure 13):</p><ul><li><p><strong>PLS setup:</strong> Weighting scheme is <em>Path</em> (path weighting scheme, the default and the recommended setting), Type of results is <em>Standardized</em>, Initial weight is <em>Default</em> (initial weights of 1) (SmartPLS GmbH, n.d.-c). The maximum number of iterations (3000) and the stop criterion (10 to the power of &#8722;7) are fixed in SmartPLS 4, so they no longer appear in the dialog.</p></li><li><p><strong>Data:</strong> set <strong>Missing value algorithm</strong> to <strong>Casewise deletion</strong> (reasons in Section 7). Leave Weighting vector at <em>None</em>.</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!dcJ2!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!dcJ2!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 424w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 848w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 1272w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!dcJ2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png" width="900" height="500" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:500,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:97927,&quot;alt&quot;:&quot;Figure 13. PLS-SEM algorithm dialog&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 13. PLS-SEM algorithm dialog" title="Figure 13. PLS-SEM algorithm dialog" srcset="https://substackcdn.com/image/fetch/$s_!dcJ2!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 424w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 848w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 1272w, https://substackcdn.com/image/fetch/$s_!dcJ2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb8ab0ffe-5789-41ce-8601-91bf179b17f6_900x500.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 13. The PLS-SEM algorithm dialog. Left, the PLS setup tab: (1) the three default settings, (3) the Start calculation button. Right, the Data tab: (2) the missing value treatment.</figcaption></figure></div><p><strong>Step 2.</strong> Click <strong>Start calculation</strong>. With <strong>Open report</strong> ticked, the report opens right away (Figure 14).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!_9s4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!_9s4!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!_9s4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png" width="1000" height="571" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:571,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:170188,&quot;alt&quot;:&quot;Figure 14. PLS-SEM algorithm report&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 14. PLS-SEM algorithm report" title="Figure 14. PLS-SEM algorithm report" srcset="https://substackcdn.com/image/fetch/$s_!_9s4!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 424w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 848w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 1272w, https://substackcdn.com/image/fetch/$s_!_9s4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbcf32969-d231-4fab-a6f5-c8be3f5369a1_1000x571.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 14. Report of the PLS-SEM algorithm: (1) the export buttons, (2) the list of results, (3) the line "Calculated for Complete data set", (4) the path coefficient of H1, (5) the R&#178; of repurchase intention.</figcaption></figure></div><p>The report screen contains:</p><ol><li><p><strong>Toolbar:</strong> export the report to Excel or HTML, save the report (Save report), create a data file from the results (Create data file) and the Compare button.</p></li><li><p><strong>List of results:</strong> Final results (path coefficients, loadings, weights and more) and Quality criteria (R&#178;, f&#178;, reliability, discriminant validity, VIF and more). Later posts read these items one by one.</p></li><li><p><strong>The line Calculated for Complete data set:</strong> the algorithm ran on the complete cases, as Casewise deletion implies.</p></li><li><p><strong>Path coefficients</strong> on the arrows and <strong>outer loadings</strong> on the lines to the items.</p></li><li><p><strong>R&#178;</strong> inside the circles of the endogenous constructs.</p></li></ol><p>To confirm that the model is drawn correctly, compare the results with numbers computed independently with cSEM on the same data:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/mcCL1/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f05a0bde-3504-4c19-9448-b725a9db666e_1220x1140.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1e05e9c6-ff2c-430b-8df4-d9b978126fd6_1220x1248.png&quot;,&quot;height&quot;:558,&quot;title&quot;:&quot;Table 3. SmartPLS results compared with cSEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/mcCL1/1/" width="730" height="558" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The two sets of numbers agree to the third decimal, and they agree with the PLS-SEM column of Post 31. The loadings of the 20 items also agree (for example CONF1 0.803, PV3 0.876, RI3 0.761). The model is drawn correctly.</p><h2>7. Choose a missing value treatment</h2><p>SmartPLS offers four options in the <strong>Missing value algorithm</strong> box (SmartPLS GmbH, n.d.-b):</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/e3Z1N/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6f788301-08a9-4b43-ac9e-c0c86335c135_1220x760.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b30821bd-6431-4b22-a814-839323f944b0_1220x830.png&quot;,&quot;height&quot;:380,&quot;title&quot;:&quot;Table 4. Missing value treatments in SmartPLS 4&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/e3Z1N/1/" width="730" height="380" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The SmartPLS guide recommends Mean replacement only when each item has less than 5% blank cells and the blanks occur at random; if an item or a respondent has more than 15% blank cells, consider removing it (SmartPLS GmbH, n.d.-b). The Marketplace X data have at most 1.12% blank cells in one item, so both options are acceptable. Figure 15 compares the results of the two.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!BCCH!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!BCCH!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!BCCH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png" width="1900" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:118212,&quot;alt&quot;:&quot;Figure 15. Path coefficients under two missing value treatments&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 15. Path coefficients under two missing value treatments" title="Figure 15. Path coefficients under two missing value treatments" srcset="https://substackcdn.com/image/fetch/$s_!BCCH!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!BCCH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F89e16afd-1486-4f4a-b272-9e074a5971cb_1900x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 15. Path coefficients under Casewise deletion (254 respondents) and Mean replacement (267 respondents), computed with cSEM. The largest difference is H6: 0.221 versus 0.201.</figcaption></figure></div><p>This series uses <strong>Casewise deletion</strong> for three reasons:</p><ul><li><p><strong>No values are filled in for the respondents:</strong> Mean replacement puts answers into the data that nobody gave, and slightly reduces the variance.</p></li><li><p><strong>Only 13 respondents (4.9%) are lost,</strong> leaving a sample of 254, still above the 250 computed in <a href="https://www.phuonglib.com/p/sample-size-calculation">Post 4, "How Large Should Your Sample Be? Formulas and Rules of Thumb for a Thesis"</a>.</p></li><li><p><strong>The results can be cross-checked:</strong> they match Post 31 (JASP and cSEM, which require complete data).</p></li></ul><p>One technical detail was checked: <strong>SmartPLS Casewise deletion drops only respondents with blanks in the items used in the model,</strong> not in the other columns of the file. The Marketplace X file has one respondent (ID 59) with a blank age group who answered all 20 items. If this respondent were also dropped, the sample would be 253, H1 would be 0.500 and the R&#178; of intention 0.317, which differs from SmartPLS. The SmartPLS results match the sample of 254, so this respondent is kept (script <code>compare_pls.R</code>).</p><p>In the AMOS posts, the model was estimated with FIML on all 267 respondents (Post 26). SmartPLS has no FIML. When you write your thesis, state the missing value treatment and the <strong>number of observations actually used</strong> (here 254), not the original sample size.</p><h2>Five common mistakes</h2><ol><li><p><strong>Not declaring the missing value code.</strong> SmartPLS treats -99 as a real answer. Sign: Missing equals 0 and Min equals -99 in the import dialog (Figure 4).</p></li><li><p><strong>Importing an SPSS or Excel file, or a file with more than 100 rows, with the Student license.</strong> SmartPLS shows <em>Feature locked</em> or marks the file as <em>not licensed</em>. Use a CSV file, or the trial or Professional license.</p></li><li><p><strong>Naming a construct after one of its items.</strong> Use names that differ from the item names (for example CONFIRM, not CONF1), so that the later posts match.</p></li><li><p><strong>Connecting a path in the wrong direction.</strong> Check each arrow against the hypothesis table before running.</p></li><li><p><strong>Forgetting to save the model.</strong> Click Save after drawing; once saved, the Save button turns pale.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Do I need to delete unused columns (ID, gender, N_MISSING and so on) before importing?</strong></p><p>No. SmartPLS uses only the items assigned to the model, and Casewise deletion only looks at those items (Section 7). Keeping the demographic variables also helps with multigroup analysis later.</p><p><strong>Should blank cells stay empty or be coded as -99?</strong></p><p>Both work (SmartPLS GmbH, n.d.-b). Coding -99 separates deliberately skipped answers from cells lost while copying data, and it is easy to check by eye. What matters is declaring the right code in the import dialog.</p><p><strong>What if my Excel file uses a decimal comma?</strong></p><p>Choose the matching Locale in the import dialog, or save the file again with a decimal point. Check the Min, Max and Mean columns after importing to make sure the numbers were read correctly.</p><p><strong>Can a drawn model be reused with other data?</strong></p><p>Yes. A project can hold several data files and several models. The small line under the word Indicators shows the data file the model is using (Figure 8).</p><div><hr></div><h2>References</h2><p>Rademaker, M. E., &amp; Schuberth, F. (2026). <em>cSEM: Composite-based structural equation modeling</em> (R package version 0.7.1) [Computer software]. https://CRAN.R-project.org/package=cSEM</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>SmartPLS GmbH. (n.d.-a). <em>Import a data file (raw data or covariance matrix)</em>. Retrieved November 6, 2026, from https://www.smartpls.com/documentation/tutorials/import-datafiles/</p><p>SmartPLS GmbH. (n.d.-b). <em>Missing values in SmartPLS</em>. Retrieved November 6, 2026, from https://www.smartpls.com/documentation/functionalities/missing-values/</p><p>SmartPLS GmbH. (n.d.-c). <em>The PLS-SEM algorithm: How SmartPLS estimates models</em>. Retrieved November 6, 2026, from https://www.smartpls.com/documentation/algorithms-and-techniques/core-algorithm/pls/</p><p>SmartPLS GmbH. (n.d.-d). <em>Pricing and licenses</em>. Retrieved November 6, 2026, from https://smartpls.com/pricing/</p><p>SmartPLS GmbH. (n.d.-e). <em>Working with large data files in SmartPLS</em>. Retrieved November 6, 2026, from https://www.smartpls.com/faq/smartpls4/large-data-files/</p><p>SmartPLS GmbH. (n.d.-f). <em>Your first PLS path model</em>. Retrieved November 6, 2026, from https://www.smartpls.com/documentation/tutorials/first-pls-path-model/</p>]]></content:encoded></item><item><title><![CDATA[Mediation Analysis with JASP: Bootstrapping Specific Indirect Effects]]></title><description><![CDATA[Define specific indirect effects in free JASP, run a 2000-sample bootstrap, read percentile confidence intervals and classify the mediation after Zhao et al.]]></description><link>https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:23 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Qojj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Qojj!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!Qojj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:622276,&quot;alt&quot;:&quot;A relay track where an orange baton reaches the finish only through hand-offs between runners while a straight shortcut lane stays empty.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A relay track where an orange baton reaches the finish only through hand-offs between runners while a straight shortcut lane stays empty." title="A relay track where an orange baton reaches the finish only through hand-offs between runners while a straight shortcut lane stays empty." srcset="https://substackcdn.com/image/fetch/$s_!Qojj!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Qojj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F775ce769-dc75-47d6-8b26-80158c6e2d5f_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Software usually prints only the total indirect effect; define each specific indirect effect as a product of labeled paths.</p></li><li><p>Judge an indirect effect by its bootstrap confidence interval, not by the Sobel test or a bootstrap <em>p</em>-value.</p></li><li><p>All three indirect paths are significant while the direct effect is not (beta = 0.091): indirect-only mediation.</p></li><li><p>The path through perceived value sits close to the threshold; its lower bound moves with the seed, the missing-data method and the interval type.</p></li><li><p>Fix the method and the seed before looking at the results, and report both.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/hypothesis-testing-sem-jasp">Post 27, "Hypothesis Testing in SEM with JASP"</a> concluded that hypothesis H4 (expectation confirmation has a direct effect on repurchase intention) is not supported, while the total effect of expectation confirmation is still large. The next question is: <strong>which route</strong> does that effect travel along, and is each route significant? That is the subject of <strong>mediation analysis</strong>.</p><p>This post splits the effect of expectation confirmation on repurchase intention into three indirect paths, tests each path with <strong>bootstrapping</strong>, classifies the type of mediation and shows a lesson that is rarely mentioned: for an indirect effect close to the threshold, the conclusion can change with the way the confidence interval is computed.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch28-mediation-analysis-bootstrap-sem">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The screenshots come from JASP 0.99.0 (JASP Team, 2026) on macOS, continuing the SEM analysis of Post 27 on the data file <code>marketplace_x_sem.sav</code> (267 respondents, FIML). All numbers in Sections 4 and 5 are those JASP printed in this run (2000 bootstrap samples, percentile intervals, random seed 2026). The numbers were checked with the lavaan package for R (Rosseel, 2012), using the script <code>mediation_bootstrap.R</code> in the repository, also with 2000 bootstrap samples; Section 6 and the unstandardized coefficients come from that script and the text says so.</p></blockquote><h2>1. Direct, indirect and total effects</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!nhmk!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!nhmk!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 424w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 848w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 1272w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!nhmk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png" width="2100" height="1120" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1120,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:191518,&quot;alt&quot;:&quot;Figure 1. Three indirect paths&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Three indirect paths" title="Figure 1. Three indirect paths" srcset="https://substackcdn.com/image/fetch/$s_!nhmk!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 424w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 848w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 1272w, https://substackcdn.com/image/fetch/$s_!nhmk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd440efe5-8d9b-49ce-8c9a-0ffbd62f6d63_2100x1120.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Three indirect paths from Expectation confirmation to Repurchase intention in the Marketplace X model.</figcaption></figure></div><p>In the Marketplace X model, expectation confirmation (CONFIRM) reaches repurchase intention (INTENT) along four routes:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/3f4m3/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e90426de-5e73-41aa-8261-e11095a260d5_1220x748.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/37a5a893-a496-4d3f-ab86-f2730bd5390a_1220x868.png&quot;,&quot;height&quot;:382,&quot;title&quot;:&quot;Table 1. Direct, indirect and total effects of expectation confirmation on repurchase intention&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/3f4m3/1/" width="730" height="382" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>A <strong>specific indirect effect</strong> is the product of the coefficients along one path. Software usually prints only the <strong>total indirect effect</strong>; to know which path is significant, you have to define each product yourself.</p><h2>2. Why bootstrapping?</h2><p>The product of two coefficients is not normally distributed, even when each coefficient is: the distribution of the product is usually skewed. The Sobel test (the delta method) assumes normality and is therefore inaccurate, especially in small and medium samples (Preacher &amp; Hayes, 2008). <strong>Bootstrapping</strong> does not need this assumption: it draws 2000 samples with replacement, computes the product in each sample and takes the 2.5th and 97.5th percentiles as the confidence interval. An indirect effect is significant when <strong>the confidence interval does not contain 0</strong>.</p><p>There are two common types of bootstrap confidence interval:</p><ul><li><p><strong>Percentile</strong>: takes the percentiles directly. Simple, and recommended in recent work because it controls the Type I error rate better (Hayes, 2022).</p></li><li><p><strong>Bias-corrected (BC)</strong>: corrects for bias in the bootstrap distribution. It has more power but sometimes gives a Type I error rate above the nominal level (Fritz et al., 2012).</p></li></ul><p>Do not rely on the bootstrap <em>p</em>-value when the software prints one (JASP and lavaan compute it from the bootstrap standard error with a normal approximation); base the conclusion on the confidence interval.</p><h2>3. Define the indirect effects and run the bootstrap in JASP</h2><p>Open the SEM analysis of Post 27 (labels h1 to h7 are already in place) and replace the defined parameter at the end of the syntax with the lines below. The complete model is:</p><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE =~ PV1 + PV2 + PV3 + PV4
SATISF =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT =~ RI1 + RI2 + RI3 + RI4
# Structural model with labels h1 to h7
VALUE ~ h1*CONFIRM
SATISF ~ h2*CONFIRM + h3*VALUE
INTENT ~ h4*CONFIRM + h5*VALUE + h6*SATISF + h7*TRUST
CONFIRM ~~ TRUST
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1
# Specific indirect effects
CONF_PV_RI := h1*h5
CONF_SAT_RI := h2*h6
CONF_PV_SAT_RI := h1*h3*h6
PV_SAT_RI := h3*h6
# Total indirect and total effects
CONF_RI_indirect := h1*h5 + h2*h6 + h1*h3*h6
CONF_RI_total := h4 + h1*h5 + h2*h6 + h1*h3*h6
diff_PV_SAT := h1*h5 - h2*h6</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!D7gO!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!D7gO!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 424w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 848w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 1272w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!D7gO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png" width="900" height="354" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:354,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:159817,&quot;alt&quot;:&quot;Figure 2. Syntax defining the indirect effects&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Syntax defining the indirect effects" title="Figure 2. Syntax defining the indirect effects" srcset="https://substackcdn.com/image/fetch/$s_!D7gO!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 424w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 848w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 1272w, https://substackcdn.com/image/fetch/$s_!D7gO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55a49bec-cf25-40e8-bd4c-fb3e315c31b6_900x354.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) The seven labeled paths, (2) the specific indirect effects, (3) the total indirect effect, the total effect and the difference between two indirect paths.</figcaption></figure></div><p><code>PV_SAT_RI</code> is the indirect effect of perceived value on repurchase intention through satisfaction (used in Section 4). The line <code>CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1</code> is needed with FIML (<a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a>). Press Cmd + Enter to apply the syntax.</p><p>Then turn on bootstrapping in <strong>Estimation Options</strong> (Figure 3):</p><ol><li><p><strong>Set</strong> <em>Standard errors</em> to <em>Bootstrap</em>.</p></li><li><p><strong>Enter</strong> 2000 in <em>Bootstrap samples</em> and <strong>set</strong> <em>Type</em> to <em>Percentile</em> (the JASP default is <em>Bias-corrected percentile</em>).</p></li><li><p><strong>Tick</strong> <em>Set a random seed</em> and enter any number (here 2026), so that a rerun gives the same results.</p></li><li><p><strong>Leave</strong> <em>Missing data handling</em> at <em>(FI)ML</em>.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!0AOo!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!0AOo!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 424w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 848w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 1272w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!0AOo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png" width="900" height="292" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:292,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:91691,&quot;alt&quot;:&quot;Figure 3. Bootstrap options in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Bootstrap options in JASP" title="Figure 3. Bootstrap options in JASP" srcset="https://substackcdn.com/image/fetch/$s_!0AOo!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 424w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 848w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 1272w, https://substackcdn.com/image/fetch/$s_!0AOo!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac14a3a0-ca18-46e5-b208-46ab885377c5_900x292.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. (1) Bootstrap standard errors, (2) number of samples and type of interval, (3) random seed, (4) FIML is kept for missing data.</figcaption></figure></div><p>JASP keeps FIML with all 267 respondents when bootstrapping, unlike AMOS, which needs the file of complete respondents. Run time depends on the computer: 2000 samples with this model take from a few minutes to more than half an hour on a slow machine. The <strong>SEM &gt; Mediation Analysis</strong> module of JASP can also test mediation, but only for observed variables (mean scores), not latent variables; that is why this post uses the SEM module.</p><p><strong>In AMOS</strong>, bootstrapping needs the complete-case file <code>marketplace_x_sem_complete.sav</code> (254 respondents) with <em>Estimate means and intercepts</em> unticked, because AMOS does not bootstrap data with empty cells. Turn it on in <strong>View &gt; Analysis Properties &gt; Bootstrap</strong>: tick <em>Perform bootstrap</em> with 2000 samples, <em>Percentile confidence intervals</em> and <em>Bias-corrected confidence intervals</em> (level 95); tick <em>Indirect, direct &amp; total effects</em> on the <strong>Output</strong> tab and set the seed on the <strong>Random #</strong> tab. After labeling the paths h1 to h7, click the line <em>Not estimating any user-defined estimand</em> in the lower-left corner of Amos Graphics, choose <em>Define new estimands</em> and type the same products with <code>=</code> instead of <code>:=</code> (for example <code>CONF_PV_RI = h1*h5</code>). After the run, <strong>View &gt; Text Output &gt; User-defined estimands</strong> with <em>Bootstrap Confidence</em> gives the percentile and bias-corrected intervals of each specific indirect effect; <em>Estimates &gt; Matrices &gt; Indirect Effects</em> gives only the total indirect effect. User-defined estimands are available from AMOS 24 onwards.</p><h2>4. Results</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!m57n!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!m57n!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 424w, https://substackcdn.com/image/fetch/$s_!m57n!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 848w, https://substackcdn.com/image/fetch/$s_!m57n!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 1272w, https://substackcdn.com/image/fetch/$s_!m57n!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!m57n!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png" width="2100" height="1260" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1260,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:171280,&quot;alt&quot;:&quot;Figure 4. Indirect effects and confidence intervals&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Indirect effects and confidence intervals" title="Figure 4. Indirect effects and confidence intervals" srcset="https://substackcdn.com/image/fetch/$s_!m57n!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 424w, https://substackcdn.com/image/fetch/$s_!m57n!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 848w, https://substackcdn.com/image/fetch/$s_!m57n!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 1272w, https://substackcdn.com/image/fetch/$s_!m57n!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb3027aa5-e98d-48fb-b340-9cded4a4ca25_2100x1260.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Standardized effects and 95% percentile bootstrap confidence intervals as printed by JASP (FIML, N = 267, 2000 samples).</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!EZdQ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!EZdQ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 424w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 848w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 1272w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!EZdQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png" width="900" height="334" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:334,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:114057,&quot;alt&quot;:&quot;Figure 5. The Defined Parameters table in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. The Defined Parameters table in JASP" title="Figure 5. The Defined Parameters table in JASP" srcset="https://substackcdn.com/image/fetch/$s_!EZdQ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 424w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 848w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 1272w, https://substackcdn.com/image/fetch/$s_!EZdQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fffb0142b-080f-4260-9264-b7ebb15b1f95_900x334.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. The Defined Parameters table after bootstrapping: (1) standardized estimates, (2) 95% percentile confidence intervals.</figcaption></figure></div><p>Because <em>Standardized estimates = All</em> is still ticked from Post 27, JASP prints the <strong>standardized</strong> effects with bootstrap confidence intervals of the standardized values. For the direct effect h4, read the <em>Regression Coefficients</em> table (Figure 6).</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/0t4k5/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/28ef5b58-d3b0-42b7-ac0c-71496e0bdc40_1220x1312.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0a8c672d-4796-407e-a7c8-35d9d68e7b7a_1220x1610.png&quot;,&quot;height&quot;:697,&quot;title&quot;:&quot;Table 2. Bootstrap results for the direct, indirect and total effects (JASP)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/0t4k5/1/" width="730" height="697" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!uJZ1!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!uJZ1!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 424w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 848w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 1272w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!uJZ1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png" width="900" height="304" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:304,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:97810,&quot;alt&quot;:&quot;Figure 6. The Regression Coefficients table in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. The Regression Coefficients table in JASP" title="Figure 6. The Regression Coefficients table in JASP" srcset="https://substackcdn.com/image/fetch/$s_!uJZ1!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 424w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 848w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 1272w, https://substackcdn.com/image/fetch/$s_!uJZ1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F30067118-ed28-4363-bbcc-b62b8f0b5a84_900x304.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. The Regression Coefficients table after bootstrapping: (1) the direct path h4, (2) its percentile confidence interval, which contains 0.</figcaption></figure></div><p>The unstandardized effects (B) are printed when <em>Standardized estimates</em> is unticked. lavaan gives B = 0.088 for the direct effect and 0.269 for the total indirect effect, out of a total effect of 0.357. On the 254 complete respondents that AMOS uses, the percentile intervals of the unstandardized effects are [&#8722;0.118, 0.262] for the direct effect, [0.003, 0.250] for indirect path 1, [0.019, 0.204] for indirect path 2, [0.013, 0.127] for indirect path 3 and [0.161, 0.426] for the total indirect effect. The two scales and both data sets lead to the same conclusion for every effect.</p><p>Three key points:</p><ol><li><p><strong>All three indirect paths are significant</strong>, and the total indirect effect (0.278) is about 75% of the total effect (0.369).</p></li><li><p><strong>The direct effect is not significant</strong> (beta = 0.091, <em>p</em> = 0.316, CI [&#8722;0.091, 0.268]). Following the classification in Section 5, this is <strong>indirect-only mediation</strong> (full mediation): expectation confirmation affects repurchase intention through perceived value and satisfaction.</p></li><li><p><strong>No path is significantly stronger than another.</strong> The difference between the path through perceived value and the path through satisfaction is 0.044, <em>p</em> = 0.624, and its confidence interval contains 0. As in Post 27, do not rank indirect paths by eye.</p></li></ol><p>For perceived value: its direct effect on repurchase intention is significant (h5: beta = 0.233, CI [0.033, 0.431] in this run) and its indirect effect through satisfaction is also significant, with the same positive sign. This is <strong>complementary mediation</strong> (partial mediation).</p><p>The bootstrap run changes the standard errors, <em>z</em>-values and <em>p</em>-values of the paths, not the estimates. For example, h4 has <em>p</em> = 0.316 here against 0.372 in the run without bootstrap (Post 27); h5 has <em>p</em> = 0.022 against 0.011. The fit (chi-square = 138.7, df = 162; CFI = 1.000; RMSEA = 0.000; SRMR = 0.035) is the same as in Post 26.</p><p><strong>A warning about the R-Squared table.</strong> In this bootstrap run, the <em>R-Squared</em> table of JASP 0.99.0 prints &#8722;0.179 for VALUE, 0.051 for SATISF and &#8722;0.241 for INTENT. These values cannot be correct (R&#178; cannot be negative), and the item rows above them look normal. Read R&#178; from the run without bootstrap (Post 26: 0.333, 0.514 and 0.443), or as 1 minus the standardized residual variance in the <em>Factor Variances</em> table of this run (0.667, 0.486 and 0.557, which give the same 0.333, 0.514 and 0.443).</p><h2>5. Types of mediation</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!T4oO!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!T4oO!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 424w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 848w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 1272w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!T4oO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png" width="2100" height="1080" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1080,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:164502,&quot;alt&quot;:&quot;Figure 7. Types of mediation&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Types of mediation" title="Figure 7. Types of mediation" srcset="https://substackcdn.com/image/fetch/$s_!T4oO!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 424w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 848w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 1272w, https://substackcdn.com/image/fetch/$s_!T4oO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0a0c0da0-08b9-4bb9-8f0d-ea78b5fdfec7_2100x1080.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Decision tree for classifying mediation after Zhao et al. (2010).</figcaption></figure></div><p>The older approach of Baron and Kenny (1986) requires a significant total effect before mediation is examined, and uses the Sobel test. Zhao et al. (2010) showed that this condition is unnecessary: what matters is <strong>whether the indirect effect is significant</strong>. The direct effect is then used to classify the mediation:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/a5qlm/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f1c20a33-39bc-4931-86c4-43fc1a6d011a_1220x710.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5c93a581-0233-4d8b-be81-3a18000500df_1220x780.png&quot;,&quot;height&quot;:353,&quot;title&quot;:&quot;Table 3. Types of mediation after Zhao et al. (2010)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/a5qlm/1/" width="730" height="353" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>A note on writing "full mediation": a non-significant direct effect <strong>does not prove</strong> that it is zero (Posts <a href="https://www.phuonglib.com/p/t-test-independent-paired-one-sample-pspp-spss">17</a> and 27). With a larger sample, a small direct effect could become significant. Many authors advise describing the result as "the effect works mainly through the mediators" rather than claiming "full" mediation (Hayes, 2022).</p><h2>6. When the conclusion depends on the method</h2><p>Indirect path 1 (through VALUE) has a lower bound close to 0. The table compares ways of computing its confidence interval. The first row is the JASP run of Section 4 (standardized); the other rows are unstandardized and come from <code>mediation_bootstrap.R</code>:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/0gZgr/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/67b77239-96fa-44b0-86c6-f2e5b79071f0_1220x742.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/33535a78-6441-46bd-a039-5f9dee54ad02_1220x862.png&quot;,&quot;height&quot;:403,&quot;title&quot;:&quot;Table 4. Confidence intervals of indirect path 1 under different methods&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/0gZgr/1/" width="730" height="403" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>In this run every method leads to the same conclusion, but the margin differs a great deal. On the 254 complete respondents (the data AMOS has to use), the percentile interval starts at only 0.003. An earlier run of the same analysis, which drew a different sequence of random bootstrap samples, gave [&#8722;0.001, 0.257] for this interval: it <strong>touches 0</strong> and would be read as not significant, while its bias-corrected interval [0.005, 0.262] did not. The coefficient h5 (VALUE to INTENT) behaves the same way: its percentile interval on the 254 respondents is [0.006, 0.385] in this run (the lower bound of 0.006 reported in Post 27) and was [&#8722;0.001, 0.398] in the earlier run. The result changes with the random seed, the handling of missing data and the type of interval: a sure sign of an effect that is <strong>close to the threshold</strong>.</p><p>How to handle it honestly:</p><ul><li><p><strong>Choose the method before looking at the results</strong> (for example FIML, percentile bootstrap, 2000 samples) and keep it.</p></li><li><p><strong>Report the random seed</strong> and the number of bootstrap samples so that others can reproduce the result.</p></li><li><p><strong>Do not switch to bias-corrected</strong> only because the percentile interval is not significant.</p></li><li><p><strong>Report this indirect path as weak evidence</strong>, together with a robustness analysis such as the table above. Paths 2 and 3 are significant with every method.</p></li></ul><h2>7. Reporting the results</h2><blockquote><p>Indirect effects were tested by bootstrapping with 2000 samples and 95% percentile confidence intervals (FIML, N = 267, random seed 2026). Expectation confirmation had a significant indirect effect on repurchase intention through perceived value (beta = 0.135, 95% CI [0.018, 0.264]), through satisfaction (beta = 0.090, [0.016, 0.175]) and through the chain of perceived value and satisfaction (beta = 0.053, [0.009, 0.107]). The total indirect effect was 0.278 ([0.163, 0.399]), while the direct effect was not significant (beta = 0.091, [&#8722;0.091, 0.268]), suggesting that the effect of expectation confirmation on repurchase intention works mainly through the two mediators (Zhao et al., 2010). The difference between the paths through perceived value and through satisfaction was not significant (0.044, [&#8722;0.131, 0.230]). The indirect path through perceived value was close to the threshold: its lower bound was 0.018 with FIML and only 0.003 (unstandardized, [0.003, 0.250]) when the analysis was restricted to the 254 complete respondents.</p></blockquote><h2>Five common mistakes</h2><ol><li><p><strong>Using the Sobel test</strong> as the main test of an indirect effect.</p></li><li><p><strong>Reporting only the total indirect effect</strong> when the model has several mediation paths.</p></li><li><p><strong>Applying the Baron and Kenny condition</strong>: stopping the analysis when the total effect is not significant.</p></li><li><p><strong>Changing the interval type or the seed</strong> until the result becomes significant.</p></li><li><p><strong>Interpreting mediation as causation</strong> with cross-sectional data. Mediation is a hypothesis about a mechanism; data from one point in time only show that the data are consistent with that mechanism.</p></li></ol><h2>Frequently asked questions</h2><p><strong>How many bootstrap samples are needed?</strong></p><p>2000 to 5000 is common. For a result close to the threshold, such as indirect path 1, 5000 or 10000 samples make the confidence interval more stable from run to run, but they do not turn a weak effect into a strong one.</p><p><strong>Does the total effect have to be significant before testing mediation?</strong></p><p>No (Zhao et al., 2010; Hayes, 2022). Competitive mediation can make the total effect close to 0 even when both the direct and the indirect effects are substantial.</p><p><strong>How is mediation with latent variables different from PROCESS in SPSS?</strong></p><p>PROCESS (Hayes, 2022) runs regressions on mean scores and ignores measurement error. SEM with latent variables corrects for that error, so the coefficients are usually larger (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a>). Both use the same bootstrap logic.</p><p><strong>How do I test moderated mediation?</strong></p><p>Moderated mediation means the indirect effect changes with a moderator. Use multigroup analysis (<a href="https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp">Post 29, "Multigroup Analysis and Moderation with JASP"</a>) when the moderator is a grouping variable, or an interaction term in PROCESS or SmartPLS (<a href="https://www.phuonglib.com/p/moderation-analysis-smartpls">Post 41, "Moderation Analysis in SmartPLS 4"</a>).</p><div><hr></div><h2>References</h2><p>Baron, R. M., &amp; Kenny, D. A. (1986). The moderator&#8211;mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. <em>Journal of Personality and Social Psychology, 51</em>(6), 1173&#8211;1182. https://doi.org/10.1037/0022-3514.51.6.1173</p><p>Fritz, M. S., Taylor, A. B., &amp; MacKinnon, D. P. (2012). Explanation of two anomalous results in statistical mediation analysis. <em>Multivariate Behavioral Research, 47</em>(1), 61&#8211;87. https://doi.org/10.1080/00273171.2012.640596</p><p>Hayes, A. F. (2022). <em>Introduction to mediation, moderation, and conditional process analysis: A regression-based approach</em> (3rd ed.). Guilford Press.</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Preacher, K. J., &amp; Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. <em>Behavior Research Methods, 40</em>(3), 879&#8211;891. https://doi.org/10.3758/BRM.40.3.879</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p><p>Zhao, X., Lynch, J. G., Jr., &amp; Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. <em>Journal of Consumer Research, 37</em>(2), 197&#8211;206. https://doi.org/10.1086/651257</p>]]></content:encoded></item><item><title><![CDATA[SEM Model Fit Indices in JASP: CMIN/df, GFI, CFI, TLI, RMSEA, SRMR]]></title><description><![CDATA[What chi-square, CFI, TLI, RMSEA and SRMR measure, where their cutoffs come from and how JASP compares three CFA models side by side]]></description><link>https://www.phuonglib.com/p/model-fit-indices-sem-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/model-fit-indices-sem-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:21 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!hdbc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!hdbc!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!hdbc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:851976,&quot;alt&quot;:&quot;Three candidate pieces held up to a gap in a large jigsaw board, the orange piece fitting exactly as small figures check it with instruments.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Three candidate pieces held up to a gap in a large jigsaw board, the orange piece fitting exactly as small figures check it with instruments." title="Three candidate pieces held up to a gap in a large jigsaw board, the orange piece fitting exactly as small figures check it with instruments." srcset="https://substackcdn.com/image/fetch/$s_!hdbc!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!hdbc!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11aec031-d087-416e-bca2-bddcaaf39fe7_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Fit indices fall into three groups (absolute, incremental and parsimony-adjusted); report several from different groups, not only those that look good.</p></li><li><p>The popular cutoffs come from the simulations of Hu and Bentler (1999) and hold only for models like those simulated.</p></li><li><p>Chi-square grows in proportion to sample size, while SRMR does not change.</p></li><li><p>A wrong model that merges perceived value and satisfaction still passes the CMIN/df, RMSEA and SRMR cutoffs; the chi-square difference test (166.52 with 4 df, p &lt; 0.001) exposes it.</p></li><li><p>JASP does not compute CMIN/df, and its SEM module compares several models in one analysis with an automatic difference test.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">Post 22, "Confirmatory Factor Analysis (CFA) in JASP"</a> ran a CFA for the five scales and noted that the model "fits the data". This post explains which indices that word "fits" rests on, what each index measures, where the usual cutoffs come from and why the cutoffs should not be used as a pass-or-fail test. To make this concrete, the post runs <strong>three measurement models</strong> on the same data, two of them deliberately wrong, and looks at how well the indices tell them apart.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch23-model-fit-indices-cfa">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>; folder <code>raw/</code> holds the JASP results as a PDF). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> Figures 6 and 7 are real screenshots of JASP 0.99.0 (JASP Team, 2026) on macOS, taken in November 2026 with the file <code>marketplace_x_sem.sav</code> prepared in <a href="https://www.phuonglib.com/p/prepare-connect-data-spss-to-jasp">Post 21, "Preparing and Connecting Data from SPSS to JASP"</a> (in the <code>data/</code> folder of the repository), continuing the analyses of Post 22. Figures 2, 3, 8 and 9 are real JASP output, exported with Export Results and cropped; nothing in them is changed. The numbers were checked with the R package lavaan (Rosseel, 2012) in the script <code>compare_model_fit.R</code> (folder <code>scripts/</code>). In AMOS, the same indices are in <strong>View &gt; Text Output &gt; Model Fit</strong> (read only the <em>Default model</em> row). SRMR is not in the Text Output: it comes from <strong>Plugins &gt; Standardized RMR</strong>, which, like the RMR, GFI table, works only on the 254 respondents who answered every item.</p></blockquote><h2>1. What is model fit?</h2><p>CB-SEM looks for the set of parameters that makes the covariance matrix <strong>reproduced by the model</strong> as close as possible to the covariance matrix <strong>of the data</strong> (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a>). Model fit measures how close they are. A well-fitting model means that its constraints (which items do not load on which factor, which errors are uncorrelated) do not clearly contradict the data.</p><p>Two things model fit does <strong>not</strong> tell you (Kline, 2023):</p><ul><li><p><strong>Whether the model is true.</strong> Many different models can fit the same data equally well.</p></li><li><p><strong>Whether the coefficients are large or significant.</strong> A model with very low loadings can still fit well if all the correlations between the items are low.</p></li></ul><h2>2. Three groups of indices</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Q0-q!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Q0-q!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Q0-q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png" width="2100" height="1000" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1000,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:174456,&quot;alt&quot;:&quot;Figure 1. Three groups of fit indices&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Three groups of fit indices" title="Figure 1. Three groups of fit indices" srcset="https://substackcdn.com/image/fetch/$s_!Q0-q!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!Q0-q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff594f701-0c4f-44cb-9fb5-a7e3e221f2c6_2100x1000.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. The three groups of fit indices and the reference cutoffs commonly used.</figcaption></figure></div><h3>Absolute fit indices</h3><ul><li><p><strong>Chi-square</strong> (AMOS calls it CMIN). Tests the null hypothesis H0: the model reproduces the population covariance matrix exactly. <em>p</em> &gt; 0.05 means H0 is not rejected. Chi-square grows in proportion to the sample size, so with samples of a few hundred respondents or more, almost every realistic model is rejected (section 5).</p></li><li><p><strong>CMIN/df</strong> (chi-square divided by the degrees of freedom). A way to reduce the dependence on sample size; the usual cutoff is below 3, and some authors accept up to 5. This index has no firm statistical basis and still grows with sample size (Kline, 2023).</p></li><li><p><strong>SRMR</strong> (standardized root mean square residual). The average difference between the observed correlations and the correlations reproduced by the model. Below 0.08 is considered good (Hu &amp; Bentler, 1999).</p></li><li><p><strong>GFI, AGFI</strong>. The proportion of variances and covariances explained by the model. Both are strongly affected by sample size and are rarely recommended today, but they still appear often in theses.</p></li></ul><h3>Incremental fit indices</h3><p>These compare your model with the <strong>independence model</strong> (also called the baseline model): a model that assumes no item is correlated with any other.</p><ul><li><p><strong>CFI</strong> (comparative fit index; Bentler, 1990). Ranges from 0 to 1; 0.90 or more is acceptable, 0.95 or more is good (Hu &amp; Bentler, 1999).</p></li><li><p><strong>TLI</strong> (Tucker-Lewis index, also called NNFI). Similar to CFI but penalizes complex models; it can exceed 1 when the chi-square is smaller than the degrees of freedom.</p></li><li><p><strong>NFI</strong>. An older version that is biased downward in small samples; use CFI instead.</p></li></ul><h3>Parsimony-adjusted indices</h3><ul><li><p><strong>RMSEA</strong> (root mean square error of approximation). Measures the misfit per degree of freedom, so it penalizes models with many parameters. Below 0.06 is good (Hu &amp; Bentler, 1999), below 0.08 is acceptable. <strong>Always read it with its 90% confidence interval</strong>: in small samples this interval is very wide.</p></li><li><p><strong>PCLOSE</strong> (in AMOS; JASP calls it <em>RMSEA p-value</em>). Tests H0: the population RMSEA is at most 0.05. PCLOSE &gt; 0.05 is a good sign.</p></li><li><p><strong>AIC, BIC</strong>. These have no cutoff; use them only to <strong>compare</strong> models on the same data: smaller is better.</p></li></ul><h3>Where do the cutoffs come from?</h3><p>The most popular cutoffs (CFI of 0.95 or more; RMSEA at most 0.06; SRMR at most 0.08) come from the simulations of Hu and Bentler (1999). The authors themselves and many later studies warn that these cutoffs hold only for the kinds of models simulated, and that they depend on sample size, the number of items and the size of the loadings (Marsh et al., 2004; McNeish &amp; Wolf, 2023). For example, with high loadings, a model with substantial misspecification can still have a low RMSEA. Therefore:</p><ul><li><p><strong>Report several indices</strong> from different groups, rather than picking the ones that look "good".</p></li><li><p><strong>Treat the cutoffs as a reference</strong>, together with the residuals, the modification indices and theory.</p></li></ul><h2>3. Finding the fit indices in JASP</h2><p>Open the CFA analysis of Post 22 again (the <em>Confirmatory Factor Analysis</em> of the <em>Factor</em> module, on <code>marketplace_x_sem.sav</code>). The results are in three tables: <em>Chi-square test</em>, <em>Fit indices</em> and <em>Other fit measures</em> (tick <em>Additional fit measures</em> in the <em>Additional Output</em> section).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!y1Ws!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!y1Ws!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 424w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 848w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 1272w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!y1Ws!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png" width="600" height="568" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:568,&quot;width&quot;:600,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:90200,&quot;alt&quot;:&quot;Figure 2. Chi-square, CFI and TLI in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Chi-square, CFI and TLI in JASP" title="Figure 2. Chi-square, CFI and TLI in JASP" srcset="https://substackcdn.com/image/fetch/$s_!y1Ws!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 424w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 848w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 1272w, https://substackcdn.com/image/fetch/$s_!y1Ws!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8cd21185-17fb-4df1-acdb-95eb536d0158_600x568.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) Chi-square of the five-factor model, (2) CFI and TLI.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!BURR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!BURR!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 424w, https://substackcdn.com/image/fetch/$s_!BURR!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 848w, https://substackcdn.com/image/fetch/$s_!BURR!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 1272w, https://substackcdn.com/image/fetch/$s_!BURR!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!BURR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png" width="330" height="918" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:918,&quot;width&quot;:330,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:112875,&quot;alt&quot;:&quot;Figure 3. RMSEA, SRMR and GFI in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. RMSEA, SRMR and GFI in JASP" title="Figure 3. RMSEA, SRMR and GFI in JASP" srcset="https://substackcdn.com/image/fetch/$s_!BURR!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 424w, https://substackcdn.com/image/fetch/$s_!BURR!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 848w, https://substackcdn.com/image/fetch/$s_!BURR!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 1272w, https://substackcdn.com/image/fetch/$s_!BURR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbfaaa162-dd9a-4302-9f1a-d96455b37873_330x918.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. (1) RMSEA and its 90% confidence interval, (2) SRMR, (3) GFI.</figcaption></figure></div><p>Three things to keep in mind when reading the JASP tables:</p><ul><li><p><strong>The Baseline model row</strong> in the <em>Chi-square test</em> table is the independence model (chi-square = 2301.467; df = 190), used to compute CFI and TLI. Read the <em>Factor model</em> row.</p></li><li><p><strong>JASP does not compute CMIN/df.</strong> Compute it by hand: 134.293 / 160 = 0.84.</p></li><li><p><strong>GFI under FIML</strong> (0.9924) is computed on a model that includes the means, so it is higher than GFI as AMOS computes it. AMOS reports GFI only when the data are complete; the corresponding value on the 254 respondents who answered every item is 0.950. When you report GFI, use the value on the complete data.</p></li></ul><h2>4. Results of the five-factor model</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/AGoS7/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d6610f0a-329b-490c-aff1-12c8018c14b8_1220x800.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/87d5f2cf-6e70-4931-aca9-8fddd6e4c1f6_1220x908.png&quot;,&quot;height&quot;:450,&quot;title&quot;:&quot;Table 1. Fit indices of the five-factor model.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/AGoS7/1/" width="730" height="450" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Every index is far beyond its cutoff: CMIN/df is below 1, TLI is greater than 1, RMSEA equals 0. Such a "perfect" result is rare with real survey data; here it happens because the data were simulated from this very five-factor structure (Post 19). When you get results this good on real data, check again: is the sample too small (a small chi-square because the test lacks power), or is some item duplicated?</p><h2>5. Chi-square depends on sample size</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!mliS!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!mliS!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 424w, https://substackcdn.com/image/fetch/$s_!mliS!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 848w, https://substackcdn.com/image/fetch/$s_!mliS!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 1272w, https://substackcdn.com/image/fetch/$s_!mliS!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!mliS!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png" width="2100" height="860" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:860,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:147715,&quot;alt&quot;:&quot;Figure 4. Chi-square by sample size&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Chi-square by sample size" title="Figure 4. Chi-square by sample size" srcset="https://substackcdn.com/image/fetch/$s_!mliS!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 424w, https://substackcdn.com/image/fetch/$s_!mliS!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 848w, https://substackcdn.com/image/fetch/$s_!mliS!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 1272w, https://substackcdn.com/image/fetch/$s_!mliS!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff48f4814-060f-4d49-b77f-92b9b3730712_2100x860.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. For the same degree of misfit between model and data, chi-square doubles when the sample size doubles.</figcaption></figure></div><p>Figure 4 uses the wrong model M4 (section 6) with the same sample covariance matrix, changing only the declared sample size. Chi-square grows exactly in proportion to <em>N</em> (295 to 591 and then 1182), while SRMR does not change (0.051) and RMSEA and CFI change much less. This is why, with large samples, the <em>p</em>-value of chi-square is almost always below 0.05 even when the model is only slightly wrong, and why other indices are used as well.</p><p>The reverse is also true: in a small sample, a chi-square with a large <em>p</em>-value does not necessarily mean the model is good; the test may simply lack power.</p><h2>6. Do the indices detect a wrong model?</h2><p>To check, the post runs two more, wrong, models on the same data:</p><ul><li><p><strong>M4</strong>: merges Perceived value (PV) and Satisfaction (SAT) into one eight-item factor. These two concepts have the highest correlation in the correct model (0.64, Post 22), so this is the easiest mistake to make.</p></li><li><p><strong>M1</strong>: all items measure a single factor.</p></li></ul><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!t6wi!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!t6wi!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 424w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 848w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 1272w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!t6wi!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png" width="2100" height="880" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:880,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:124262,&quot;alt&quot;:&quot;Figure 5. Comparison of three models&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Comparison of three models" title="Figure 5. Comparison of three models" srcset="https://substackcdn.com/image/fetch/$s_!t6wi!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 424w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 848w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 1272w, https://substackcdn.com/image/fetch/$s_!t6wi!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F69327bfa-9506-4445-a02a-29fba9ef48a4_2100x880.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. CFI, TLI, RMSEA and SRMR of the three models. The dashed lines are the cutoffs of Hu and Bentler (1999).</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/rxszd/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f128c2a3-c120-4f93-b55a-f772232db55f_1220x730.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f814c2f4-329a-42ef-a7f1-702883e39213_1220x866.png&quot;,&quot;height&quot;:414,&quot;title&quot;:&quot;Table 2. Fit indices of the three measurement models.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/rxszd/1/" width="730" height="414" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The most striking result is <strong>M4</strong>: a wrong model that <strong>passes</strong> many of the familiar cutoffs. CMIN/df = 1.83 (below 3), CFI = 0.935 (above 0.90), RMSEA = 0.056 (below 0.06), SRMR = 0.048 (below 0.08). Only CFI and TLI fall short of the "good" level of 0.95. A researcher who only checks cutoffs could conclude that M4 is "acceptable" and overlook the fact that PV and SAT are two different concepts.</p><p>What separates M4 from M5 is a <strong>direct comparison of the two models</strong>:</p><ul><li><p><strong>The chi-square difference test.</strong> M4 is nested in M5 (M4 is M5 with the VALUE&#8211;SATISF correlation fixed to 1), so the difference in chi-square follows a chi-square distribution: chi-square difference = 300.81 &#8722; 134.29 = 166.52; difference in degrees of freedom = 164 &#8722; 160 = 4; <em>p</em> &lt; 0.001. Merging the two factors makes the model clearly worse.</p></li><li><p><strong>The AIC</strong> of M5 is 159 units smaller than that of M4.</p></li><li><p><strong>The VALUE&#8211;SATISF correlation in M5</strong> is 0.64, with a 95% confidence interval [0.54, 0.73] far from 1, so the data do not support treating them as one concept.</p></li></ul><p>M1 is rejected by every index (CFI 0.64; RMSEA 0.13; SRMR 0.11). This comparison is also a rough check for <strong>common method variance</strong>: if the one-factor model fits almost as well as the multi-factor model, the answers may be driven by the way the questions were asked rather than by their content (<a href="https://www.phuonglib.com/p/common-method-bias-jasp-harman-clf">Post 30, "Common Method Bias (CMB) with JASP"</a>).</p><h3>Comparing models in JASP</h3><p>The <strong>SEM</strong> module lets you specify several models in one analysis and computes the chi-square difference test automatically:</p><ol><li><p><strong>Open</strong> the <em>Structural Equation Modeling</em> analysis of Post 22 (the five-factor model is already in the <em>Model 1</em> tab).</p></li><li><p><strong>Click</strong> the green plus button next to the tab to add <em>Model 2</em>; type model M4 and press Cmd+Enter (Ctrl+Enter on Windows) (Figure 6). Repeat for M1.</p></li><li><p><strong>Double-click</strong> a tab name to rename the tabs M5, M4 and M1.</p></li><li><p><strong>Choose</strong> <em>Missing data handling = Listwise deletion</em> in <em>Estimation Options</em> (Figure 7). Reason: with FIML, the SEM module of JASP 0.99.0 also estimates the factor means and reports the wrong degrees of freedom (Post 22). Listwise deletion uses the 254 respondents who answered every item, so the numbers differ slightly from the table above.</p></li><li><p><strong>Tick</strong> <em>Additional fit measures</em> in <em>Output Options</em> (Figure 7).</p></li></ol><p>The syntax of M4 (the merged factor is named VALSAT) and M1 (one factor named GENERAL):</p><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALSAT =~ PV1 + PV2 + PV3 + PV4 + SAT1 + SAT2 + SAT3 + SAT4
TRUST =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT =~ RI1 + RI2 + RI3 + RI4</code></pre><pre><code>GENERAL =~ CONF1 + CONF2 + CONF3 + CONF4 + PV1 + PV2 + PV3 + PV4 + SAT1 + SAT2 + SAT3 + SAT4 + TRU1 + TRU2 + TRU3 + TRU4R + RI1 + RI2 + RI3 + RI4</code></pre><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!kPXg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 424w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 848w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 1272w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!kPXg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png" width="760" height="195" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:195,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:64618,&quot;alt&quot;:&quot;Figure 6. Model M4 in the second tab&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Model M4 in the second tab" title="Figure 6. Model M4 in the second tab" srcset="https://substackcdn.com/image/fetch/$s_!kPXg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 424w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 848w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 1272w, https://substackcdn.com/image/fetch/$s_!kPXg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F15349fdf-d403-42ad-b1ec-1cd812920434_760x195.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 6. (1) The M4 model tab, (2) the button that adds a model, (3) the factor VALSAT with the eight items of PV and SAT.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!h7-Q!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!h7-Q!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 424w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 848w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 1272w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!h7-Q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png" width="760" height="296" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:296,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:75835,&quot;alt&quot;:&quot;Figure 7. Estimation and output options for the comparison&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Estimation and output options for the comparison" title="Figure 7. Estimation and output options for the comparison" srcset="https://substackcdn.com/image/fetch/$s_!h7-Q!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 424w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 848w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 1272w, https://substackcdn.com/image/fetch/$s_!h7-Q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc95c05a2-bb8e-4f78-83ad-6b9aad1a6c64_760x296.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. (1) Missing data handling set to Listwise deletion, (2) Additional fit measures.</figcaption></figure></div><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!VEtW!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 424w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 848w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 1272w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!VEtW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png" width="1000" height="203" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:203,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:78843,&quot;alt&quot;:&quot;Figure 8. The Model Fit table comparing the three models&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. The Model Fit table comparing the three models" title="Figure 8. The Model Fit table comparing the three models" srcset="https://substackcdn.com/image/fetch/$s_!VEtW!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 424w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 848w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 1272w, https://substackcdn.com/image/fetch/$s_!VEtW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F132ba90a-3d65-4989-89d2-e4e5fd9b0253_1000x203.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 8. (1) Chi-square, df and p of each model, (2) the difference test against the model in the row above.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!wcDC!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!wcDC!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 424w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 848w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 1272w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!wcDC!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png" width="560" height="952" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:952,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:162208,&quot;alt&quot;:&quot;Figure 9. Fit indices of the three models&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Fit indices of the three models" title="Figure 9. Fit indices of the three models" srcset="https://substackcdn.com/image/fetch/$s_!wcDC!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 424w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 848w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 1272w, https://substackcdn.com/image/fetch/$s_!wcDC!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d2e59bb-d116-4f01-8f99-4fb176499922_560x952.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. (1) CFI and TLI, (2) RMSEA, its confidence interval, RMSEA p-value and SRMR of M5, M4 and M1.</figcaption></figure></div><p>With the 254 complete respondents (the note under the table says that 13 cases were removed because of missing values): chi-square difference = 158.4 (4 degrees of freedom) when M4 is compared with M5, and chi-square difference = 630.9 (6 degrees of freedom) when M1 is compared with M4, both <em>p</em> &lt; 0.001. The indices tell the same story as with FIML: CFI = 1.000, 0.9362 and 0.6330; RMSEA = 0.000, 0.05617 and 0.1323; SRMR = 0.03512, 0.05103 and 0.1166 for M5, M4 and M1. JASP prints the RMSEA p-value of M1 in scientific notation (8.216 &#215; 10 to the power &#8722;15, practically 0). Note that JASP names the column <em>Baseline test</em> for the chi-square test of each model, which is different from the meaning of "baseline model" in the table of the CFA module. The <em>Difference test</em> column compares each model with the model directly above it, so the order of the tabs matters: put the most complex model in the first tab. Below the <em>Fit Indices</em> table, JASP 0.99.0 also prints a <em>T-Size Fit Indices</em> table, which this post does not use.</p><h2>7. When the fit is not good enough</h2><p>With real data, a CFI below 0.90 or an RMSEA above 0.08 can happen. A sensible order of work (Brown, 2015; Kline, 2023):</p><ol><li><p><strong>Check the data and the model specification.</strong> Have the reverse-worded items been recoded, has an item been put on the wrong factor, how were the missing data handled?</p></li><li><p><strong>Look at the standardized residuals and the modification indices</strong> (Post 22) to see where the misfit is concentrated: one item, one pair of items, or a whole factor.</p></li><li><p><strong>Modify only with a theoretical reason.</strong> For example, two items with very similar wording may be allowed to have correlated errors. Dropping an item with a low loading and a cross-loading is also a reasonable choice if it does not remove content.</p></li><li><p><strong>Report every change</strong> and the fit before and after the change.</p></li></ol><p>Do not keep modifying the model by the largest modification index until it passes the cutoffs: the final model will fit the peculiarities of the sample and will be hard to replicate (MacCallum et al., 1992).</p><h2>8. Reporting the results</h2><blockquote><p>The five-factor measurement model fitted the data: chi-square = 134.29; df = 160; <em>p</em> = 0.931; CMIN/df = 0.84; CFI = 1.000; TLI = 1.014; RMSEA = 0.000 (90% CI [0.000, 0.008]); SRMR = 0.032. An alternative model that merged Perceived value and Satisfaction into one factor fitted significantly worse, chi-square difference = 166.52 with 4 degrees of freedom; <em>p</em> &lt; 0.001; CFI = 0.935; RMSEA = 0.056; the one-factor model fitted poorly, CFI = 0.638; RMSEA = 0.130; SRMR = 0.109.</p></blockquote><p>At a minimum, report chi-square with df and <em>p</em>, CFI, TLI, RMSEA with its 90% confidence interval and SRMR. CMIN/df, GFI and AGFI can be added if your committee or journal asks for them.</p><h2>Five common mistakes</h2><ol><li><p><strong>Reporting only the indices that pass the cutoffs</strong> and leaving out those that do not.</p></li><li><p><strong>Concluding that the model is "true" because it passes the cutoffs.</strong> Model M4 is wrong but passes many of them.</p></li><li><p><strong>Ignoring the confidence interval of RMSEA</strong>, especially with a small sample or few degrees of freedom.</p></li><li><p><strong>Using GFI and AGFI as the main criteria.</strong> Both depend heavily on sample size.</p></li><li><p><strong>Modifying the model by the modification indices to pass the cutoffs</strong> without a theoretical reason.</p></li></ol><h2>Frequently asked questions</h2><p><strong>If chi-square is significant (p &lt; 0.05), is the model rejected?</strong></p><p>Statistically, yes, but with samples of a few hundred respondents, a very small misfit is enough to make <em>p</em> &lt; 0.05. In practice, the model is also judged by CFI, RMSEA and SRMR, and the chi-square is still reported. Kline (2023) advises always looking at the residuals to see whether the misfit is substantial, rather than ignoring the chi-square.</p><p><strong>Is a TLI greater than 1 an error?</strong></p><p>No. TLI exceeds 1 when the chi-square is smaller than the degrees of freedom, which is common with a model that fits very well or with a small sample. Report the value as printed.</p><p><strong>Is a CMIN/df below 1 good?</strong></p><p>It does not mean "the smaller the better". A CMIN/df below 1 means that the chi-square is smaller than expected if the model were true; this can come from a model that fits very well, a small sample or a model with too many parameters.</p><p><strong>Does the structural model need to fit as well as the measurement model?</strong></p><p>Yes. The structural model (<a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a>) usually has fewer paths than the CFA model (the covariances between factors are replaced by a limited set of effects), so its fit cannot be better than that of the CFA. Comparing the two models shows how much fit the structural constraints cost.</p><div><hr></div><h2>References</h2><p>Bentler, P. M. (1990). Comparative fit indexes in structural models. <em>Psychological Bulletin, 107</em>(2), 238&#8211;246. https://doi.org/10.1037/0033-2909.107.2.238</p><p>Brown, T. A. (2015). <em>Confirmatory factor analysis for applied research</em> (2nd ed.). Guilford Press.</p><p>Hu, L., &amp; Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. <em>Structural Equation Modeling: A Multidisciplinary Journal, 6</em>(1), 1&#8211;55. https://doi.org/10.1080/10705519909540118</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Kline, R. B. (2023). <em>Principles and practice of structural equation modeling</em> (5th ed.). Guilford Press.</p><p>MacCallum, R. C., Roznowski, M., &amp; Necowitz, L. B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. <em>Psychological Bulletin, 111</em>(3), 490&#8211;504. https://doi.org/10.1037/0033-2909.111.3.490</p><p>Marsh, H. W., Hau, K.-T., &amp; Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler's (1999) findings. <em>Structural Equation Modeling: A Multidisciplinary Journal, 11</em>(3), 320&#8211;341. https://doi.org/10.1207/s15328007sem1103_2</p><p>McNeish, D., &amp; Wolf, M. G. (2023). Dynamic fit index cutoffs for confirmatory factor analysis models. <em>Psychological Methods, 28</em>(1), 61&#8211;88. https://doi.org/10.1037/met0000425</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p>]]></content:encoded></item><item><title><![CDATA[Discriminant Validity in SmartPLS 4 with HTMT and Fornell-Larcker]]></title><description><![CDATA[Read the HTMT matrix against the 0.85 and 0.90 thresholds with bootstrap intervals, and see why Fornell-Larcker and cross loadings rarely fail in PLS-SEM]]></description><link>https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls</link><guid isPermaLink="false">https://www.phuonglib.com/p/htmt-discriminant-validity-smartpls</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:19 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Us3G!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Us3G!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!Us3G!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:916400,&quot;alt&quot;:&quot;Two similar islands separated by a clear channel of water, a small rowing boat with an orange sail measuring the gap between them.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Two similar islands separated by a clear channel of water, a small rowing boat with an orange sail measuring the gap between them." title="Two similar islands separated by a clear channel of water, a small rowing boat with an orange sail measuring the gap between them." srcset="https://substackcdn.com/image/fetch/$s_!Us3G!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!Us3G!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F489f8c9d-e6b4-4a6b-99e2-0f620fa82b59_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>In PLS-SEM, HTMT is the main criterion for discriminant validity; the Fornell-Larcker criterion and cross loadings are supplementary information.</p></li><li><p>HTMT should be below 0.85 for clearly distinct concepts and below 0.90 for concepts close in theory; a stricter test requires the 95th bootstrap percentile to stay below the threshold.</p></li><li><p>On the Marketplace X data the highest HTMT is 0.661 (Value and Satisfaction), with a 95th percentile of 0.738, so every pair has discriminant validity.</p></li><li><p>The Fornell-Larcker criterion is easy to pass in PLS-SEM, because the square root of the AVE is inflated and the latent correlations are deflated.</p></li><li><p>When HTMT exceeds the threshold, check the theory and the item content before merging constructs; do not drop items just to push HTMT down.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/evaluate-reflective-measurement-model-smartpls">Post 34</a> showed that the five scales of the Marketplace X model are reliable and have convergent validity. The last step in assessing the measurement model is <strong>discriminant validity</strong>: does each scale measure a concept that is <strong>different</strong> from the other scales? For the Marketplace X study, the pair that raises the most concern is Perceived value and Satisfaction, two concepts that are closely related in theory.</p><p><a href="https://www.phuonglib.com/p/convergent-discriminant-validity-cfa-jasp">Post 25</a> introduced HTMT and the Fornell-Larcker criterion in CB-SEM. This post focuses on how to read the three discriminant validity tables in SmartPLS, and on why, in PLS-SEM, <strong>HTMT is the main criterion</strong>, while the Fornell-Larcker criterion and cross loadings serve only as supplementary information.</p><blockquote><p><strong>Sample data and scripts.</strong> The data file and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch35-htmt-discriminant-validity">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> All figures are real screenshots of the SmartPLS 4.1.2.1 report (Ringle et al., 2026; 30-day trial), taken in November 2026, for the model <code>marketplace-x-model</code> of <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33</a>, with casewise deletion, 254 respondents. Every number was cross-checked in R with the script <code>discriminant_validity.R</code> (in the repository's <code>scripts/</code> folder) and matches SmartPLS to the third decimal place. The bootstrap confidence intervals of HTMT were computed in R with 5,000 samples. SmartPLS lists the constructs alphabetically (CONFIRM, INTENT, SATISF, TRUST, VALUE), so the rows and columns of its tables follow that order.</p></blockquote><h2>1. Open the discriminant validity tables</h2><p>In the report of the PLS-SEM algorithm (Post 33, Section 6), expand the <strong>Quality criteria</strong> group, then click the arrow next to <strong>Discriminant validity</strong>. This entry has six options:</p><ul><li><p><strong>Heterotrait-monotrait ratio (HTMT)</strong>, shown in three forms: Matrix, List (one row per pair) and Bar chart.</p></li><li><p><strong>Congruence coefficient</strong>: another index, not used in this series.</p></li><li><p><strong>Fornell-Larcker criterion</strong>.</p></li><li><p><strong>Cross loadings</strong>.</p></li></ul><h2>2. HTMT: the main criterion</h2><p><strong>HTMT</strong> (Henseler et al., 2015) compares the correlations between items <strong>of two different scales</strong> with the correlations between items <strong>within the same scale</strong>. An HTMT close to 1 means that the items of the two scales correlate almost as strongly with each other as items within one scale: the two scales are hard to tell apart. Post 25 explains the calculation in detail.</p><p><strong>Step 1.</strong> Select <strong>Heterotrait-monotrait ratio (HTMT) - Matrix</strong> (Figure 1).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!BKHF!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!BKHF!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!BKHF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png" width="1000" height="838" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:838,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:200491,&quot;alt&quot;:&quot;Figure 1. HTMT matrix in SmartPLS&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. HTMT matrix in SmartPLS" title="Figure 1. HTMT matrix in SmartPLS" srcset="https://substackcdn.com/image/fetch/$s_!BKHF!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!BKHF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40c20c8e-6ac6-407d-a8ab-c6eede13c9c8_1000x838.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. HTMT matrix: (1) the HTMT - Matrix entry, (2) HTMT of the ten construct pairs in the lower half of the matrix, (3) the highest pair, Value and Satisfaction (row VALUE, column SATISF), (4) the other tables of the Discriminant validity entry.</figcaption></figure></div><p><strong>Thresholds</strong> (Henseler et al., 2015):</p><ul><li><p><strong>Below 0.85</strong> for concepts that are clearly distinct in theory.</p></li><li><p><strong>Below 0.90</strong> for concepts that are close in theory, such as Perceived value and Satisfaction.</p></li></ul><p>A stricter approach tests whether HTMT is <strong>significantly lower than the threshold</strong>, using a bootstrap confidence interval: the upper bound of the one-sided 95% confidence interval (the 95th percentile) must be lower than the threshold (Franke &amp; Sarstedt, 2019; Hair et al., 2019).</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/WW8rA/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8ccb298f-82df-4c75-bb1a-c5112f07fad9_1220x1176.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8471ed22-e901-4c7f-87d7-42784e155a93_1220x1334.png&quot;,&quot;height&quot;:586,&quot;title&quot;:&quot;Table 1. HTMT of the ten construct pairs with bootstrap confidence intervals&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/WW8rA/1/" width="730" height="586" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!wv9c!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!wv9c!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!wv9c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png" width="1900" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:1900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:120204,&quot;alt&quot;:&quot;Figure 2. HTMT and bootstrap confidence intervals&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. HTMT and bootstrap confidence intervals" title="Figure 2. HTMT and bootstrap confidence intervals" srcset="https://substackcdn.com/image/fetch/$s_!wv9c!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 424w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 848w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 1272w, https://substackcdn.com/image/fetch/$s_!wv9c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70597f20-1701-492a-aeaf-2843b01d49ab_1900x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. HTMT of the ten construct pairs (dots) and 95% bootstrap confidence intervals (horizontal lines). The dashed lines mark the thresholds 0.85 and 0.90.</figcaption></figure></div><p><strong>Result:</strong> the highest HTMT is 0.661 (Value and Satisfaction), with a 95th percentile of 0.738, both far below the 0.85 threshold. <strong>Every construct pair has discriminant validity.</strong> The result is close to Post 25 (0.653 for the same pair); the small difference arises because Post 25 used 267 respondents with pairwise correlations.</p><p>HTMT depends only on the correlations between the items, not on how the model is estimated. HTMT in SmartPLS, JASP or R therefore gives the same result on the same data, and it is not biased the way AVE is (Post 34).</p><p><a href="https://www.phuonglib.com/p/bootstrapping-smartpls-path-significance">Post 39</a> explains how bootstrapping works inside SmartPLS and how to read its confidence intervals.</p><h2>3. The Fornell-Larcker criterion</h2><p>Fornell and Larcker (1981): a construct has discriminant validity when <strong>the square root of its AVE is greater than its correlation with every other construct</strong>.</p><p><strong>Step 2.</strong> Select <strong>Fornell-Larcker criterion</strong> (Figure 3). The diagonal shows the square root of the AVE; the other cells are the correlations between two latent variables.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!IkgB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!IkgB!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!IkgB!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png" width="1000" height="838" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:838,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:199122,&quot;alt&quot;:&quot;Figure 3. Fornell-Larcker table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Fornell-Larcker table" title="Figure 3. Fornell-Larcker table" srcset="https://substackcdn.com/image/fetch/$s_!IkgB!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!IkgB!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4d1d846b-fbb9-4cea-b50c-88b61eac655d_1000x838.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. Fornell-Larcker criterion: (1) the Fornell-Larcker criterion entry, (2) the table with the square root of the AVE on the diagonal, (3) the highest correlation, between Value and Satisfaction.</figcaption></figure></div><p><strong>Result:</strong> the square roots of the AVE range from 0.782 (Intention) to 0.826 (Satisfaction), all greater than the highest correlation, 0.556 (Value and Satisfaction). The criterion is met.</p><p><strong>Why not rely on the Fornell-Larcker criterion in PLS-SEM?</strong> In PLS-SEM, this criterion is biased towards "easy to pass" from both ends:</p><ul><li><p><strong>The square root of the AVE is inflated</strong>, because the loadings are inflated (Post 34).</p></li><li><p><strong>The correlations between latent variables are deflated</strong>, because composites contain measurement error (<a href="https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem">Post 31</a>).</p></li></ul><p>The simulations of Henseler et al. (2015) show that the Fornell-Larcker criterion often <strong>fails to detect</strong> a lack of discriminant validity between two concepts. Hair et al. (2019) therefore recommend HTMT instead of the Fornell-Larcker criterion. Many thesis committees still ask for the Fornell-Larcker table; you can report it alongside, but base your conclusion on HTMT.</p><h2>4. Cross loadings</h2><p><strong>Cross loadings</strong> are the correlations of each item with the scores of <strong>every</strong> latent variable. The requirement: each item correlates <strong>more strongly</strong> with its own construct than with any other construct.</p><p><strong>Step 3.</strong> Select <strong>Cross loadings</strong> (Figure 4). If the table does not show all 20 rows, drag the zoom slider in the bottom-right corner to shrink the table.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!WEjM!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!WEjM!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!WEjM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png" width="1000" height="838" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:838,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:322558,&quot;alt&quot;:&quot;Figure 4. Cross loadings table&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Cross loadings table" title="Figure 4. Cross loadings table" srcset="https://substackcdn.com/image/fetch/$s_!WEjM!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 424w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 848w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 1272w, https://substackcdn.com/image/fetch/$s_!WEjM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F43ea8ba4-9aff-4d19-ac88-90f9a89374b5_1000x838.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Cross loadings: (1) the Cross loadings entry, (2) the correlations of the 20 items with the five latent variables (items in alphabetical order), (3) item PV4, with the smallest gap between its own loading and its cross loadings.</figcaption></figure></div><p><strong>Result:</strong> every item has its highest loading on its own construct. The smallest gap is item PV4: 0.769 on Value versus 0.471 on Satisfaction, a difference of 0.298. The criterion is met.</p><p>As with the Fornell-Larcker criterion, cross loadings very rarely detect discriminant validity problems in PLS-SEM (Henseler et al., 2015), so use them only as supplementary information: if an item has a higher cross loading than its own loading, that is a clear sign that the content of the item needs another look.</p><h2>5. What to do when HTMT exceeds the threshold</h2><p>The Marketplace X data do not have this problem, but when the HTMT of a pair exceeds 0.85 or 0.90, consider the following, in order:</p><ol><li><p><strong>Check the theory:</strong> are the two concepts really different, or are they two names for the same concept?</p></li><li><p><strong>Look at the cross loadings</strong> to find items that correlate highly with the other construct. Items with overlapping content are candidates for review.</p></li><li><p><strong>Consider merging the two constructs</strong> if theory allows it, then compare the models (<a href="https://www.phuonglib.com/p/model-fit-indices-sem-jasp">Post 23</a> tried this with Value and Satisfaction in CB-SEM).</p></li></ol><p>Do not drop items just to push HTMT below the threshold without considering their content.</p><h2>6. Reporting the results</h2><blockquote><p>Discriminant validity was assessed with the HTMT ratio (Henseler et al., 2015). The HTMT values of the ten construct pairs ranged from 0.133 to 0.661, all below the conservative threshold of 0.85. Bootstrap confidence intervals (5,000 samples) showed that the 95th percentile of the highest HTMT (Perceived value and Satisfaction) was 0.738, lower than 0.85 (Franke &amp; Sarstedt, 2019). The Fornell-Larcker criterion and the cross loadings led to the same conclusion. The scales have discriminant validity.</p></blockquote><p>The table should present the HTMT matrix (as in Figure 1), with a column of confidence intervals for the pairs with the highest HTMT.</p><h2>Four common mistakes</h2><ol><li><p><strong>Reporting only the Fornell-Larcker criterion</strong> in a study that uses PLS-SEM. This criterion is easy to pass and misses problems.</p></li><li><p><strong>Using the 0.90 threshold for every pair</strong> to "be safe". The 0.90 threshold is only for concepts that are close in theory; state the reason when you use it.</p></li><li><p><strong>Misreading the upper half of the matrix.</strong> SmartPLS fills in only the lower half; an empty cell does not mean zero.</p></li><li><p><strong>Using HTMT for formative scales.</strong> HTMT is for reflective scales (<a href="https://www.phuonglib.com/p/formative-measurement-model-smartpls">Post 36</a> discusses formative scales).</p></li></ol><h2>Frequently asked questions</h2><p><strong>Is HTMT in SmartPLS the same as in JASP (Post 25)?</strong></p><p>SmartPLS uses the original HTMT of Henseler et al. (2015). JASP uses HTMT2, a variant that uses the geometric mean instead of the arithmetic mean (Post 25). The two usually differ very little.</p><p><strong>Do I need to run bootstrapping to report HTMT?</strong></p><p>It is advisable. Comparing HTMT with the threshold is enough for a preliminary conclusion, but the confidence interval shows whether the conclusion holds up against sampling variation (Franke &amp; Sarstedt, 2019). For the Marketplace X data, both approaches give the same conclusion.</p><p><strong>Value and Satisfaction correlate at 0.556. Is that high?</strong></p><p>The two concepts are related in theory (perceived value increases satisfaction, hypothesis H3), so a moderate correlation is reasonable. What needs checking is whether the two scales measure <strong>the same thing</strong>, and an HTMT of 0.661 shows that they do not.</p><div><hr></div><h2>References</h2><p>Fornell, C., &amp; Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. <em>Journal of Marketing Research, 18</em>(1), 39&#8211;50. https://doi.org/10.2307/3151312</p><p>Franke, G., &amp; Sarstedt, M. (2019). Heuristics versus statistics in discriminant validity testing: A comparison of four procedures. <em>Internet Research, 29</em>(3), 430&#8211;447. https://doi.org/10.1108/IntR-12-2017-0515</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>Henseler, J., Ringle, C. M., &amp; Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. <em>Journal of the Academy of Marketing Science, 43</em>(1), 115&#8211;135. https://doi.org/10.1007/s11747-014-0403-8</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p>]]></content:encoded></item><item><title><![CDATA[Confirmatory Factor Analysis (CFA) in JASP: Output and Mardia's Test]]></title><description><![CDATA[Run a five-factor CFA in free JASP with FIML, read the loadings, correlations and modification indices, then test multivariate normality with Mardia's coefficient]]></description><link>https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:05 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!FkdE!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!FkdE!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!FkdE!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:772012,&quot;alt&quot;:&quot;A large translucent blueprint held over five stone pillars, its drawn lines matching the stones, one pillar glowing orange.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A large translucent blueprint held over five stone pillars, its drawn lines matching the stones, one pillar glowing orange." title="A large translucent blueprint held over five stone pillars, its drawn lines matching the stones, one pillar glowing orange." srcset="https://substackcdn.com/image/fetch/$s_!FkdE!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!FkdE!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42062d53-badb-42f0-970c-620b61f6e9ee_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>A CFA states in advance which items measure which factor; five factors with 20 items give 160 degrees of freedom.</p></li><li><p>In JASP, FIML keeps all 267 respondents, and the modification indices come from the same run.</p></li><li><p>All standardized loadings are significant and range from 0.65 to 0.85; the highest factor correlation is 0.64.</p></li><li><p>Small modification indices without a theoretical reason are no basis for changing the model.</p></li><li><p>Mardia's test is in the SEM module, not the Factor module; a critical ratio of 4.62 is below Bentler's cutoff of 5.</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/exploratory-factor-analysis-efa-pspp-spss">Post 12, "Exploratory Factor Analysis (EFA) in PSPP (SPSS)"</a> used EFA to <em>explore</em> how the 20 items group together and found the five factors of the design. <strong>Confirmatory factor analysis</strong> (CFA) works the other way round: you state in advance which items measure which concept, then test whether the data support that structure (Brown, 2015). CFA is a required step before SEM: if the measurement model does not hold up, conclusions about the hypotheses in the structural model cannot be trusted (Anderson &amp; Gerbing, 1988).</p><p>This post runs a CFA for the five scales of the Marketplace X model (20 items), reads the core results and carries out <strong>Mardia's test of multivariate normality</strong>, as promised in the FAQ of <a href="https://www.phuonglib.com/p/descriptive-statistics-spss-pspp">Post 9, "Descriptive Statistics in PSPP (SPSS)"</a>. The next three posts go deeper into each group of indices: model fit (<a href="https://www.phuonglib.com/p/model-fit-indices-sem-jasp">Post 23, "SEM Model Fit Indices in JASP"</a>), CR and AVE (<a href="https://www.phuonglib.com/p/composite-reliability-ave-cfa-jasp">Post 24, "Composite Reliability (CR) and Average Variance Extracted (AVE) in JASP"</a>) and discriminant validity (<a href="https://www.phuonglib.com/p/convergent-discriminant-validity-cfa-jasp">Post 25, "Convergent and Discriminant Validity in JASP"</a>).</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch22-confirmatory-factor-analysis-cfa">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> Figures 1, 2 and 10 to 12 are real screenshots of JASP 0.99.0 (JASP Team, 2026) on macOS, taken in November 2026 with the file <code>marketplace_x_sem.sav</code> prepared in <a href="https://www.phuonglib.com/p/prepare-connect-data-spss-to-jasp">Post 21, "Preparing and Connecting Data from SPSS to JASP"</a> (in the <code>data/</code> folder of the repository). Figures 5 to 8 and 13 are real JASP output, exported with Export Results and cropped; nothing in them is changed. The numbers were checked with the R package lavaan (Rosseel, 2012) in the script <code>cfa_5_factors.R</code> (folder <code>scripts/</code> of the repository), which also computes the results JASP does not provide (item skewness and kurtosis, Mahalanobis distances, standardized residuals and the Bollen-Stine bootstrap). In AMOS, the same CFA is drawn with <strong>Plugins &gt; Draw Covariances</strong> for the 10 factor covariances and run with <strong>Analyze &gt; Calculate Estimates</strong>. Modification indices, <em>Tests for normality and outliers</em> and the Bollen-Stine bootstrap do not run when the data have missing cells (Byrne, 2016), so AMOS users run these checks on the 254 complete respondents (<code>marketplace_x_sem_complete.sav</code>) and take the reported estimates from a run with <em>Estimate means and intercepts</em> on all 267.</p></blockquote><h2>1. The measurement model to be tested</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Y8ErK/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5ebb9981-9100-416e-b254-0897e1965922_1220x774.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/de26231f-ac08-46e8-9a83-a22f59967a64_1220x894.png&quot;,&quot;height&quot;:368,&quot;title&quot;:&quot;Table 1. The five factors of the measurement model and their items.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:false}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Y8ErK/1/" width="730" height="368" frameborder="0" scrolling="no"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>In a CFA, each item loads on <strong>only one</strong> factor, the measurement errors are uncorrelated and <strong>every pair of factors is allowed to correlate</strong> (10 covariances). The factor names differ from every column name in the data file (Posts <a href="https://www.phuonglib.com/p/first-model-jasp-free-amos-alternative">20</a> and 21).</p><p>With 20 items, the data contain 20 &#215; 21 / 2 = 210 variances and covariances. The model estimates 50 parameters: 15 loadings (the other 5 are fixed to 1), 20 error variances, 5 factor variances and 10 covariances between the factors. The degrees of freedom are 210 &#8722; 50 = <strong>160</strong>.</p><h2>2. Running the CFA in JASP</h2><h3>Specify the five factors</h3><ol><li><p><strong>Open</strong> the file <code>marketplace_x_sem.sav</code> (the items are already Scale variables after the preparation in Post 21) and choose <strong>Factor &gt; Confirmatory Factor Analysis</strong>.</p></li><li><p><strong>Move</strong> CONF1 to CONF4 into the <em>Factor 1</em> box, double-click the box title and rename it CONFIRM.</p></li><li><p><strong>Click</strong> the green plus button below the factor boxes to add a new factor; repeat for VALUE, SATISF, TRUST and INTENT (Figure 1).</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Z5rV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Z5rV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 424w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 848w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 1272w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Z5rV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png" width="700" height="683" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:683,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:110705,&quot;alt&quot;:&quot;Figure 1. Specifying the five factors in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Specifying the five factors in JASP" title="Figure 1. Specifying the five factors in JASP" srcset="https://substackcdn.com/image/fetch/$s_!Z5rV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 424w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 848w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 1272w, https://substackcdn.com/image/fetch/$s_!Z5rV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbd2f1586-5968-48f1-a28a-72bd3436fe21_700x683.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. (1) Factor name, (2) arrow button that moves items into the factor, (3) buttons that add or remove a factor.</figcaption></figure></div><p>By default, JASP lets every pair of factors correlate and fixes the first loading of each factor to 1 (section <em>Model Options</em>), as AMOS does. In the variable list on the left, items already assigned to a factor stay visible, because JASP allows an item to load on several factors; take care not to put one item into two factors by mistake.</p><h3>Choose the options</h3><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/lX6NU/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ff706926-73c5-48d3-8f8e-14705c72e3e6_1220x780.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4cbb4c92-2b06-4e63-abf5-7a650e0d25c4_1220x850.png&quot;,&quot;height&quot;:389,&quot;title&quot;:&quot;Table 2. CFA options to set in JASP.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/lX6NU/1/" width="730" height="389" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!qMHR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!qMHR!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 424w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 848w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 1272w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!qMHR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png" width="700" height="547" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:547,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:106220,&quot;alt&quot;:&quot;Figure 2. Output options of the CFA in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Output options of the CFA in JASP" title="Figure 2. Output options of the CFA in JASP" srcset="https://substackcdn.com/image/fetch/$s_!qMHR!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 424w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 848w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 1272w, https://substackcdn.com/image/fetch/$s_!qMHR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7f4d8057-d92b-470a-82a1-580c3d69ffef_700x547.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) Additional fit measures, (2) R-Squared, (3) Modification indices with the cutoff 3.84, (4) Model plot with standardized parameter estimates.</figcaption></figure></div><p>Unlike AMOS, JASP computes modification indices with FIML, so a single run with all 267 respondents is enough. The <em>Information criteria</em> table confirms that FIML is in use: it reports 70 free parameters, the 50 parameters above plus 20 item intercepts.</p><h2>3. Reading the output in six steps</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!tvAJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!tvAJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 424w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 848w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 1272w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!tvAJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png" width="2000" height="760" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:760,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:138505,&quot;alt&quot;:&quot;Figure 3. Six steps for reading CFA output&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Six steps for reading CFA output" title="Figure 3. Six steps for reading CFA output" srcset="https://substackcdn.com/image/fetch/$s_!tvAJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 424w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 848w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 1272w, https://substackcdn.com/image/fetch/$s_!tvAJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff6b3b202-a7a7-4bde-a2f3-49d577fc3c15_2000x760.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. Six steps for checking CFA output before assessing fit, reliability and discriminant validity.</figcaption></figure></div><h3>Step 1. Check that the model runs</h3><p>The <em>Chi-square test</em> table in the results pane on the right shows the chi-square with the note <em>The estimator is ML</em>. If lavaan fails to converge or produces improper estimates, JASP shows a red warning just below the title of the analysis.</p><p>With the sample data the model runs normally. Chi-square = 134.293; df = 160; <em>p</em> = 0.931; CFI = 1.000; RMSEA = 0.000; SRMR = 0.032. Post 23 analyzes these indices; here it is enough to note that the model fits the data. The near-perfect fit arises because the data were simulated from exactly this structure (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a>).</p><h3>Step 2. Check for improper estimates</h3><p>Look for two warning signs (Kline, 2023):</p><ul><li><p><strong>Negative variances</strong> of an error or a factor (a Heywood case).</p></li><li><p><strong>Standardized estimates greater than 1</strong> in absolute value.</p></li></ul><p>Neither occurs in the sample data: the smallest standardized error variance is 0.28 (PV3, <em>Residual variances</em> table).</p><h3>Step 3. Read the standardized loadings</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!AUMO!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!AUMO!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 424w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 848w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 1272w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!AUMO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png" width="2200" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/aa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2200,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:326355,&quot;alt&quot;:&quot;Figure 4. CFA diagram with standardized estimates&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. CFA diagram with standardized estimates" title="Figure 4. CFA diagram with standardized estimates" srcset="https://substackcdn.com/image/fetch/$s_!AUMO!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 424w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 848w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 1272w, https://substackcdn.com/image/fetch/$s_!AUMO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Faa6666eb-d762-48a5-8cc6-06fbf12f7542_2200x1440.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Standardized loadings, R&#178; of each item and the correlations between the five factors. The diagram is redrawn from the results for readability.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!0FUV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!0FUV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 424w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 848w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 1272w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!0FUV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png" width="760" height="593" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:593,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:197260,&quot;alt&quot;:&quot;Figure 5. Standardized factor loadings in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Standardized factor loadings in JASP" title="Figure 5. Standardized factor loadings in JASP" srcset="https://substackcdn.com/image/fetch/$s_!0FUV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 424w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 848w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 1272w, https://substackcdn.com/image/fetch/$s_!0FUV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d261f84-5201-4198-8f7a-15afd14dc4a0_760x593.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. The Factor loadings table with Standardization = All: the Std. estimate column holds the standardized loadings.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/kQY8Q/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6bd60d1b-eb92-4e10-9c42-39eba722cbb4_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f8da5986-a804-4e57-8585-033c38ef7014_1220x616.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 3. Standardized loadings and R&#178; of the 20 items.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/kQY8Q/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><ul><li><p><strong>Every loading is significant</strong> (<em>p</em> &lt; 0.001; the z-values in the standardized table range from 15.56 to 32.28).</p></li><li><p><strong>Every standardized loading is 0.65 or higher.</strong> The usual minimum is 0.5, and 0.7 or higher is ideal (Hair et al., 2019). Five items lie between 0.65 and 0.69 (CONF3, PV4, RI2, RI3, RI4). This is acceptable; do not drop an item only because its loading is below 0.7 if the scale still meets the CR and AVE criteria (Post 24) and the item matters for content.</p></li><li><p><strong>R&#178;</strong> is the square of the standardized loading: 72% of the variance of PV3 is explained by Perceived value, but only 43% of PV4.</p></li></ul><p>The order of the loadings is quite similar to the EFA in Post 12 (for example, PV3 is the highest in the Perceived value scale), but the values are not identical: unlike EFA, CFA does not allow items to have secondary loadings on other factors.</p><h3>Step 4. Read the correlations between the factors</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!6yPt!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!6yPt!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 424w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 848w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 1272w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!6yPt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png" width="760" height="352" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:352,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:114416,&quot;alt&quot;:&quot;Figure 6. Factor covariances in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Factor covariances in JASP" title="Figure 6. Factor covariances in JASP" srcset="https://substackcdn.com/image/fetch/$s_!6yPt!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 424w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 848w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 1272w, https://substackcdn.com/image/fetch/$s_!6yPt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac60ad31-2509-46ed-824f-1d6b4a82cba7_760x352.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. The Factor Covariances table; with Standardization = All, the Std. estimate column holds the correlations.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/OrFsA/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c9a1684d-1fd3-45cd-9fe6-3453e3705072_1220x504.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fabaed3b-e01b-4a99-a5bd-e7e78b251636_1220x624.png&quot;,&quot;height&quot;:264,&quot;title&quot;:&quot;Table 4. Correlations between the five factors (standardized estimates).&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/OrFsA/1/" width="730" height="264" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><ul><li><p><strong>The highest correlation is 0.64</strong> (Perceived value and Satisfaction), well below the cutoff of 0.85 often used to warn that two concepts overlap (Kline, 2023). Post 25 tests discriminant validity in more detail with the Fornell-Larcker criterion and HTMT.</p></li><li><p><strong>Perceived value and Trust are almost uncorrelated</strong> (0.13; <em>p</em> = 0.067 in JASP's standardized table, 95% CI [&#8722;0.009, 0.273]), in line with the Pearson correlation between the mean scores in <a href="https://www.phuonglib.com/p/pearson-correlation-pspp-spss">Post 13, "Pearson Correlation in PSPP (SPSS)"</a> (<em>r</em> = 0.12; <em>p</em> = 0.051). The test of the unstandardized covariance in the R script gives <em>p</em> = 0.076, the value AMOS reports; both are above 0.05.</p></li><li><p>The correlations in the CFA are <strong>higher</strong> than the correlations between the mean scores in Post 13, because CFA removes measurement error (Post 19).</p></li></ul><h3>Step 5. Look at the modification indices</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!EPw6!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!EPw6!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 424w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 848w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 1272w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!EPw6!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png" width="560" height="629" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:629,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:77625,&quot;alt&quot;:&quot;Figure 7. Modification indices in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Modification indices in JASP" title="Figure 7. Modification indices in JASP" srcset="https://substackcdn.com/image/fetch/$s_!EPw6!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 424w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 848w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 1272w, https://substackcdn.com/image/fetch/$s_!EPw6!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F63c805fd-8fe0-4ae1-acd9-50376c421c26_560x629.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. The Modification Indices tables: Cross-loadings and Residual covariances (cutoff 3.84).</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!9cl2!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!9cl2!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 424w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 848w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 1272w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!9cl2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png" width="760" height="454" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:454,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:127332,&quot;alt&quot;:&quot;Figure 8. The model plot drawn by JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. The model plot drawn by JASP" title="Figure 8. The model plot drawn by JASP" srcset="https://substackcdn.com/image/fetch/$s_!9cl2!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 424w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 848w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 1272w, https://substackcdn.com/image/fetch/$s_!9cl2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fad587d42-b227-43ca-a9a7-b8b1608e04d7_760x454.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. The model plot with standardized estimates. With 20 items the JASP plot is crowded and shortens the names (CONFI, VAL, SATI, TRUS, INT; TRU4 for TRU4R); enlarge it or redraw it as in Figure 4 for a report.</figcaption></figure></div><p>A modification index (MI) estimates how much the chi-square would drop if a fixed parameter were freed (Brown, 2015). JASP lists nine MIs above the cutoff of 3.84; the five largest are:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/fFCu8/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2d253e74-c169-44ea-b13d-ff0aece44b52_1220x578.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/843d59ce-9a21-411e-be0b-15ab64b31581_1220x648.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 5. The five largest modification indices.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/fFCu8/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>JASP prints the expected change in unstandardized form. In the Cross-loadings table, a row such as "SATISF, PV4" means the factor SATISF would get an extra loading on item PV4.</p><p><strong>Do not modify the model on the basis of these MIs.</strong> Three reasons:</p><ol><li><p><strong>The MIs are small.</strong> The largest, 6.12, is only just above 3.84 (the critical chi-square with 1 degree of freedom at the 0.05 level). With 190 parameters that could be freed, a few MIs above 3.84 by chance are normal.</p></li><li><p><strong>There is no theoretical reason.</strong> Nothing suggests that the error of an item about delivery service (CONF3) should correlate with the error of an item about the plan to keep buying (RI1).</p></li><li><p><strong>The model already fits well.</strong> Further changes would only fit the model to the peculiarities of this sample and would be hard to replicate in another one (MacCallum et al., 1992).</p></li></ol><p>An additional check is the <strong>standardized residuals</strong> (computed by the R script): only 1 of the 190 residuals exceeds 2.58 in absolute value (the largest is 2.82), so no pair of items is poorly reproduced by the model.</p><h2>4. Testing multivariate normality</h2><p>ML estimation assumes the data are <strong>multivariate normal</strong>. When multivariate kurtosis is high, the chi-square is inflated and the standard errors are underestimated, which leads to rejecting the model and finding significance more often than is warranted (Kline, 2023). This is why the FAQ of Post 9 recommended checking normality <strong>at the CFA stage</strong> rather than testing each item with Shapiro-Wilk.</p><h3>Distribution of each item</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!lzJQ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!lzJQ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 424w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 848w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 1272w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!lzJQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png" width="2000" height="1040" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1040,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:116282,&quot;alt&quot;:&quot;Figure 9. Skewness and kurtosis of the 20 items&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Skewness and kurtosis of the 20 items" title="Figure 9. Skewness and kurtosis of the 20 items" srcset="https://substackcdn.com/image/fetch/$s_!lzJQ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 424w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 848w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 1272w, https://substackcdn.com/image/fetch/$s_!lzJQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F328d98a0-058d-4c28-ba6f-576a7a3edf24_2000x1040.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. Skewness and kurtosis of each item (N = 254 complete respondents), computed by the R script.</figcaption></figure></div><ul><li><p><strong>Skewness</strong> ranges from &#8722;0.99 (PV3) to &#8722;0.50 (TRU4R): every item is slightly left-skewed, because most respondents chose "Agree". <strong>Kurtosis</strong> ranges from &#8722;0.53 to 1.05.</p></li><li><p>Both are far from the usual limits for single variables: |skewness| &lt; 2 and |kurtosis| &lt; 7 (Kline, 2023).</p></li><li><p>The critical ratios (c.r.) of skewness range from &#8722;6.46 to &#8722;3.27, all beyond 1.96. This does <strong>not</strong> mean the distributions are severely skewed: the c.r. is the skewness divided by its standard error, and the standard error shrinks as the sample grows, so with 254 respondents even a small skewness gives a large c.r. Read the size of the skewness, not the c.r. of each item.</p></li></ul><h3>Mardia's coefficient</h3><p>The <em>Factor</em> module of JASP has no Mardia's test. The <strong>SEM</strong> module does:</p><ol><li><p><strong>Enable</strong> the SEM module with the plus button at the right end of the toolbar and tick <em>SEM</em>.</p></li><li><p><strong>Choose</strong> <strong>SEM &gt; Structural Equation Modeling</strong> (Figure 10).</p></li><li><p><strong>Type</strong> the model in lavaan syntax into the editor and press Cmd+Enter (Ctrl+Enter on Windows) (Figure 11):</p></li></ol><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE =~ PV1 + PV2 + PV3 + PV4
SATISF =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT =~ RI1 + RI2 + RI3 + RI4</code></pre><ol><li><p><strong>Tick</strong> <em>Mardia's coefficient</em> in the <em>Output Options</em> section (Figure 12).</p></li></ol><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!GXt5!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 424w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 848w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 1272w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!GXt5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png" width="520" height="240" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:240,&quot;width&quot;:520,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:58168,&quot;alt&quot;:&quot;Figure 10. The SEM menu in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. The SEM menu in JASP" title="Figure 10. The SEM menu in JASP" srcset="https://substackcdn.com/image/fetch/$s_!GXt5!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 424w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 848w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 1272w, https://substackcdn.com/image/fetch/$s_!GXt5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc715270-fa75-47d1-bc3a-e32e993d0069_520x240.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 10. (1) The SEM module after it is enabled, (2) Structural Equation Modeling.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!jUqL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!jUqL!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 424w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 848w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 1272w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!jUqL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png" width="700" height="373" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8297d346-068b-4c88-8799-40f13af90700_700x373.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:373,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:60249,&quot;alt&quot;:&quot;Figure 11. The CFA model in lavaan syntax in the SEM module&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 11. The CFA model in lavaan syntax in the SEM module" title="Figure 11. The CFA model in lavaan syntax in the SEM module" srcset="https://substackcdn.com/image/fetch/$s_!jUqL!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 424w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 848w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 1272w, https://substackcdn.com/image/fetch/$s_!jUqL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8297d346-068b-4c88-8799-40f13af90700_700x373.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 11. (1) The CFA model written in lavaan syntax, (2) the hint to apply the syntax (Cmd+Enter on a Mac).</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!IFTX!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!IFTX!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 424w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 848w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 1272w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!IFTX!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png" width="700" height="390" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:390,&quot;width&quot;:700,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:98951,&quot;alt&quot;:&quot;Figure 12. The Mardia's coefficient option in the SEM module&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 12. The Mardia's coefficient option in the SEM module" title="Figure 12. The Mardia's coefficient option in the SEM module" srcset="https://substackcdn.com/image/fetch/$s_!IFTX!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 424w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 848w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 1272w, https://substackcdn.com/image/fetch/$s_!IFTX!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F315f7527-9d59-4485-a5e5-7a98bb80c7c7_700x390.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 12. (1) The Mardia's coefficient option in Output Options.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Tw9F!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Tw9F!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 424w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 848w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 1272w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Tw9F!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png" width="760" height="531" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/cf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:531,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:79153,&quot;alt&quot;:&quot;Figure 13. Mardia's test in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 13. Mardia's test in JASP" title="Figure 13. Mardia's test in JASP" srcset="https://substackcdn.com/image/fetch/$s_!Tw9F!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 424w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 848w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 1272w, https://substackcdn.com/image/fetch/$s_!Tw9F!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fcf121e7b-dbe1-4928-8001-53069f8c7d91_760x531.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 13. The Model Fit table and the Mardia's Coefficients table: multivariate skewness and kurtosis.</figcaption></figure></div><p>JASP computes Mardia's coefficients on the 254 complete respondents (the R script reproduces both values from these 254 rows): the kurtosis coefficient (<em>Kurtosis</em>) = 453.6, <em>z</em> = 3.657. The value 453.6 is Mardia's b2; its expected value for normal data is <em>p</em>(<em>p</em> + 2) = 20 &#215; 22 = 440. AMOS reports the <strong>excess</strong> over 440, so to compare with the way theses using AMOS usually report it, take 453.6 &#8722; 440 = 13.6. With the original formula of Mardia (1970), which uses the covariance matrix divided by <em>N</em> instead of <em>N</em> &#8722; 1 as JASP does:</p><blockquote><p>Multivariate kurtosis = 17.21; critical ratio = 4.62.</p></blockquote><p>Bentler (2005) suggests treating the data as <strong>not multivariate normal</strong> when the critical ratio of Mardia's coefficient is <strong>greater than 5</strong>. At 4.62, the sample data are below that cutoff, so ML estimation is appropriate.</p><p>JASP's <em>z</em> = 3.657 is lower than 4.62 because of the different denominator described above; both are below 5. The JASP table also has a test of multivariate skewness (<em>Skewness</em>: chi-square = 1934; df = 1540; <em>p</em> &lt; 0.001). This test is very sensitive to sample size and to the common direction of skew in Likert scales. ML estimation is more sensitive to kurtosis than to skewness, which is why Bentler's rule is based on kurtosis (Kline, 2023). Still, significant skewness is a reason to double-check with the Bollen-Stine bootstrap below.</p><p>A further note: the <em>Model Fit</em> table of the SEM module shows df = 155 instead of 160 (chi-square = 134.3; <em>p</em> = 0.884). With FIML, the SEM module of JASP 0.99.0 also estimates the means of the five factors (the <em>Intercepts</em> table shows them as approximately 0); the chi-square does not change. The table also carries a note that fitting the model resulted in warnings. For these reasons, this post uses the <em>Factor</em> module for the main CFA results and the SEM module only for Mardia's coefficients.</p><h3>Multivariate outliers</h3><p>The Mahalanobis distance (<em>d*&#178;) measures how far each respondent is from the centroid on all 20 items at once. The five respondents farthest from the centroid have *d*&#178; from 38.65 to 43.53 (IDs 112, 189, 82, 267 and 135), with *p</em> from 0.002 to 0.007. Nobody has <em>p</em> &lt; 0.001, the usual cutoff for a multivariate outlier (Tabachnick &amp; Fidell, 2019), which agrees with the conclusion of <a href="https://www.phuonglib.com/p/data-cleaning-missing-values-outliers-spss-pspp">Post 8, "Data Cleaning in PSPP (SPSS)"</a>.</p><h3>The Bollen-Stine bootstrap</h3><p>When the critical ratio of Mardia's coefficient exceeds 5, you can keep the ML estimates but replace the <em>p</em>-value of the chi-square with the <strong>Bollen-Stine bootstrap</strong> <em>p</em>-value (Bollen &amp; Stine, 1992). This method transforms the data so that the model fits them exactly, draws 2000 bootstrap samples and counts the proportion of samples with a chi-square larger than the observed value.</p><p>The sample data do not need it, but the post runs it as an illustration: on the 254 complete respondents, chi-square = 137.05 and the Bollen-Stine <em>p</em> = 0.922 (2000 samples), close to the <em>p</em> = 0.931 of the usual test. The conclusion about fit does not change.</p><p>JASP 0.99.0 has no Bollen-Stine option in its interface; the script <code>cfa_5_factors.R</code> in the repository computes it directly with lavaan. AMOS users choose it under <strong>Analysis Properties &gt; Bootstrap</strong>.</p><h2>5. Reporting the CFA in a thesis</h2><blockquote><p>A confirmatory factor analysis with maximum likelihood estimation and FIML for missing data (N = 267) showed that the five-factor measurement model fitted the data, chi-square = 134.29; df = 160; <em>p</em> = 0.931; CFI = 1.000; TLI = 1.014; RMSEA = 0.000 (90% CI [0.000, 0.008]); SRMR = 0.032. All standardized loadings were statistically significant (<em>p</em> &lt; 0.001) and ranged from 0.65 to 0.85. The correlations between the factors ranged from 0.13 to 0.64. Mardia's coefficient of multivariate kurtosis was 17.21 (c.r. = 4.62), below the cutoff of 5 proposed by Bentler (2005); the Bollen-Stine bootstrap <em>p</em> was 0.922 (2000 samples), leading to the same conclusion as the usual chi-square test.</p></blockquote><p>Attach the table of standardized loadings (section 3, step 3) and the table of factor correlations (step 4). CR, AVE and discriminant validity will be added to the same tables after Posts 24 and 25.</p><h2>Five common mistakes</h2><ol><li><p><strong>Forgetting the covariances between the factors.</strong> The model becomes five independent factors, the fit is very poor and the loadings are wrong.</p></li><li><p><strong>Putting one item into two factors</strong> by dragging it twice. Check each group of items before running.</p></li><li><p><strong>Dropping items only because their loadings are below 0.7</strong> without considering CR, AVE and the content of the scale.</p></li><li><p><strong>Adding error covariances based on MIs until the indices look "nice".</strong> Every modification needs a theoretical reason and must be reported.</p></li><li><p><strong>Reading the critical ratio of each item's skewness and concluding that the data are not normal.</strong> With a sample of a few hundred, read the size of the skewness and kurtosis and Mardia's coefficient.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Do I need to run a separate CFA for each scale first?</strong></p><p>It is not required. Running each scale separately (as in Post 20) helps catch problem items early, but the main result is the CFA of the whole measurement model, because it also tests the discriminant validity between the concepts.</p><p><strong>At what standardized loading should an item be dropped?</strong></p><p>Consider dropping items with a standardized loading below 0.5, or items that pull the AVE of the scale below 0.5 (Hair et al., 2019). Before dropping an item, check whether doing so removes an important aspect of the concept's content, and report clearly which items were dropped.</p><p><strong>What should I do if Mardia's coefficient exceeds 5?</strong></p><p>There are three options: use the Bollen-Stine bootstrap <em>p</em> for the test of fit, use the bootstrap for the confidence intervals of the coefficients or use a robust estimator (for example MLR in lavaan and JASP). You do not need to switch to PLS-SEM just because the data are not normal (Kline, 2023).</p><p><strong>Simulated data give results that are too good. What about real data?</strong></p><p>With real survey data, a significant chi-square, a CFI around 0.90 to 0.95, a few loadings below 0.6 and larger MIs are common. Post 23 shows how to read fit in those situations.</p><div><hr></div><h2>References</h2><p>Anderson, J. C., &amp; Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. <em>Psychological Bulletin, 103</em>(3), 411&#8211;423. https://doi.org/10.1037/0033-2909.103.3.411</p><p>Bentler, P. M. (2005). <em>EQS 6 structural equations program manual</em>. Multivariate Software.</p><p>Bollen, K. A., &amp; Stine, R. A. (1992). Bootstrapping goodness-of-fit measures in structural equation models. <em>Sociological Methods &amp; Research, 21</em>(2), 205&#8211;229. https://doi.org/10.1177/0049124192021002004</p><p>Brown, T. A. (2015). <em>Confirmatory factor analysis for applied research</em> (2nd ed.). Guilford Press.</p><p>Byrne, B. M. (2016). <em>Structural equation modeling with AMOS: Basic concepts, applications, and programming</em> (3rd ed.). Routledge.</p><p>Hair, J. F., Black, W. C., Babin, B. J., &amp; Anderson, R. E. (2019). <em>Multivariate data analysis</em> (8th ed.). Cengage Learning.</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Kline, R. B. (2023). <em>Principles and practice of structural equation modeling</em> (5th ed.). Guilford Press.</p><p>MacCallum, R. C., Roznowski, M., &amp; Necowitz, L. B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. <em>Psychological Bulletin, 111</em>(3), 490&#8211;504. https://doi.org/10.1037/0033-2909.111.3.490</p><p>Mardia, K. V. (1970). Measures of multivariate skewness and kurtosis with applications. <em>Biometrika, 57</em>(3), 519&#8211;530. https://doi.org/10.1093/biomet/57.3.519</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p><p>Tabachnick, B. G., &amp; Fidell, L. S. (2019). <em>Using multivariate statistics</em> (7th ed.). Pearson.</p>]]></content:encoded></item><item><title><![CDATA[Hypothesis Testing in SEM with JASP: Coefficients, p-Values and f²]]></title><description><![CDATA[Label the paths in free JASP, read standardized coefficients, p-values and confidence intervals, compare two coefficients and compute f&#178; for every hypothesis]]></description><link>https://www.phuonglib.com/p/hypothesis-testing-sem-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/hypothesis-testing-sem-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:48:02 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!lwSM!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!lwSM!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!lwSM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:870164,&quot;alt&quot;:&quot;Seven slender rope bridges across a pale canyon, small figures rolling stone balls over them, one bridge sagging, an orange balance scale on the rim.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Seven slender rope bridges across a pale canyon, small figures rolling stone balls over them, one bridge sagging, an orange balance scale on the rim." title="Seven slender rope bridges across a pale canyon, small figures rolling stone balls over them, one bridge sagging, an orange balance scale on the rim." srcset="https://substackcdn.com/image/fetch/$s_!lwSM!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!lwSM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4936512e-3ec2-43c8-9477-2510438ef171_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Read each hypothesis in five steps: direction, significance, size, comparison and conclusion.</p></li><li><p>Six of the seven hypotheses are supported; H4 (beta = 0.09, <em>p</em> = 0.372) means no evidence of a direct effect, not "no effect".</p></li><li><p>State whether a <em>p</em>-value tests B or beta: JASP with <em>Standardized estimates = All</em> tests beta and AMOS tests B; the two differ slightly.</p></li><li><p>f&#178; measures the unique contribution of a direct path: VALUE is significant (beta = 0.23) yet its f&#178; is only 0.012.</p></li><li><p>Do not rank coefficients by eye: trust (0.34) and satisfaction (0.23) do not differ significantly (<em>p</em> = 0.379).</p></li></ul></blockquote><div><hr></div><p><a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a> ran the structural model and found that six of the seven paths were significant. This post goes deeper into the question that the results chapter of every thesis has to answer: <strong>is each hypothesis supported or not, and how strong is the effect?</strong> It shows how to read unstandardized and standardized coefficients, <em>p</em>-values and confidence intervals correctly, how to measure effect size with f&#178;, how to compare two coefficients with a test instead of by eye and how to write the conclusion for a hypothesis that is not supported.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch27-hypothesis-testing-sem">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The screenshots come from JASP 0.99.0 (JASP Team, 2026) on macOS, continuing the SEM analysis of Post 26 on the data file <code>marketplace_x_sem.sav</code>. The result tables are taken from JASP's <em>Export Results</em>. The numbers were checked with the lavaan package for R (Rosseel, 2012), using the script <code>hypothesis_testing.R</code> in the repository.</p></blockquote><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!9Ino!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!9Ino!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 424w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 848w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 1272w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!9Ino!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png" width="2000" height="740" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/0552e329-dfca-46d0-981a-998e12b41745_2000x740.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:740,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:101462,&quot;alt&quot;:&quot;Figure 1. Five steps to read a hypothesis&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Five steps to read a hypothesis" title="Figure 1. Five steps to read a hypothesis" srcset="https://substackcdn.com/image/fetch/$s_!9Ino!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 424w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 848w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 1272w, https://substackcdn.com/image/fetch/$s_!9Ino!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0552e329-dfca-46d0-981a-998e12b41745_2000x740.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Five steps to read a hypothesis in SEM.</figcaption></figure></div><h2>1. Unstandardized and standardized coefficients</h2><p>Each path has two coefficients:</p><ul><li><p><strong>Unstandardized coefficient</strong> (B): when the cause increases by one unit, the outcome changes by B units, holding the other variables constant. A latent variable takes the unit of the item whose loading is fixed to 1 (<a href="https://www.phuonglib.com/p/first-model-jasp-free-amos-alternative">Post 20, "Your First Model in JASP"</a>), so B is hard to interpret directly. <strong>Significance tests</strong> (S.E., C.R., <em>p</em>) are computed on B.</p></li><li><p><strong>Standardized coefficient</strong> (beta): when the cause increases by one standard deviation, the outcome changes by beta standard deviations. Beta is used to <strong>compare the size</strong> of paths within the same model.</p></li></ul><p>Kline (2023) recommends reporting both, or at least beta together with <em>p</em>.</p><h2>2. Label the paths and read the output in JASP</h2><p>Open the SEM analysis of Post 26 and edit the structural part of the model to <strong>add labels</strong> h1 to h7 in front of the names of the causes (Figure 2):</p><pre><code>VALUE ~ h1*CONFIRM
SATISF ~ h2*CONFIRM + h3*VALUE
INTENT ~ h4*CONFIRM + h5*VALUE + h6*SATISF + h7*TRUST
CONFIRM ~~ TRUST
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1
# Compare two coefficients: Trust and Satisfaction to Intention
diff_TRU_SAT := h7 - h6</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!pUcg!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!pUcg!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 424w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 848w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 1272w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!pUcg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png" width="900" height="432" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:432,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:138425,&quot;alt&quot;:&quot;Figure 2. The model with labels for the seven paths&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. The model with labels for the seven paths" title="Figure 2. The model with labels for the seven paths" srcset="https://substackcdn.com/image/fetch/$s_!pUcg!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 424w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 848w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 1272w, https://substackcdn.com/image/fetch/$s_!pUcg!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfd97d74-1c7a-45fd-b46f-74aa9bd0f03e_900x432.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) Labels h1 to h7 placed in front of the causes, (2) a defined parameter to compare h7 with h6 (Section 6).</figcaption></figure></div><p>Labels do not change the results; they only give the parameters names so that you can reuse them. A <code>:=</code> line defines a new parameter computed from labeled parameters; JASP prints it in the <em>Defined Parameters</em> table. Press Cmd + Enter to apply the syntax.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!jAQV!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!jAQV!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 424w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 848w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 1272w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!jAQV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png" width="900" height="329" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:329,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:112949,&quot;alt&quot;:&quot;Figure 3. The regression coefficients table in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. The regression coefficients table in JASP" title="Figure 3. The regression coefficients table in JASP" srcset="https://substackcdn.com/image/fetch/$s_!jAQV!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 424w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 848w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 1272w, https://substackcdn.com/image/fetch/$s_!jAQV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F335d195f-d269-46a9-b4b7-3c2ae178610c_900x329.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. The Regression Coefficients table with Standardized estimates = All. (1) The label of each path, (2) the p-value of the standardized estimate, (3) the 95% confidence interval of the standardized coefficient.</figcaption></figure></div><p>With <em>Standardized estimates</em> ticked and <em>All</em> selected in <em>Output Options</em>, the <em>Std. estimate</em> column is beta, and the standard error, <em>z</em>-value, <em>p</em> and confidence interval refer to beta. JASP lists the outcomes in the order INTENT, SATISF, VALUE, so read the labels rather than the row order. Untick <em>Standardized estimates</em> to see B with its own standard error, <em>z</em>-value (the equivalent of AMOS's C.R.), <em>p</em> and confidence interval. The test of beta and the test of B give slightly different <em>p</em>-values: lavaan's test of B, which AMOS also reports, gives 0.375 for h4, 0.015 for h5 and 0.026 for h6, against .372, .011 and .023 in Figure 3. The conclusions are the same. This post reports the numbers JASP prints with <em>Standardized estimates = All</em>.</p><p>For bootstrap confidence intervals, set <em>Standard errors</em> to <em>Bootstrap</em> in <em>Estimation Options</em> and enter the number of <em>Bootstrap samples</em> (<a href="https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp">Post 28, "Mediation Analysis with JASP"</a> uses 2000, <em>Type</em> Percentile, random seed 2026). With 267 respondents and FIML, each run takes a few minutes.</p><p><strong>In AMOS</strong>, the same output is in <strong>View &gt; Text Output</strong> (F10) &gt; <em>Estimates</em>: the <em>Regression Weights</em> table gives B, S.E., C.R. (B divided by S.E.) and P, and <em>Standardized Regression Weights</em> gives beta. For H1 to H7, B = 0.627, 0.412, 0.389, 0.088, 0.208, 0.212 and 0.316, with C.R. = 7.22, 4.87, 5.02, 0.89, 2.44, 2.23 and 4.46. The P column prints three asterisks (<code>***</code>) when <em>p</em> &lt; 0.001; report it as "<em>p</em> &lt; 0.001". AMOS prints confidence intervals only after a bootstrap (<strong>Analysis Properties &gt; Bootstrap</strong>), which needs the complete-case file <code>marketplace_x_sem_complete.sav</code>. To label a path, double-click the arrow and type the label in the <em>Regression weight</em> box of <em>Object Properties</em>, or label every parameter with <strong>Plugins &gt; Name Parameters</strong>.</p><h2>3. Results for the seven hypotheses</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!IMHt!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!IMHt!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 424w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 848w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!IMHt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png" width="2000" height="1000" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1000,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:126948,&quot;alt&quot;:&quot;Figure 4. Seven hypotheses&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Seven hypotheses" title="Figure 4. Seven hypotheses" srcset="https://substackcdn.com/image/fetch/$s_!IMHt!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 424w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 848w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!IMHt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c337838-f9a4-40f6-964c-8b0094ae0779_2000x1000.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Standardized coefficients and 95% confidence intervals of the seven hypotheses.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/yB9sv/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8f4bca3a-02df-425b-899e-6e98172ced7c_1220x986.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f07abb43-1ebe-4cc1-b697-23f08233c95b_1220x1200.png&quot;,&quot;height&quot;:568,&quot;title&quot;:&quot;Table 1. Results for the seven hypotheses (standardized coefficients, JASP)&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/yB9sv/1/" width="730" height="568" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Read each hypothesis with the five steps of Figure 1, for example H7:</p><ol><li><p><strong>Read the direction.</strong> beta = 0.34 is positive, in the hypothesized direction (higher trust goes with higher repurchase intention).</p></li><li><p><strong>Read the significance.</strong> <em>z</em> = 4.95 &gt; 1.96; <em>p</em> &lt; 0.001; the confidence interval [0.20, 0.47] does not contain 0.</p></li><li><p><strong>Read the size.</strong> When trust increases by one standard deviation, repurchase intention increases by 0.34 standard deviations, holding the other three variables constant.</p></li><li><p><strong>Compare.</strong> It is the largest coefficient among the four predictors of intention, but it is not necessarily significantly larger (Section 6).</p></li><li><p><strong>Conclude.</strong> H7 is supported.</p></li></ol><h3>Interpreting H4, which is not supported</h3><p>H4 has <em>p</em> = 0.372 and its confidence interval contains 0. The correct wording is: "there is <em>no evidence</em> of a <strong>direct</strong> effect of expectation confirmation on repurchase intention once perceived value, satisfaction and trust are taken into account". Do not write "expectation confirmation does not affect repurchase intention", because:</p><ul><li><p>A non-significant result does not prove that the effect is zero (<a href="https://www.phuonglib.com/p/t-test-independent-paired-one-sample-pspp-spss">Post 17, "T-Tests in PSPP (SPSS)"</a>). The confidence interval shows that the true effect could lie anywhere from &#8722;0.11 to 0.29.</p></li><li><p>Expectation confirmation still has a substantial <strong>total effect</strong> (0.37) through perceived value and satisfaction (Post 26). This result suggests <strong>full mediation</strong>, which Post 28 tests.</p></li></ul><p>The sample data were simulated with a direct effect of zero (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a>), so not supporting H4 here matches the truth of the simulated population.</p><h2>4. Confidence intervals and bootstrapping</h2><p>The <em>p</em>-values and confidence intervals above rely on the assumption that the estimates are normally distributed. <strong>Bootstrapping</strong> does not need this assumption: it draws 2000 samples with replacement from the data, re-estimates the model each time and takes the 2.5th and 97.5th percentiles of the coefficients as the confidence interval (Kline, 2023).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Ri2x!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Ri2x!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 424w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 848w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 1272w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Ri2x!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png" width="2000" height="960" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/42d56064-e97c-465f-b432-64045439edd0_2000x960.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:960,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:114503,&quot;alt&quot;:&quot;Figure 5. Normal-theory and bootstrap confidence intervals&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Normal-theory and bootstrap confidence intervals" title="Figure 5. Normal-theory and bootstrap confidence intervals" srcset="https://substackcdn.com/image/fetch/$s_!Ri2x!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 424w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 848w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 1272w, https://substackcdn.com/image/fetch/$s_!Ri2x!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42d56064-e97c-465f-b432-64045439edd0_2000x960.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. 95% confidence intervals of the unstandardized coefficients: from standard errors (FIML, 267 respondents) and percentile bootstrap (2000 samples, 254 complete respondents).</figcaption></figure></div><p>Both methods lead to the same conclusion for all seven hypotheses. The bootstrap lower bound of <strong>H5</strong> is only 0.006, very close to 0: the evidence for the effect of perceived value on repurchase intention is weaker than for the other hypotheses, and the conclusion could change in another sample. This is worth reporting, and the <em>p</em>-value of 0.011 alone does not show it.</p><p>Bootstrapping is especially advisable when the data are non-normal (Mardia's coefficient above 5, <a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">Post 22, "Confirmatory Factor Analysis (CFA) in JASP"</a>) or when testing indirect effects (Post 28).</p><h2>5. Effect size: f&#178;</h2><p>Beta tells you the size of each path, but it does not directly tell you how much each variable <strong>contributes</strong> to R&#178;. The <strong>f&#178;</strong> effect size (Cohen, 1988) measures how much R&#178; drops when one path is removed:</p><p>f&#178; = (full R&#178; &#8722; R&#178; without the path) / (1 &#8722; full R&#178;)</p><p>Cohen's thresholds: 0.02 is small, 0.15 is medium and 0.35 is large. Remove one path to INTENT at a time, rerun the model and record the R&#178; of INTENT:</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!CXHI!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!CXHI!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 424w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 848w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 1272w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!CXHI!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png" width="1800" height="780" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ebf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:780,&quot;width&quot;:1800,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:78321,&quot;alt&quot;:&quot;Figure 6. Effect size f&#178;&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Effect size f&#178;" title="Figure 6. Effect size f&#178;" srcset="https://substackcdn.com/image/fetch/$s_!CXHI!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 424w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 848w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 1272w, https://substackcdn.com/image/fetch/$s_!CXHI!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Febf1f803-c023-4366-a6cb-74312d1b1489_1800x780.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. f&#178; of the four predictors of Repurchase intention.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/Hd9qL/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/bd56a19e-98bc-4fcd-a384-6d7e7bd5b0f9_1220x508.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/66ae3133-a115-4474-8649-9e41d3e18510_1220x666.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 2. Effect size f&#178; of the four predictors of repurchase intention&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/Hd9qL/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>A notable result: VALUE has beta = 0.23 and is significant, yet its f&#178; is only 0.012. The reason is that when the direct path from VALUE to INTENT is removed, most of VALUE's effect still flows through SATISF (VALUE affects SATISF with beta = 0.41), so the R&#178; of INTENT barely drops. f&#178; measures the <strong>unique contribution</strong> of the direct path, not the overall importance of the variable.</p><p>f&#178; is a common index in PLS-SEM (Hair et al., 2022) and is rarely used in CB-SEM; neither AMOS nor JASP computes it. The file <code>hypothesis_testing.R</code> computes it automatically; in JASP, do it by hand: delete one predictor from the <code>INTENT ~</code> line, run the model again and read the R&#178; of INTENT in the <em>R-Squared</em> table. In AMOS, delete one arrow at a time and read <em>Squared Multiple Correlations</em>.</p><h2>6. Comparing two coefficients: do not compare by eye</h2><p>TRUST has beta = 0.34 and SATISF has beta = 0.23. Can you conclude that "trust matters more than satisfaction in predicting repurchase intention"? Only if the <strong>difference between the two coefficients</strong> is significant. The two confidence intervals [0.20, 0.47] and [0.03, 0.42] overlap a great deal, which suggests the answer is no.</p><p>In JASP, the line <code>diff_TRU_SAT := h7 - h6</code> added in Section 2 creates a defined parameter equal to the difference between the two coefficients. JASP prints a <em>z</em> test for this parameter:</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!xOak!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 424w, https://substackcdn.com/image/fetch/$s_!xOak!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 848w, https://substackcdn.com/image/fetch/$s_!xOak!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 1272w, https://substackcdn.com/image/fetch/$s_!xOak!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!xOak!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png" width="760" height="170" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:170,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:30473,&quot;alt&quot;:&quot;Figure 7. The defined parameter in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. The defined parameter in JASP" title="Figure 7. The defined parameter in JASP" srcset="https://substackcdn.com/image/fetch/$s_!xOak!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 424w, https://substackcdn.com/image/fetch/$s_!xOak!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 848w, https://substackcdn.com/image/fetch/$s_!xOak!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 1272w, https://substackcdn.com/image/fetch/$s_!xOak!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7b3a445c-55b8-4542-b028-65b8b6e99c3d_760x170.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 7. (1) The difference between the coefficients of Trust and Satisfaction on Repurchase intention.</figcaption></figure></div><p>With <em>Standardized estimates = All</em>, JASP prints the difference between the two betas and its test; with <em>Standardized estimates</em> unticked, the difference between the two Bs. Both lead to the same conclusion here (lavaan gives B(TRUST) &#8722; B(SATISF) = 0.104, S.E. = 0.119, <em>z</em> = 0.87, <em>p</em> = 0.384).</p><p>In AMOS, label the paths and tick <em>Critical ratios for differences</em> in <strong>Analysis Properties &gt; Output</strong>. The <em>Pairwise Parameter Comparisons</em> section of the Text Output then prints the critical ratio for the difference between every pair of parameters; find the cell where h6 and h7 meet. An absolute value above 1.96 means a significant difference at the 0.05 level. A second way is to add a model with the constraint <code>h6 = h7</code> in the <em>Models</em> pane and read the chi-square difference in <em>Model Comparison</em>. Both compare the unstandardized coefficients, which is meaningful here because TRUST and SATISF are measured on similar five-point scales.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/JYZVQ/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2f309dc7-e58f-48f2-a58e-2c76384f8271_1220x314.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/79021654-fdee-40b9-b22d-7970ffad96ac_1220x472.png&quot;,&quot;height&quot;:213,&quot;title&quot;:&quot;Table 3. Test of the difference between the effects of trust and satisfaction&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/JYZVQ/1/" width="730" height="213" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The difference is <strong>not significant</strong>. The correct conclusion: trust and satisfaction both have a significant effect on repurchase intention, and there is <strong>no evidence</strong> that the effect of trust is stronger. Writing "trust is the most important factor" only because 0.34 is larger than 0.23 is a very common mistake (Gelman &amp; Stern, 2006).</p><h2>7. Reporting the results</h2><p>A results table should contain the hypothesis, path, B, S.E., C.R. (<em>z</em>), <em>p</em>, beta and conclusion; Section 3 shows the standardized part as JASP prints it, and unticking <em>Standardized estimates</em> gives the B columns. Add confidence intervals if you have them. A sample paragraph:</p><blockquote><p>The structural model supported six of the seven hypotheses. Expectation confirmation had a positive effect on perceived value (H1: beta = 0.58, <em>p</em> &lt; 0.001) and satisfaction (H2: beta = 0.40, <em>p</em> &lt; 0.001); perceived value affected satisfaction (H3: beta = 0.41, <em>p</em> &lt; 0.001). Repurchase intention was affected by trust (H7: beta = 0.34, <em>p</em> &lt; 0.001), perceived value (H5: beta = 0.23, <em>p</em> = 0.011) and satisfaction (H6: beta = 0.23, <em>p</em> = 0.023). There was no evidence of a direct effect of expectation confirmation on repurchase intention (H4: beta = 0.09, <em>p</em> = 0.372, 95% CI [&#8722;0.11, 0.29]). Bootstrap confidence intervals (2000 samples) led to the same conclusions, although the lower bound for H5 was very close to 0. The difference between the effects of trust and satisfaction on repurchase intention was not significant (difference in beta = 0.11, <em>p</em> = 0.379).</p></blockquote><h2>Five common mistakes</h2><ol><li><p><strong>Writing that a hypothesis is "proved" or "confirmed".</strong> Cross-sectional data only allow you to say that a hypothesis is <strong>supported</strong> or <strong>not supported</strong>.</p></li><li><p><strong>Interpreting H4 as "no effect"</strong> and ignoring the indirect effect.</p></li><li><p><strong>Comparing coefficients by eye</strong> to rank "the most important factor".</p></li><li><p><strong>Reporting a </strong><em><strong>p</strong></em><strong>-value without saying which coefficient it belongs to.</strong> AMOS tests the unstandardized coefficient B (C.R. = B / S.E.); JASP with <em>Standardized estimates = All</em> tests beta. The two <em>p</em>-values are close but not identical (H5: <em>p</em> = 0.015 for B, 0.011 for beta), so state whether <em>p</em> comes from B or beta; use beta to compare the size of paths.</p></li><li><p><strong>Drawing causal conclusions</strong> from a one-time survey. A path in SEM represents a hypothesized relationship; causality comes from theory and research design, not from the software.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Is a hypothesis supported when p = 0.051?</strong></p><p>Not by the 0.05 threshold. But report the exact value with its confidence interval, and avoid both extremes: do not write "marginally significant", and do not conclude "there is no effect". State the 0.05 threshold before the analysis.</p><p><strong>One-tailed or two-tailed test?</strong></p><p>AMOS and JASP print two-tailed <em>p</em>-values. A directional hypothesis (for example "a positive effect") allows a one-tailed <em>p</em> (half the two-tailed <em>p</em> when the coefficient has the expected sign), but most journals and examiners are used to two-tailed <em>p</em>-values. If you use a one-tailed test, say so in the methods section.</p><p><strong>Is a standardized coefficient above 1 a problem?</strong></p><p>A standardized path coefficient can exceed 1 when the predictors are highly correlated (multicollinearity), and it usually signals that discriminant validity should be checked again (<a href="https://www.phuonglib.com/p/convergent-discriminant-validity-cfa-jasp">Post 25, "Convergent and Discriminant Validity in JASP"</a>). In this model, the largest coefficient is 0.58.</p><p><strong>Should the diagram show beta or B?</strong></p><p>Diagrams usually show beta (standardized) with asterisks for the significance level. The table should include B, S.E. and beta.</p><div><hr></div><h2>References</h2><p>Cohen, J. (1988). <em>Statistical power analysis for the behavioral sciences</em> (2nd ed.). Lawrence Erlbaum Associates.</p><p>Gelman, A., &amp; Stern, H. (2006). The difference between "significant" and "not significant" is not itself statistically significant. <em>The American Statistician, 60</em>(4), 328&#8211;331. https://doi.org/10.1198/000313006X152649</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Kline, R. B. (2023). <em>Principles and practice of structural equation modeling</em> (5th ed.). Guilford Press.</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p>]]></content:encoded></item><item><title><![CDATA[What Is SmartPLS? PLS-SEM vs CB-SEM and When to Use Each]]></title><description><![CDATA[Compare PLS-SEM and CB-SEM on the same data, from path coefficients and loadings to R&#178;, and choose the right method for your thesis]]></description><link>https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem</link><guid isPermaLink="false">https://www.phuonglib.com/p/what-is-smartpls-pls-sem-vs-cb-sem</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:47:56 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/c6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!J87Y!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!J87Y!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!J87Y!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:909320,&quot;alt&quot;:&quot;Two stone arches side by side, one carved from a single block and one built of small bricks, both holding up the same orange sphere.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Two stone arches side by side, one carved from a single block and one built of small bricks, both holding up the same orange sphere." title="Two stone arches side by side, one carved from a single block and one built of small bricks, both holding up the same orange sphere." srcset="https://substackcdn.com/image/fetch/$s_!J87Y!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!J87Y!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6177cca-19d1-4f2c-9c3d-da97d3b884c0_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>CB-SEM treats a construct as a common factor; PLS-SEM treats it as a composite, a weighted sum of its items.</p></li><li><p>On the same data PLS-SEM gives smaller paths (H1: 0.501 against 0.579), lower R&#178; and higher loadings than CB-SEM.</p></li><li><p>Consistent PLS (PLSc) corrects these biases and lands within 0.02 of CB-SEM.</p></li><li><p>Both methods reach the same conclusions about direction and significance; only the size of the coefficients differs.</p></li><li><p>A small sample or non-normal data is not a sufficient reason to choose PLS-SEM; justify the choice by the constructs and the research goal.</p></li></ul></blockquote><div><hr></div><p>The AMOS posts (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19</a> to <a href="https://www.phuonglib.com/p/common-method-bias-jasp-harman-clf">Post 30</a>) estimated the Marketplace X model with <strong>CB-SEM</strong>. This post opens the SmartPLS series with the question every thesis committee asks: why did you choose PLS-SEM, or why not? Instead of just listing pros and cons, the post estimates <strong>the same model on the same data set</strong> in four ways and puts the results side by side.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch31-what-is-smartpls-pls-sem-vs-cb-sem">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The numbers were computed with the R packages cSEM 0.7.1 (Rademaker &amp; Schuberth, 2026) and lavaan 0.7.3 (Rosseel, 2012), using the script <code>compare_pls_cbsem.R</code> in the repository, on the 254 respondents who answered all 20 items (<code>marketplace_x_sem_complete.sav</code>). The screenshots come from the Partial Least Squares SEM module of JASP 0.99.0 (JASP Team, 2026) on macOS; this module is built on the cSEM package, and its PLS-SEM estimates match the R results to the fourth decimal. There are no SmartPLS screenshots in this post.</p></blockquote><h2>1. What is SmartPLS?</h2><p>SmartPLS is software for estimating structural equation models with <strong>partial least squares</strong> (PLS-SEM), developed by Christian M. Ringle, Sven Wende and Jan-Michael Becker since 2005. The current version is <strong>SmartPLS 4.1.2.1</strong>, released on 25 October 2026 (Ringle et al., 2026). Its main features:</p><ul><li><p><strong>Drag-and-drop interface</strong>: you draw the model with the mouse and read Excel, CSV and SPSS files directly.</p></li><li><p><strong>The full PLS-SEM toolbox</strong>: bootstrapping, PLSpredict, multigroup analysis (MGA), IPMA, higher-order models and moderators. SmartPLS 4 also offers CB-SEM and regression.</p></li><li><p><strong>Commercial license</strong>: there is a 30-day trial with all features and a free Student license with reduced functionality (CSV files only, at most 100 rows of data; see <a href="https://www.phuonglib.com/p/install-smartpls-4-interface">Post 32, "How to Install SmartPLS 4 and Find Your Way Around the Interface"</a>). The SmartPLS release notes state that the Student license now includes basic CB-SEM functions and bootstrapping for CB-SEM.</p></li><li><p><strong>Operating systems</strong>: Windows and macOS.</p></li></ul><h2>2. How does PLS-SEM differ from CB-SEM?</h2><p>The core difference lies in <strong>how a construct is understood</strong> (Sarstedt et al., 2016):</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!0TpP!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!0TpP!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!0TpP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png" width="2100" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:121079,&quot;alt&quot;:&quot;Figure 1. Common factor and composite&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Common factor and composite" title="Figure 1. Common factor and composite" srcset="https://substackcdn.com/image/fetch/$s_!0TpP!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!0TpP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F365b20e6-9318-46ba-843a-97b2ca91dfad_2100x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. One construct, two ways to model it: a common factor (CB-SEM) and a composite (PLS-SEM).</figcaption></figure></div><ul><li><p><strong>CB-SEM</strong> treats a construct as a <strong>common factor</strong>: satisfaction is the cause that makes respondents give high scores on SAT1 to SAT4; the rest of each item is measurement error and is separated out. The method estimates all parameters at once so that the covariance matrix implied by the model fits the data as closely as possible.</p></li><li><p><strong>PLS-SEM</strong> treats a construct as a <strong>composite</strong>: a weighted sum of its four items. The algorithm alternates between two steps, finding a weight for each item and then running regressions between the composites, until the weights are stable (Wold, 1982; Hair et al., 2022). Its goal is to <strong>maximize the R&#178;</strong> of the dependent constructs.</p></li></ul><p>Because a composite also contains measurement error, PLS-SEM has two well-known biases when the construct really is a common factor (Dijkstra &amp; Henseler, 2015):</p><ul><li><p><strong>Loadings are inflated</strong>, because each item is part of the composite itself.</p></li><li><p><strong>Path coefficients and R&#178; are attenuated</strong>, because measurement error dilutes the relationships between constructs, just as in a regression on mean scores (Post 19).</p></li></ul><p><strong>Consistent PLS</strong> (PLSc) corrects these two biases using the reliability of each construct, which gives estimates close to CB-SEM (Dijkstra &amp; Henseler, 2015).</p><h2>3. Comparison on the Marketplace X data</h2><p>The model is the one from <a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a>: five constructs and seven hypotheses, H1 to H7. The data are the 254 respondents with complete answers, because cSEM and JASP require data without missing cells (how SmartPLS handles missing cells is covered in <a href="https://www.phuonglib.com/p/import-data-draw-model-smartpls-4">Post 33, "Importing Data and Drawing a Model in SmartPLS 4"</a>).</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/JA2MY/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/545d58d6-8487-4a2f-a519-dcc9c29d790a_1220x954.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fbcfdbb3-fe43-47d2-9e7d-7025bd147786_1220x1112.png&quot;,&quot;height&quot;:523,&quot;title&quot;:&quot;Table 1. Standardized path coefficients from four estimation approaches&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/JA2MY/1/" width="730" height="523" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!8XVt!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!8XVt!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!8XVt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png" width="2100" height="1000" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1000,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:118659,&quot;alt&quot;:&quot;Figure 2. Path coefficients from four estimation approaches&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Path coefficients from four estimation approaches" title="Figure 2. Path coefficients from four estimation approaches" srcset="https://substackcdn.com/image/fetch/$s_!8XVt!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 424w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 848w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 1272w, https://substackcdn.com/image/fetch/$s_!8XVt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6de18590-a084-4162-ae29-557a2406f2ca_2100x1000.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. Standardized path coefficients from four estimation approaches.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!d8kY!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!d8kY!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 424w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 848w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 1272w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!d8kY!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png" width="2100" height="880" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:880,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:182356,&quot;alt&quot;:&quot;Figure 3. Loadings and R&#178;&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Loadings and R&#178;" title="Figure 3. Loadings and R&#178;" srcset="https://substackcdn.com/image/fetch/$s_!d8kY!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 424w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 848w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 1272w, https://substackcdn.com/image/fetch/$s_!d8kY!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1877fde2-fef8-445b-ae68-0fb14899b2bf_2100x880.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. Standardized loadings of the 20 items and R&#178; of the three dependent constructs.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/ZwcXB/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7a59c786-72bd-4886-92ba-2085748ed0a8_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3271800c-e03f-4b3b-ba1e-50cbcfc80342_1220x666.png&quot;,&quot;height&quot;:302,&quot;title&quot;:&quot;Table 2. Loadings, R&#178; and SRMR under PLS-SEM, PLSc-SEM and CB-SEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/ZwcXB/1/" width="730" height="302" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>How to read the results:</p><ul><li><p><strong>PLS-SEM almost coincides with regression on mean scores</strong>: the path coefficients differ by at most 0.02. This makes sense: a composite with nearly equal weights is the mean score.</p></li><li><p><strong>Compared with CB-SEM, PLS-SEM gives smaller path coefficients</strong> (for example H1: 0.501 versus 0.579; H7: 0.280 versus 0.338), <strong>lower R&#178;</strong> (repurchase intention: 0.318 versus 0.444) and <strong>higher loadings</strong> (mean 0.809 versus 0.736). These are exactly the two biases described in Section 2.</p></li><li><p><strong>PLSc-SEM is close to CB-SEM</strong>: the path coefficients differ by at most 0.02, and the mean loadings are almost equal.</p></li><li><p><strong>The test conclusions are the same.</strong> Bootstrapping PLS-SEM with 5,000 samples: H4 is not significant (<em>p</em> = 0.086; 95% confidence interval [&#8722;0.017, 0.210]), and the other six hypotheses are significant, as in CB-SEM (H4: <em>p</em> = 0.443). The <em>p</em> values of PLS-SEM are smaller on some paths, for example H5: 0.007 versus 0.027.</p></li></ul><p>The lesson: with reflective constructs such as those in the Marketplace X model, PLS-SEM and CB-SEM usually lead to <strong>the same conclusions about direction and significance</strong>, but the <strong>size</strong> of the coefficients differs in a systematic way. When you compare your results with earlier studies, check which method those studies used.</p><h3>Estimating the model in JASP</h3><p>You can reproduce the PLS-SEM column for free in JASP:</p><ol><li><p>Open <code>marketplace_x_sem_complete.sav</code> in JASP.</p></li><li><p>Choose <strong>SEM &gt; Partial Least Squares SEM</strong> (Figure 4).</p></li><li><p>Type the model syntax below in the model box and press Cmd + Enter to apply it (the box shows this hint).</p></li><li><p>Expand <strong>Estimation</strong>, untick <strong>Consistent partial least squares</strong> and keep <strong>Inner weighting scheme</strong> at <strong>Path</strong> (Figure 5).</p></li><li><p>Expand <strong>Output</strong> and tick <strong>R-squared</strong>.</p></li></ol><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE   =~ PV1 + PV2 + PV3 + PV4
SATISF  =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST   =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT  =~ RI1 + RI2 + RI3 + RI4
VALUE   ~ CONFIRM
SATISF  ~ CONFIRM + VALUE
INTENT  ~ CONFIRM + VALUE + SATISF + TRUST</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!1V26!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!1V26!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 424w, https://substackcdn.com/image/fetch/$s_!1V26!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 848w, https://substackcdn.com/image/fetch/$s_!1V26!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 1272w, https://substackcdn.com/image/fetch/$s_!1V26!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!1V26!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png" width="440" height="326" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:326,&quot;width&quot;:440,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:63021,&quot;alt&quot;:&quot;Figure 4. SEM menu in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. SEM menu in JASP" title="Figure 4. SEM menu in JASP" srcset="https://substackcdn.com/image/fetch/$s_!1V26!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 424w, https://substackcdn.com/image/fetch/$s_!1V26!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 848w, https://substackcdn.com/image/fetch/$s_!1V26!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 1272w, https://substackcdn.com/image/fetch/$s_!1V26!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3b6f576-7450-4f49-aed9-284f578ae36c_440x326.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. The SEM menu of JASP 0.99.0: (1) Partial Least Squares SEM.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!L_Yp!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!L_Yp!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 424w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 848w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 1272w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!L_Yp!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png" width="1400" height="1014" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1014,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:372439,&quot;alt&quot;:&quot;Figure 5. Partial Least Squares SEM in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Partial Least Squares SEM in JASP" title="Figure 5. Partial Least Squares SEM in JASP" srcset="https://substackcdn.com/image/fetch/$s_!L_Yp!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 424w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 848w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 1272w, https://substackcdn.com/image/fetch/$s_!L_Yp!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26301abc-3bc6-427a-b818-f7bff115383b_1400x1014.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. The Marketplace X model in JASP: (1) untick Consistent partial least squares to get PLS-SEM as in SmartPLS, (2) the path weighting scheme, (3) the R-squared option, (4) the R-squared table, which matches the PLS-SEM column of the second table above (0.2506, 0.3986, 0.3176).</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!P47d!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!P47d!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 424w, https://substackcdn.com/image/fetch/$s_!P47d!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 848w, https://substackcdn.com/image/fetch/$s_!P47d!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 1272w, https://substackcdn.com/image/fetch/$s_!P47d!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!P47d!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png" width="620" height="365" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:365,&quot;width&quot;:620,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:74943,&quot;alt&quot;:&quot;Figure 6. Path coefficients in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Path coefficients in JASP" title="Figure 6. Path coefficients in JASP" srcset="https://substackcdn.com/image/fetch/$s_!P47d!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 424w, https://substackcdn.com/image/fetch/$s_!P47d!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 848w, https://substackcdn.com/image/fetch/$s_!P47d!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 1272w, https://substackcdn.com/image/fetch/$s_!P47d!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F26175f6a-67f3-40d0-8f97-d3649002aaae_620x365.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. The Regression Coefficients table of the JASP output: (1) the path coefficients, which match the PLS-SEM column of the first table above. JASP sorts the predictors of INTENT as CONFIRM, TRUST, VALUE, SATISF. The f&#178; and VIF columns are explained in <a href="https://www.phuonglib.com/p/evaluate-structural-model-smartpls">Post 37, "Evaluating the Structural Model in SmartPLS 4"</a>.</figcaption></figure></div><h2>4. PLS-SEM or CB-SEM?</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!PA27!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!PA27!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 424w, https://substackcdn.com/image/fetch/$s_!PA27!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 848w, https://substackcdn.com/image/fetch/$s_!PA27!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 1272w, https://substackcdn.com/image/fetch/$s_!PA27!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!PA27!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png" width="2100" height="1040" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1040,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:157012,&quot;alt&quot;:&quot;Figure 7. Choosing a method&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Choosing a method" title="Figure 7. Choosing a method" srcset="https://substackcdn.com/image/fetch/$s_!PA27!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 424w, https://substackcdn.com/image/fetch/$s_!PA27!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 848w, https://substackcdn.com/image/fetch/$s_!PA27!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 1272w, https://substackcdn.com/image/fetch/$s_!PA27!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9937fd5-b0d2-4447-8164-e1aff29f79d6_2100x1040.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. Three questions for choosing an estimation method.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/i16nM/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/56170aad-1665-4f22-a0b2-7c727b25d669_1220x898.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/41710bb4-efac-4b05-bfe0-8ec14fc20e23_1220x968.png&quot;,&quot;height&quot;:450,&quot;title&quot;:&quot;Table 3. When to use PLS-SEM, PLSc-SEM or CB-SEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/i16nM/1/" width="730" height="450" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Two reasons that are <strong>not sufficient</strong> for choosing PLS-SEM, even though they are very common in theses:</p><ul><li><p><strong>"The sample is small."</strong> PLS-SEM runs with small samples, but that does not make the estimates more accurate; a small sample still gives low statistical power. Hair et al. (2019) themselves warn against using a small sample as the main reason, and R&#246;nkk&#246; et al. (2016) strongly criticize this argument. The sample size should be determined by a power analysis or the inverse square root method (Hair et al., 2022; <a href="https://www.phuonglib.com/p/sample-size-calculation">Post 4, "How Large Should Your Sample Be? Formulas and Rules of Thumb for a Thesis"</a>).</p></li><li><p><strong>"The data are not normal."</strong> CB-SEM has ways to deal with non-normal data: robust estimation (MLR) and the Bollen-Stine bootstrap (<a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">Post 22, "Confirmatory Factor Analysis (CFA) in JASP"</a>).</p></li></ul><p>The debate about PLS-SEM continues. Its supporters stress its role in prediction and in composite models (Hair et al., 2019; Sarstedt et al., 2016). Its critics argue that PLS-SEM is often used for common factor constructs, where it gives biased estimates (R&#246;nkk&#246; et al., 2016). The safe approach in a thesis: state why you chose the method based on <strong>how you understand the constructs and on the research goal</strong>, not on sample size or distribution.</p><h2>5. Software for PLS-SEM</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/y341u/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1d6e90c3-d10a-47aa-8441-5809efb6d29e_1220x654.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d0efbada-5e0e-44d5-98a7-4654e603839d_1220x724.png&quot;,&quot;height&quot;:326,&quot;title&quot;:&quot;Table 4. Software for PLS-SEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/y341u/1/" width="730" height="326" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>This series follows the workflow in SmartPLS 4 and cross-checks it with JASP or R, so that readers without a SmartPLS license can still practice.</p><h2>6. Roadmap of the SmartPLS series</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/blZyG/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/745b45de-679f-4740-b12d-6dfa4c942595_1220x1118.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/b8978879-df97-40ad-b3b7-bc808c380589_1220x1188.png&quot;,&quot;height&quot;:560,&quot;title&quot;:&quot;Table 5. Roadmap of the SmartPLS series&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/blZyG/1/" width="730" height="560" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><h2>Frequently asked questions</h2><p><strong>SmartPLS and AMOS give different results. Which should I report?</strong></p><p>Report the results of the method you chose and explain why you chose it. Do not run both and then pick the "nicer" results. If you want to check robustness, report both, as in Section 3.</p><p><strong>Do I need model fit with PLS-SEM?</strong></p><p>PLS-SEM has no chi-square test like CB-SEM. SmartPLS reports SRMR; Hair et al. (2022) advise caution when using SRMR for PLS-SEM and recommend focusing on R&#178;, f&#178;, Q&#178; and predictive power.</p><p><strong>Should I choose PLS or PLSc in SmartPLS?</strong></p><p>If your constructs are reflective and you care about the size of the coefficients, PLSc gives less biased estimates. If the goal is prediction or the constructs are composites, use PLS. State your choice in the methods section.</p><p><strong>Can JASP replace SmartPLS?</strong></p><p>For estimating the model, bootstrapping and PLSpredict, yes, because JASP uses the cSEM package. Some tools, such as drawing the model with the mouse or IPMA, are not available in the PLS-SEM module of JASP 0.99.0 (the module has only the sections Model, Estimation, Output and Prediction).</p><div><hr></div><h2>References</h2><p>Dijkstra, T. K., &amp; Henseler, J. (2015). Consistent partial least squares path modeling. <em>MIS Quarterly, 39</em>(2), 297&#8211;316. https://doi.org/10.25300/MISQ/2015/39.2.02</p><p>Hair, J. F., Hult, G. T. M., Ringle, C. M., &amp; Sarstedt, M. (2022). <em>A primer on partial least squares structural equation modeling (PLS-SEM)</em> (3rd ed.). SAGE Publications.</p><p>Hair, J. F., Risher, J. J., Sarstedt, M., &amp; Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. <em>European Business Review, 31</em>(1), 2&#8211;24. https://doi.org/10.1108/EBR-11-2018-0203</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Rademaker, M. E., &amp; Schuberth, F. (2026). <em>cSEM: Composite-based structural equation modeling</em> (R package version 0.7.1) [Computer software]. https://CRAN.R-project.org/package=cSEM</p><p>Ringle, C. M., Wende, S., &amp; Becker, J.-M. (2026). <em>SmartPLS 4</em> (Version 4.1.2.1) [Computer software]. SmartPLS GmbH. https://www.smartpls.com</p><p>R&#246;nkk&#246;, M., McIntosh, C. N., Antonakis, J., &amp; Edwards, J. R. (2016). Partial least squares path modeling: Time for some serious second thoughts. <em>Journal of Operations Management, 47&#8211;48</em>, 9&#8211;27. https://doi.org/10.1016/j.jom.2016.05.002</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p><p>Sarstedt, M., Hair, J. F., Ringle, C. M., Thiele, K. O., &amp; Gudergan, S. P. (2016). Estimation issues with PLS and CBSEM: Where the bias lies! <em>Journal of Business Research, 69</em>(10), 3998&#8211;4010. https://doi.org/10.1016/j.jbusres.2016.06.007</p><p>Wold, H. (1982). Soft modeling: The basic design and some extensions. In K. G. J&#246;reskog &amp; H. Wold (Eds.), <em>Systems under indirect observation: Causality, structure, prediction</em> (Part II, pp. 1&#8211;54). North-Holland.</p>]]></content:encoded></item><item><title><![CDATA[Structural Equation Model (SEM) in JASP: Specify, Run and Read Results]]></title><description><![CDATA[Turn a CFA into a structural model in free JASP with lavaan syntax, compare its fit and read the paths and R-square]]></description><link>https://www.phuonglib.com/p/structural-equation-model-sem-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/structural-equation-model-sem-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:47:45 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!sW2h!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!sW2h!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!sW2h!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:647668,&quot;alt&quot;:&quot;Stone water channels carrying water from two springs through basins into one orange pool, small figures adjusting sluice gates.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Stone water channels carrying water from two springs through basins into one orange pool, small figures adjusting sluice gates." title="Stone water channels carrying water from two springs through basins into one orange pool, small figures adjusting sluice gates." srcset="https://substackcdn.com/image/fetch/$s_!sW2h!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!sW2h!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6df8bfd8-fcd4-47c5-80fe-3d24cc56198f_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>The SEM replaces the 10 factor covariances with 7 hypothesized paths and keeps the covariance between the two exogenous factors, so the degrees of freedom rise from 160 to 162.</p></li><li><p>With FIML, add a line that fixes the factor means at 0; otherwise JASP 0.99.0 reports 157 degrees of freedom instead of 162.</p></li><li><p>The structural constraints cost little fit: chi-square rises by 4.38 with 2 degrees of freedom (p = 0.112).</p></li><li><p>Six of the seven paths are significant; the direct path from confirmation to repurchase intention (H4) is not.</p></li><li><p>The model explains 33% of the variance in perceived value, 51% in satisfaction and 44% in repurchase intention.</p></li></ul></blockquote><div><hr></div><p>Posts <a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">22</a> to <a href="https://www.phuonglib.com/p/convergent-discriminant-validity-cfa-jasp">25</a> showed that the measurement model holds up: the five scales fit the data and are reliable, convergent (except the AVE of Repurchase intention) and discriminant. Following the two-step approach (Anderson &amp; Gerbing, 1988), we can now move to the second step: the <strong>structural model</strong>, where the two-headed arrows between the factors are replaced by <strong>directional paths</strong> that state the hypotheses.</p><p>This post shows how to turn the CFA model into an SEM, run it and read the overall results: model fit, path coefficients and R&#178;. <a href="https://www.phuonglib.com/p/hypothesis-testing-sem-jasp">Post 27, "Hypothesis Testing in SEM with JASP"</a> goes deeper into testing each hypothesis and interpreting the standardized coefficients.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch26-structural-equation-model-sem">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The screenshots and result tables come from JASP 0.99.0 (JASP Team, 2026) on macOS, run on the file <code>marketplace_x_sem.sav</code> from <a href="https://www.phuonglib.com/p/prepare-connect-data-spss-to-jasp">Post 21, "Preparing and Connecting Data from SPSS to JASP"</a> (267 respondents, missing cells kept). The numbers were checked with the R package lavaan (Rosseel, 2012) in the script <code>structural_model.R</code> (folder <code>scripts/</code> of the repository).</p></blockquote><h2>1. From CFA to SEM: what changes?</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!NIn1!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!NIn1!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 424w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 848w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 1272w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!NIn1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png" width="2200" height="980" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:980,&quot;width&quot;:2200,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:235780,&quot;alt&quot;:&quot;Figure 1. From the measurement model to the structural model&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. From the measurement model to the structural model" title="Figure 1. From the measurement model to the structural model" srcset="https://substackcdn.com/image/fetch/$s_!NIn1!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 424w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 848w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 1272w, https://substackcdn.com/image/fetch/$s_!NIn1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe335f3a6-ae39-464d-b8ab-a607fbe2e8a4_2200x980.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Left: CFA, every pair of factors is freely correlated. Right: SEM, the correlations are replaced by seven hypothesized paths.</figcaption></figure></div><h3>Exogenous and endogenous variables</h3><ul><li><p>An <strong>exogenous variable</strong> has no arrow pointing at it: CONFIRM and TRUST. The model does not explain them; it only uses them to explain other variables.</p></li><li><p>An <strong>endogenous variable</strong> has at least one arrow pointing at it: VALUE, SATISF and INTENT. Each endogenous variable needs a <strong>structural residual</strong> (d, disturbance): the part of its variance that the variables before it do not explain.</p></li></ul><p>SATISF is both an outcome (of CONFIRM and VALUE) and a cause (of INTENT). Ordinary regression cannot do this in one equation, but SEM estimates all the equations at once (<a href="https://www.phuonglib.com/p/what-is-amos-cb-sem">Post 19, "What Is AMOS? CB-SEM and When to Use It Instead of SPSS"</a>).</p><h3>The seven hypotheses</h3><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/I0RAf/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/55fcd62d-c2fc-4de2-8acc-6c25ea27990c_1220x694.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a8f22380-6fd9-46a7-a1f1-4a57fe7807e8_1220x764.png&quot;,&quot;height&quot;:377,&quot;title&quot;:&quot;Table 1. The seven hypotheses of the structural model.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/I0RAf/1/" width="730" height="377" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p><strong>Note on numbering.</strong> Posts <a href="https://www.phuonglib.com/p/quantitative-research-process-7-steps">2</a> and <a href="https://www.phuonglib.com/p/multiple-linear-regression-pspp-spss">14</a> use six hypotheses, in which H4 is Perceived value to Repurchase intention, H5 is Satisfaction to Repurchase intention and H6 is Trust to Repurchase intention. From this post on, the model adds a direct path from Confirmation to Repurchase intention (to test the direct effect alongside the indirect paths through Perceived value and Satisfaction, <a href="https://www.phuonglib.com/p/mediation-analysis-bootstrap-jasp">Post 28, "Mediation Analysis with JASP"</a>), so this path takes the number H4 and the next three hypotheses move to H5 (Perceived value), H6 (Satisfaction) and H7 (Trust); H1 to H3 stay the same.</p><h3>Where do the extra degrees of freedom come from?</h3><p>The CFA has 10 covariances among the five factors. The SEM replaces them with 7 paths and <strong>keeps 1 covariance</strong>, between the two exogenous factors CONFIRM and TRUST. The SEM therefore has 2 parameters fewer than the CFA: the two relations that are not hypothesized, TRUST to VALUE and TRUST to SATISF (the dashed orange lines in Figure 1). The degrees of freedom rise from 160 to <strong>162</strong>.</p><p><strong>Important rule:</strong> always let the exogenous variables correlate, unless theory says they are independent (Kline, 2023). AMOS <strong>does not draw</strong> this covariance for you; forgetting it is a very common mistake (Section 5).</p><h2>2. Specify the structural model in JASP</h2><p>The <em>Factor</em> module only runs CFA. The structural model needs the <strong>SEM</strong> module (switch it on with the plus button at the right end of the toolbar, Post 22):</p><ol><li><p><strong>Open</strong> <code>marketplace_x_sem.sav</code> and choose <strong>SEM &gt; Structural Equation Modeling</strong>.</p></li><li><p><strong>Type</strong> the model in the editor in lavaan syntax, then press <strong>Cmd+Enter</strong> (Figure 2):</p></li></ol><pre><code># Measurement model
CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE =~ PV1 + PV2 + PV3 + PV4
SATISF =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT =~ RI1 + RI2 + RI3 + RI4
# Structural model
VALUE ~ CONFIRM
SATISF ~ CONFIRM + VALUE
INTENT ~ CONFIRM + VALUE + SATISF + TRUST
CONFIRM ~~ TRUST
# Fix the factor means at 0 (as in AMOS)
CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!a_y-!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!a_y-!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 424w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 848w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 1272w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!a_y-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png" width="760" height="372" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:372,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:98108,&quot;alt&quot;:&quot;Figure 2. The SEM model written in lavaan syntax in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. The SEM model written in lavaan syntax in JASP" title="Figure 2. The SEM model written in lavaan syntax in JASP" srcset="https://substackcdn.com/image/fetch/$s_!a_y-!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 424w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 848w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 1272w, https://substackcdn.com/image/fetch/$s_!a_y-!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a31fce6-5b66-47d3-9fe4-0d9b79056d25_760x372.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) Measurement model, (2) structural model and the covariance between the two exogenous factors, (3) the line that fixes the factor means.</figcaption></figure></div><p>Three operators to remember:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/2tqtO/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/273f103e-0a4d-463c-99df-9a4aaf552f21_1220x468.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8e8e56fa-4472-4c64-b72a-ab1d3435ced7_1220x538.png&quot;,&quot;height&quot;:243,&quot;title&quot;:&quot;Table 2. lavaan operators and their AMOS equivalents.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/2tqtO/1/" width="730" height="243" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>lavaan <strong>adds a residual automatically</strong> to every endogenous variable, so you do not need to declare d1, d2 and d3. The latent variable names must be spelled exactly as in the measurement part; comment lines start with <code>#</code>.</p><p><strong>The last line is required with FIML.</strong> As in Post 22, the SEM module of JASP 0.99.0 also estimates the factor means when FIML is used, which makes the degrees of freedom come out wrong (157 instead of 162). Fixing the factor means at 0, as AMOS and the CFA module do, gives the correct degrees of freedom and chi-square.</p><ol><li><p><strong>Check</strong> that <em>Missing data handling</em> under <em>Estimation Options</em> is <strong>FIML</strong> (the default of the SEM module when the data have missing cells; the output notes that "Missing data method 'FIML' forces meanstructure").</p></li><li><p><strong>Tick</strong> under <em>Output Options</em> (Figure 3): <em>Additional fit measures</em>, <em>R-squared</em>, <em>Standardized estimates</em> (choose <em>All</em>) and <em>Path diagram</em> with <em>Show parameter estimates</em> and <em>Standardized</em>.</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Z9Oa!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 424w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 848w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 1272w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png" width="760" height="431" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:431,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:124678,&quot;alt&quot;:&quot;Figure 3. Output options in the SEM module&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Output options in the SEM module" title="Figure 3. Output options in the SEM module" srcset="https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 424w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 848w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 1272w, https://substackcdn.com/image/fetch/$s_!Z9Oa!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4e3f1d88-fe51-420d-ae70-7a9f5c692f9d_760x431.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. (1) R&#178; of the endogenous variables, (2) standardized estimates, (3) path diagram with standardized estimates.</figcaption></figure></div><h2>3. Read the results</h2><h3>Model fit: SEM compared with CFA</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!W_tJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!W_tJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 424w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 848w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 1272w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!W_tJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png" width="900" height="673" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/a04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:673,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:166000,&quot;alt&quot;:&quot;Figure 4. Fit of the SEM model in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Fit of the SEM model in JASP" title="Figure 4. Fit of the SEM model in JASP" srcset="https://substackcdn.com/image/fetch/$s_!W_tJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 424w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 848w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 1272w, https://substackcdn.com/image/fetch/$s_!W_tJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa04c8064-c89b-41aa-bf78-4588b272b5eb_900x673.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. Top: the Model Fit table, (1) chi-square, df and p of the structural model. Bottom: Fit Indices, (2) CFI and TLI, (3) RMSEA with its 90% confidence interval, (4) SRMR.</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/CHWmu/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fea05bbb-4733-4dd4-b070-40f0ba24df0a_1220x546.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e9cf30ca-9c60-477c-96ad-14539ccf832c_1220x654.png&quot;,&quot;height&quot;:321,&quot;title&quot;:&quot;Table 3. Fit of the CFA and the SEM.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/CHWmu/1/" width="730" height="321" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>A structural model <strong>cannot fit better</strong> than the CFA, because it is the CFA with extra constraints (Anderson &amp; Gerbing, 1988). The right question is whether those constraints cost much fit. Comparing the two nested models, the chi-square rises by 4.38 with 2 degrees of freedom, <em>p</em> = 0.112 (computed in lavaan from the unrounded values 134.29 and 138.67). The two omitted relations (Trust to Value and Trust to Satisfaction) are <strong>not needed</strong> to explain the data. The other indices hardly change.</p><p>The standardized loadings in the SEM differ from the CFA by at most 0.002. If the loadings changed noticeably when moving to the SEM, that would signal a misspecified structural model that is "pulling" the measurement model (Kline, 2023).</p><h3>Path coefficients</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!Eeda!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!Eeda!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 424w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 848w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 1272w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!Eeda!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png" width="900" height="344" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/daf76a65-b404-42df-a079-ec4c851bf657_900x344.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:344,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:109046,&quot;alt&quot;:&quot;Figure 5. Standardized regression coefficients in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Standardized regression coefficients in JASP" title="Figure 5. Standardized regression coefficients in JASP" srcset="https://substackcdn.com/image/fetch/$s_!Eeda!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 424w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 848w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 1272w, https://substackcdn.com/image/fetch/$s_!Eeda!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdaf76a65-b404-42df-a079-ec4c851bf657_900x344.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. The Regression Coefficients table with standardization set to All: (1) the Std. estimate column holds the standardized path coefficients, (2) their p-values.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ka2_!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ka2_!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 424w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 848w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 1272w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ka2_!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png" width="2100" height="1240" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1240,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:193200,&quot;alt&quot;:&quot;Figure 6. Structural model results&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Structural model results" title="Figure 6. Structural model results" srcset="https://substackcdn.com/image/fetch/$s_!ka2_!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 424w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 848w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 1272w, https://substackcdn.com/image/fetch/$s_!ka2_!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6036fd50-2532-4985-8540-5d7ba1bcd49a_2100x1240.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. Standardized path coefficients, the p-values JASP prints for them and R&#178; of the three endogenous variables (redrawn from the JASP output).</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/YMFWS/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/20846ee0-79e9-452b-bd0c-bfdcd45ae1de_1220x986.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/6cf9f57f-b68c-473b-993d-26953238766c_1220x1122.png&quot;,&quot;height&quot;:530,&quot;title&quot;:&quot;Table 4. Standardized path coefficients.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/YMFWS/1/" width="730" height="530" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Six of the seven paths are significant at the 0.05 level; H4 is not. With standardization set to All, JASP tests the <strong>standardized</strong> coefficients, so its z and p differ slightly from the test of the unstandardized coefficients that AMOS reports as C.R. (for example H5: <em>p</em> = 0.011 here, 0.015 for the unstandardized coefficient). The conclusions are the same. This result <strong>agrees</strong> with the regression on mean scores in Post 14 in direction and significance, but the standardized coefficients in the SEM are larger because measurement error has been removed (Post 19). Post 27 interprets each hypothesis, compares the size of the coefficients and shows how to write the conclusions.</p><p><strong>In AMOS.</strong> The unstandardized coefficients are under <strong>Estimates &gt; Scalars &gt; Regression Weights</strong> in the Text Output (the same table as the loadings; the rows look like <code>VALUE &lt;--- CONFIRM</code>), with C.R. as the z-value; the standardized coefficients are under <strong>Standardized Regression Weights</strong> (Byrne, 2016). With the AMOS settings (ML, FIML, N = 267), the unstandardized estimates, with S.E., C.R. and <em>p</em>, are: H1 0.627 (0.087; 7.22; &lt; 0.001); H2 0.412 (0.085; 4.87; &lt; 0.001); H3 0.389 (0.077; 5.02; &lt; 0.001); H4 0.088 (0.099; 0.89; 0.375); H5 0.208 (0.085; 2.44; 0.015); H6 0.212 (0.095; 2.23; 0.026); H7 0.316 (0.071; 4.46; &lt; 0.001). The standardized coefficients are the same as in the table above.</p><p>The covariance between CONFIRM and TRUST corresponds to a correlation of 0.39 (<em>p</em> &lt; 0.001, Factor Covariances table), which confirms that letting the two exogenous factors correlate is necessary.</p><h3>R&#178; of the endogenous variables</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!XQZh!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!XQZh!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 424w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 848w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 1272w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!XQZh!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png" width="300" height="990" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:990,&quot;width&quot;:300,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:104591,&quot;alt&quot;:&quot;Figure 7. R-square in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. R-square in JASP" title="Figure 7. R-square in JASP" srcset="https://substackcdn.com/image/fetch/$s_!XQZh!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 424w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 848w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 1272w, https://substackcdn.com/image/fetch/$s_!XQZh!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F214ec6c4-b069-4029-95a2-8780ec6cb14a_300x990.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. The R-Squared table: (1) R&#178; of the three endogenous variables. The rows above are the R&#178; of each item (as in the CFA).</figcaption></figure></div><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/xUa7x/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f636a64c-61ca-400f-b8f0-7145e09982f2_1220x590.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/ead22bcc-6650-478b-b69f-f009c4c2b1ec_1220x660.png&quot;,&quot;height&quot;:293,&quot;title&quot;:&quot;Table 5. R&#178; of the endogenous variables.&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/xUa7x/1/" width="730" height="293" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The standardized residual variance equals 1 &#8722; R&#178; (for example 0.557 for INTENT, Factor Variances table). The R&#178; of INTENT (0.443) is clearly higher than in the regression of Post 14 (0.315), again because SEM removes measurement error.</p><h3>The JASP path diagram</h3><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!IuKy!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!IuKy!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 424w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 848w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 1272w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!IuKy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png" width="900" height="558" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:558,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:201412,&quot;alt&quot;:&quot;Figure 8. Path diagram of the structural model drawn by JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Path diagram of the structural model drawn by JASP" title="Figure 8. Path diagram of the structural model drawn by JASP" srcset="https://substackcdn.com/image/fetch/$s_!IuKy!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 424w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 848w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 1272w, https://substackcdn.com/image/fetch/$s_!IuKy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c5d4af2-cf71-421f-af35-28b0b3e1c22c_900x558.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. The diagram JASP draws with standardized estimates. Factor names are shortened (CONFI, VAL, SATI, TRUS, INT).</figcaption></figure></div><p>The JASP diagram (drawn with the semPlot package) is good enough to check the specification: the right number of arrows, the right directions and a two-headed arrow between CONFI and TRUS (0.39). For a thesis, redraw it more cleanly, as in Figure 6.</p><h2>4. Indirect and total effects</h2><p>With <em>Standardized estimates</em> ticked, JASP also prints the <em>Indirect Effects</em> and <em>Total Effects</em> tables, with standard errors computed by the delta method.</p><p>CONFIRM affects INTENT through three indirect paths: via VALUE (0.135, <em>p</em> = 0.014), via SATISF (0.090, <em>p</em> = 0.038) and via VALUE then SATISF (0.053, <em>p</em> = 0.038). The direct effect is small and not significant (0.09), but the <strong>total effect</strong> is 0.37 (<em>p</em> &lt; 0.001, 95% CI [0.231, 0.508]). This is a sign of <strong>mediation</strong>: confirmation of expectations affects repurchase intention mainly through perceived value and satisfaction. A rigorous test of indirect effects needs bootstrapping, the topic of Post 28.</p><h2>5. A common mistake: forgetting the covariance between exogenous variables</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!hGek!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!hGek!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 424w, https://substackcdn.com/image/fetch/$s_!hGek!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 848w, https://substackcdn.com/image/fetch/$s_!hGek!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 1272w, https://substackcdn.com/image/fetch/$s_!hGek!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!hGek!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png" width="2200" height="760" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/dfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:760,&quot;width&quot;:2200,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:115746,&quot;alt&quot;:&quot;Figure 9. Fit when the covariance is left out&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Fit when the covariance is left out" title="Figure 9. Fit when the covariance is left out" srcset="https://substackcdn.com/image/fetch/$s_!hGek!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 424w, https://substackcdn.com/image/fetch/$s_!hGek!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 848w, https://substackcdn.com/image/fetch/$s_!hGek!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 1272w, https://substackcdn.com/image/fetch/$s_!hGek!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbb1c10-c0eb-48e2-a8e1-5b9f6bf8ec1a_2200x760.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. Chi-square, CFI, RMSEA and SRMR of the CFA, the correct SEM and the SEM without the covariance between CONFIRM and TRUST (lavaan).</figcaption></figure></div><p>If the two-headed arrow between CONFIRM and TRUST is left out, the model assumes that the two concepts are <strong>uncorrelated</strong>. On the sample data, the chi-square rises by 30.32 with 1 degree of freedom (<em>p</em> &lt; 0.001) and <strong>SRMR rises to 0.084, above the 0.08 cutoff</strong>, while CFI (0.997) and RMSEA (0.012) still look excellent. This illustrates the advice in <a href="https://www.phuonglib.com/p/model-fit-indices-sem-jasp">Post 23, "SEM Model Fit Indices in JASP"</a>: read several indices. SRMR is sensitive to this kind of error because it measures the gap between correlations directly.</p><p>lavaan lets <strong>latent</strong> exogenous variables correlate by default, so the line <code>CONFIRM ~~ TRUST</code> is in fact redundant in JASP. The post keeps it to make the model explicit and to match the AMOS diagram. AMOS users must draw this arrow themselves.</p><h2>6. Reporting the results</h2><blockquote><p>The structural model fit the data: chi-square = 138.67, df = 162, <em>p</em> = 0.908; CFI = 1.000; TLI = 1.013; RMSEA = 0.000 (90% CI [0.000, 0.012]); SRMR = 0.035. Compared with the measurement model, the structural model did not significantly worsen fit (chi-square difference = 4.38 with 2 degrees of freedom, <em>p</em> = 0.112). The model explained 33.3% of the variance in perceived value, 51.4% in satisfaction and 44.3% in repurchase intention.</p></blockquote><p>The table of path coefficients and the diagram with standardized coefficients are reported together with the conclusion for each hypothesis in Post 27.</p><h2>Five common mistakes</h2><ol><li><p><strong>Running the SEM before evaluating the measurement model.</strong> Path coefficients cannot be trusted if the scales do not hold up.</p></li><li><p><strong>Forgetting the residual of an endogenous variable</strong> (AMOS reports an error) or <strong>forgetting the covariance between exogenous variables</strong> (AMOS reports no error; the model just fits worse).</p></li><li><p><strong>Adding paths because of modification indices</strong> to improve fit without a hypothesis.</p></li><li><p><strong>Skipping the comparison with the CFA.</strong> SEM fit must be read relative to the CFA.</p></li><li><p><strong>Reporting only the significant coefficients</strong> and leaving out rejected hypotheses such as H4.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Should I run SEM on mean scores (path analysis) instead of latent variables?</strong></p><p>Path analysis on mean scores is simultaneous regression and still ignores measurement error, as in Post 14. With multi-item scales that have passed a CFA, full SEM is the better choice; path analysis suits samples that are small relative to the number of parameters.</p><p><strong>Unstandardized or standardized coefficients?</strong></p><p>Report both, or at least the standardized coefficients with <em>p</em>. The significance test that AMOS reports (C.R., <em>p</em>) is based on the unstandardized coefficients. Post 27 discusses this in detail.</p><p><strong>How high should R&#178; be?</strong></p><p>There is no general cutoff; it depends on the field. In consumer behavior research, an R&#178; of 0.3 to 0.5 for the main outcome is common. Chin (1998) suggests 0.19, 0.33 and 0.67 as weak, moderate and substantial, but these levels were proposed for PLS-SEM.</p><p><strong>What about control variables (gender, income)?</strong></p><p>Add them as observed exogenous variables, draw arrows to the outcome and let them correlate with the other exogenous variables. For a nominal variable with several groups, use dummy variables (<a href="https://www.phuonglib.com/p/dummy-variables-moderation-regression-pspp-spss">Post 16, "Dummy Variables and Moderator Variables in Regression with PSPP (SPSS)"</a>) or multigroup analysis (<a href="https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp">Post 29, "Multigroup Analysis and Moderation with JASP"</a>).</p><div><hr></div><h2>References</h2><p>Anderson, J. C., &amp; Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. <em>Psychological Bulletin, 103</em>(3), 411&#8211;423. https://doi.org/10.1037/0033-2909.103.3.411</p><p>Byrne, B. M. (2016). <em>Structural equation modeling with AMOS: Basic concepts, applications, and programming</em> (3rd ed.). Routledge.</p><p>Chin, W. W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (Ed.), <em>Modern methods for business research</em> (pp. 295&#8211;336). Lawrence Erlbaum Associates.</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Kline, R. B. (2023). <em>Principles and practice of structural equation modeling</em> (5th ed.). Guilford Press.</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p>]]></content:encoded></item><item><title><![CDATA[Preparing SPSS Data for CFA and SEM in JASP, a Free AMOS Alternative]]></title><description><![CDATA[Check value ranges, count missing data, run Little's MCAR test, choose FIML and open a clean .sav file in free JASP]]></description><link>https://www.phuonglib.com/p/prepare-connect-data-spss-to-jasp</link><guid isPermaLink="false">https://www.phuonglib.com/p/prepare-connect-data-spss-to-jasp</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:47:35 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/fe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!btZr!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!btZr!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!btZr!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!btZr!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!btZr!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!btZr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:786898,&quot;alt&quot;:&quot;A cook sorts vegetables on a sunny kitchen counter before cooking, setting aside a few gaps marked by small orange tags.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="A cook sorts vegetables on a sunny kitchen counter before cooking, setting aside a few gaps marked by small orange tags." title="A cook sorts vegetables on a sunny kitchen counter before cooking, setting aside a few gaps marked by small orange tags." srcset="https://substackcdn.com/image/fetch/$s_!btZr!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!btZr!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!btZr!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!btZr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe6b9b8e-fb89-4c8d-9297-564b35b096dd_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Most errors in a first CFA or SEM run come from the data file: out-of-range values, reverse-worded items that were never recoded, wrong variable types or clashing names.</p></li><li><p>In the sample data, 254 of the 267 respondents answered all 20 items, and the 21 missing cells are only 0.39% of the data.</p></li><li><p>Little's MCAR test (chi-square = 223.26, df = 239, <em>p</em> = 0.760) gives no evidence that the data are missing in a systematic way.</p></li><li><p>FIML, listwise deletion and mean substitution give almost identical results here, but FIML stays the default and mean substitution should not be used.</p></li><li><p>Declare the items as Scale in SPSS so that JASP estimates with ML, and use Sync Data in JASP when the <code>.sav</code> file changes.</p></li></ul></blockquote><div><hr></div><p>In <a href="https://www.phuonglib.com/p/first-model-jasp-free-amos-alternative">Post 20, "Your First Model in JASP"</a>, the first model ran after changing the variable types in JASP and choosing FIML. In practice, most errors in a first CFA or SEM run do not come from the model but from the <strong>data file</strong>: variable names that clash with latent variables, reverse-worded items that were never recoded, variable types that make the software use the wrong estimator or blank cells handled by a default you did not intend. This post shows how to prepare a clean data file for CFA and SEM, decide how to handle missing data and then open and sync that file in JASP.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch21-prepare-data-amos-jasp">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The data preparation part uses PSPP 2.1.2 (Free Software Foundation, 2026) with real output from the syntax file <code>prepare_sem_data.sps</code> (in the <code>scripts/</code> folder of the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch21-prepare-data-amos-jasp">repository</a>; Figures 2 and 3 come from <code>figure_tables.sps</code>, which runs the same commands with short labels so the tables stay narrow). The missing-data analysis uses R and the lavaan package (Rosseel, 2012), file <code>compare_missing_data.R</code>. The JASP screenshots were taken in JASP 0.99.0 on macOS (JASP Team, 2026).</p></blockquote><h2>1. Six steps to prepare the data file</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!9sP2!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!9sP2!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 424w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 848w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 1272w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!9sP2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png" width="2000" height="740" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:740,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:124058,&quot;alt&quot;:&quot;Figure 1. Six steps to prepare the data&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Six steps to prepare the data" title="Figure 1. Six steps to prepare the data" srcset="https://substackcdn.com/image/fetch/$s_!9sP2!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 424w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 848w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 1272w, https://substackcdn.com/image/fetch/$s_!9sP2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7009a96e-e171-43ba-a361-c3b5e0c9f37f_2000x740.png 1456w" sizes="100vw"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Six steps to prepare the data file in SPSS/PSPP before taking it to JASP or AMOS.</figcaption></figure></div><h3>Step 1. Keep the valid responses</h3><p>Use only the responses that passed the cleaning in <a href="https://www.phuonglib.com/p/data-cleaning-missing-values-outliers-spss-pspp">Post 8, "Data Cleaning in PSPP (SPSS)"</a>: drop those that failed the screening question or the attention check, or gave the same answer to every item. The file <code>marketplace_x_analysis.sav</code> keeps <strong>267 responses</strong>.</p><h3>Step 2. Check the value range</h3><p>Every Likert item must lie between 1 and 5. A single value of 55 that slips through (like the error in item SAT2 found in Post 8) distorts the whole covariance matrix, and SEM software gives no warning about it.</p><pre><code>DESCRIPTIVES VARIABLES=CONF1 CONF2 CONF3 CONF4 PV1 PV2 PV3 PV4
    SAT1 SAT2 SAT3 SAT4 TRU1 TRU2 TRU3 TRU4R RI1 RI2 RI3 RI4
  /STATISTICS=MEAN STDDEV MIN MAX.</code></pre><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!dDmR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!dDmR!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 424w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 848w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 1272w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!dDmR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png" width="560" height="554" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/f8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:554,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:106088,&quot;alt&quot;:&quot;Figure 2. Descriptive statistics of the 20 items&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Descriptive statistics of the 20 items" title="Figure 2. Descriptive statistics of the 20 items" srcset="https://substackcdn.com/image/fetch/$s_!dDmR!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 424w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 848w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 1272w, https://substackcdn.com/image/fetch/$s_!dDmR!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff8672ff0-94d9-42be-bd73-358a1684cc9f_560x554.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. Output in PSPP. The Minimum and Maximum columns show that every item lies between 1 and 5; the N column shows how many people answered each item.</figcaption></figure></div><p>The <em>N</em> column tells you how many respondents answered each item: from 264 (CONF3, RI1) to 267. Note the two rows at the bottom of the table in PSPP 2.1.2: <em>Valid N (listwise)</em> shows 267 and <em>Missing N (listwise)</em> shows 13. The 267 does not match the number of respondents who answered all 20 items (only 254, see Step 4), while 13 equals the number of respondents with at least one blank item. To get an exact count, this post uses the <code>COUNT</code> command.</p><p>(In <code>prepare_sem_data.sps</code> the items keep their full labels and value labels, so PSPP 2.1.2 prints the item wording as row names and shows Minimum and Maximum as the labels "Strongly disagree" and "Strongly agree"; the numbers are the same as in Figure 2.)</p><h3>Step 3. Reverse-code the reverse-worded item</h3><p>Item TRU4 ("I worry that marketplace X does not keep its promises") runs in the opposite direction from the other Trust items. Use <strong>TRU4R</strong> (reverse-coded in <a href="https://www.phuonglib.com/p/recode-compute-composite-scores-pspp-spss">Post 10, "Recoding Variables and Computing Composite Scores in PSPP (SPSS)"</a>) in the model, not TRU4. If TRU4 goes into the model by mistake, its loading will be negative and the reliability of the scale drops sharply, as seen with Cronbach's alpha in <a href="https://www.phuonglib.com/p/cronbach-alpha-spss-pspp">Post 11, "Cronbach's Alpha in PSPP (SPSS)"</a>.</p><h3>Step 4. Count the missing data</h3><pre><code>COUNT N_MISSING = CONF1 CONF2 CONF3 CONF4 PV1 PV2 PV3 PV4
    SAT1 SAT2 SAT3 SAT4 TRU1 TRU2 TRU3 TRU4R RI1 RI2 RI3 RI4 (MISSING).
FREQUENCIES N_MISSING.</code></pre><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!NfWf!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 424w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 848w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 1272w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!NfWf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png" width="560" height="196" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:196,&quot;width&quot;:560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:38654,&quot;alt&quot;:&quot;Figure 3. Number of blank items per respondent&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Number of blank items per respondent" title="Figure 3. Number of blank items per respondent" srcset="https://substackcdn.com/image/fetch/$s_!NfWf!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 424w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 848w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 1272w, https://substackcdn.com/image/fetch/$s_!NfWf!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6869159-dc89-4b8e-8bd8-1e9cfdf71243_560x196.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 3. Distribution of the number of blank items (out of the 20 model items) per respondent.</figcaption></figure></div><ul><li><p><strong>254 respondents</strong> (95.1%) answered all 20 items.</p></li><li><p><strong>5 respondents</strong> left 1 item blank, <strong>8 respondents</strong> left 2 items blank.</p></li><li><p>In total there are <strong>21 missing cells</strong> out of 5,340 (267 x 20), that is <strong>0.39%</strong>.</p></li></ul><h3>Step 5. Use names that do not clash</h3><p>The file <code>marketplace_x_analysis.sav</code> contains the columns CONF, PV, SAT, TRU and RI (the mean scores created in Post 10). The latent variables in this series are named CONFIRM, VALUE, SATISF, TRUST and INTENT, so there is no clash; but if you keep mean-score columns, you must not give those names to factors, because lavaan (inside JASP) reports a name clash (Post 20). The simplest fix is to <strong>drop the mean-score columns</strong> from the file used for CFA and SEM. Also:</p><ul><li><p><strong>Use plain names without spaces or special characters</strong> (CONF1, TRU4R), at most a few dozen characters long.</p></li><li><p><strong>Convert user-defined missing codes</strong> (for example 9 or 99 meaning "no answer") into blank cells (system-missing). The sample file has no such codes, but this is a common mistake when data are entered in Excel.</p></li></ul><h3>Step 6. Save a lean file for CFA and SEM</h3><pre><code>SAVE OUTFILE='../data/marketplace_x_sem.sav'
  /KEEP=ID GENDER OCCUP INCOME AGEGROUP
        CONF1 CONF2 CONF3 CONF4 PV1 PV2 PV3 PV4 SAT1 SAT2 SAT3 SAT4
        TRU1 TRU2 TRU3 TRU4R RI1 RI2 RI3 RI4 N_MISSING.</code></pre><p>The syntax file <code>prepare_sem_data.sps</code> creates the three files in Table 1; all three are in the <code>data</code> folder of the repository.</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/TuHwv/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1eb6bf9e-7de5-48e4-a4c6-e4537e1f18a4_1220x724.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3c1f81a5-98b4-4d64-9fc0-267a3c2b8a24_1220x794.png&quot;,&quot;height&quot;:342,&quot;title&quot;:&quot;Table 1. The three data files for CFA and SEM&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/TuHwv/1/" width="730" height="342" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The grouping variables (GENDER, OCCUP, INCOME, AGEGROUP) are kept for the multigroup analysis in <a href="https://www.phuonglib.com/p/multigroup-analysis-moderation-jasp">Post 29, "Multigroup Analysis and Moderation with JASP"</a>. The syntax file also declares the 20 items as quantitative variables (<code>VARIABLE LEVEL ... (SCALE)</code>). This step <strong>matters for JASP</strong>: JASP reads the variable type from the <code>.sav</code> file, and if the items are <em>Ordinal</em>, JASP estimates with DWLS instead of ML (Post 20). Declaring the type in SPSS saves you from changing it by hand in JASP.</p><h2>2. Are the data missing at random?</h2><p>How to handle missing data depends on the <strong>mechanism</strong> behind it (Enders, 2022):</p><ul><li><p><strong>MCAR</strong> (missing completely at random): missingness is unrelated to any variable. For example, a respondent skips a line by accident.</p></li><li><p><strong>MAR</strong> (missing at random): missingness is related to variables that <em>are</em> in the data. For example, older respondents skip questions about the mobile app more often.</p></li><li><p><strong>MNAR</strong> (missing not at random): missingness is related to the missing value itself. For example, very dissatisfied respondents deliberately skip the satisfaction questions.</p></li></ul><p><strong>Little's MCAR test</strong> (Little, 1988) compares the means of groups of respondents who share the same missing-data pattern. The null hypothesis H0 is that the data are missing completely at random. JASP, AMOS and PSPP do not offer this test; SPSS has it in the <em>Missing Value Analysis</em> module. The file <code>compare_missing_data.R</code> computes it directly from Little's formula:</p><blockquote><p>Chi-square = 223.26; df = 239; <em>p</em> = 0.760.</p></blockquote><p><em>p</em> is larger than 0.05, so we <strong>do not reject</strong> the MCAR hypothesis: there is no evidence that the data are missing in any systematic way. Note that this test can only detect departures from MCAR; it cannot prove that the data are MCAR, and it cannot tell MAR from MNAR.</p><h2>3. Choose how to handle missing data</h2><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!hrD3!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!hrD3!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 424w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 848w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 1272w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!hrD3!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png" width="2000" height="1060" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/d4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1060,&quot;width&quot;:2000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:162007,&quot;alt&quot;:&quot;Figure 4. Choosing how to handle missing data&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Choosing how to handle missing data" title="Figure 4. Choosing how to handle missing data" srcset="https://substackcdn.com/image/fetch/$s_!hrD3!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 424w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 848w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 1272w, https://substackcdn.com/image/fetch/$s_!hrD3!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4e2d299-4b55-459e-8853-2f075a764eef_2000x1060.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. How to choose a missing-data method when running CFA and SEM. The "Yes" branch is a limitation of AMOS only; JASP does not have it.</figcaption></figure></div><p>There are three common methods:</p><ol><li><p><strong>FIML</strong> (full information maximum likelihood). Estimates the model from all the information available for each respondent, without dropping anyone. This is the recommended method when data are MCAR or MAR (Enders, 2022). In JASP: check that <em>Missing data handling</em> in the <strong>Advanced</strong> section of the analysis is set to <em>FIML</em>.</p></li><li><p><strong>Listwise deletion.</strong> Drops every respondent with at least one blank cell. Simple and unbiased when data are MCAR, but it loses sample size.</p></li><li><p><strong>Mean substitution.</strong> Fills each blank cell with the mean of that item. This shrinks variances and correlations, so it <strong>should not be used</strong> (Enders, 2022).</p></li></ol><h3>Comparing the three methods on the sample data</h3><p>The file <code>compare_missing_data.R</code> fits the five-factor measurement model (20 items, analyzed in detail in <a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">Post 22, "Confirmatory Factor Analysis (CFA) in JASP"</a>) in each of the three ways (Table 2).</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/DFJcf/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/afe24cb9-9564-4648-b2b3-263970b2d58f_1220x570.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3ac2f387-a1ad-49b8-a829-7966c1c218b6_1220x690.png&quot;,&quot;height&quot;:315,&quot;title&quot;:&quot;Table 2. Three ways to handle missing data on the sample data&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/DFJcf/1/" width="730" height="315" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>With only 0.39% missing cells and MCAR data, the three methods give almost identical results: the standardized loadings differ by at most 0.014 between listwise deletion and FIML. This conclusion <strong>does not carry over</strong> to a high missing rate (usually 5% or more) or to data that are not missing at random: then a poor choice can bias the coefficients and standard errors noticeably.</p><h3>A note on AMOS</h3><p>AMOS uses FIML when you tick <em>Estimate means and intercepts</em>, but when the data have blank cells AMOS does not provide modification indices, bootstrapping or some indices such as GFI and AGFI (Byrne, 2016). AMOS users must then run the model on the complete-case file (<code>marketplace_x_sem_complete.sav</code>) or impute the data with <strong>Analyze &gt; Data Imputation</strong>, which offers <em>Regression imputation</em>, <em>Stochastic regression imputation</em> and <em>Bayesian imputation</em> (prefer the last two, because they keep the natural variability of the data). JASP (through lavaan) computes modification indices with FIML, so the JASP series uses <strong>FIML with the 267-respondent file</strong> throughout. The 254-respondent file is still useful to check the robustness of the results.</p><p>When you write your thesis, state which method you chose and why.</p><h2>4. Open and sync the data in JASP</h2><h3>Open the prepared file</h3><ol><li><p><strong>Click</strong> the three-line menu icon, choose <strong>Open &gt; Computer &gt; Browse</strong> and select <code>marketplace_x_sem.sav</code> (as in Post 20).</p></li><li><p><strong>Check</strong> the type icon at the top of the item columns: it must be the <strong>ruler</strong> (Scale), thanks to the <code>VARIABLE LEVEL</code> declaration in Step 6 (Figure 5, box 1). There is no need to change the variable type by hand as in Post 20.</p></li><li><p><strong>Check</strong> the blank cells: JASP shows a missing value as a dot (Figure 5, box 2).</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!HdEJ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!HdEJ!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 424w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 848w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 1272w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!HdEJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png" width="1000" height="480" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:480,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:147655,&quot;alt&quot;:&quot;Figure 5. The prepared data opened in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. The prepared data opened in JASP" title="Figure 5. The prepared data opened in JASP" srcset="https://substackcdn.com/image/fetch/$s_!HdEJ!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 424w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 848w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 1272w, https://substackcdn.com/image/fetch/$s_!HdEJ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F25887b77-68eb-4dec-8e2c-03730a1d867a_1000x480.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. (1) Item CONF1 is already Scale when the file opens, (2) the blank CONF1 cell of respondent 13 is shown as a dot (JASP 0.99.0 on macOS also prints "nan" in grey where the value label would be).</figcaption></figure></div><h3>Check the missing data with Descriptives</h3><p><strong>Choose</strong> <strong>Descriptives &gt; Descriptive Statistics</strong>, move CONF1 to RI4 into the <em>Variables</em> box and tick <em>Transpose descriptives table</em> so that each item is on its own row.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!2pLz!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!2pLz!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 424w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 848w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 1272w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!2pLz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png" width="1000" height="391" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/fe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:391,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:152027,&quot;alt&quot;:&quot;Figure 6. Descriptive statistics and missing cells in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Descriptive statistics and missing cells in JASP" title="Figure 6. Descriptive statistics and missing cells in JASP" srcset="https://substackcdn.com/image/fetch/$s_!2pLz!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 424w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 848w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 1272w, https://substackcdn.com/image/fetch/$s_!2pLz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe564663-70f2-49f3-b252-5acdd2664b50_1000x391.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. (1) Transpose descriptives table, (2) the Missing column gives the number of blank cells for each item and matches the PSPP output in Figure 2.</figcaption></figure></div><p>The <em>Missing</em> column shows that CONF3 and RI1 have the most blank cells (3 each); the total is 21, the same as in PSPP. The <em>Valid</em> column (264 to 267) equals the <em>N</em> column of PSPP, and the means agree after rounding (for example CONF1: 3.875 in JASP, 3.88 in PSPP).</p><h3>Sync when the data change</h3><p>JASP keeps a copy of the data inside the <code>.jasp</code> file. If you later change the <code>.sav</code> file in SPSS (add responses, fix data-entry errors), JASP can <strong>sync</strong> the data and rerun every analysis:</p><ol><li><p><strong>Click</strong> the three-line menu icon and choose <strong>Sync Data</strong> (Figure 7, box 1).</p></li><li><p><strong>Click</strong> the name of the source file under <em>Current File</em> (Figure 7, box 2), or press Cmd+Y (Ctrl+Y on Windows).</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!BOz1!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!BOz1!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 424w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 848w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 1272w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!BOz1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png" width="900" height="393" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:393,&quot;width&quot;:900,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:67233,&quot;alt&quot;:&quot;Figure 7. Syncing data in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. Syncing data in JASP" title="Figure 7. Syncing data in JASP" srcset="https://substackcdn.com/image/fetch/$s_!BOz1!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 424w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 848w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 1272w, https://substackcdn.com/image/fetch/$s_!BOz1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14109c42-5d69-4f63-ae22-5a5d421bbed4_900x393.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 7. (1) The Sync Data menu, (2) the source file linked to the analysis.</figcaption></figure></div><p>If you rename a variable in SPSS, that variable disappears from the analyses that use it in JASP; you have to add the renamed variable again.</p><h3>If you use AMOS</h3><p>In AMOS, open the prepared file with <strong>File &gt; Data Files</strong> (Ctrl+D): click <em>File Name</em>, choose <code>marketplace_x_sem.sav</code>, check that the <em>N</em> column shows <strong>267/267</strong> and click <em>OK</em>. If the file is open in SPSS Statistics, <strong>Analyze &gt; IBM SPSS Amos</strong> opens AMOS and connects it to the saved file on disk. <strong>View &gt; Variables in Dataset</strong> (Ctrl+Shift+D) should list CONF1 to RI4 (with TRU4R instead of TRU4) and no longer list CONF, PV, SAT, TRU and RI. AMOS reads the <code>.sav</code> file every time it runs, so after a change you only save the file in SPSS and run the model again; a renamed variable must also be renamed in the AMOS diagram.</p><h3>Save a .jasp file</h3><p>After opening and checking the data, <strong>save</strong> your work with Cmd+S (Ctrl+S on Windows), for example as <code>post21-data.jasp</code>. You can open this file in the next posts to run CFA (Post 22) and SEM (<a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a>), or start again from <code>marketplace_x_sem.sav</code> in the <code>data</code> folder.</p><h2>5. Five common mistakes when connecting data</h2><ol><li><p><strong>Keeping a mean-score column with the same name as a factor.</strong> lavaan in JASP reports a name clash.</p></li><li><p><strong>Using an item that was not reverse-coded.</strong> The loading is negative and reliability drops.</p></li><li><p><strong>Leaving the items as Ordinal.</strong> JASP estimates with DWLS instead of ML, and the results do not match AMOS or the posts that use ML.</p></li><li><p><strong>Using 9 or 99 for blank cells</strong> without converting them to system-missing. These values are counted as real answers.</p></li><li><p><strong>Leaving Missing data handling at Listwise deletion.</strong> JASP then drops the 13 respondents with blank cells without telling you; choose FIML in the Advanced section of each analysis.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Should I just delete respondents with blank cells to keep things simple?</strong></p><p>With a very low missing rate and MCAR data, as in this sample, listwise deletion gives almost the same results as FIML. FIML is still the better default because it keeps the sample size and stays unbiased when data are MAR. Use the complete-case file to check the robustness of your results.</p><p><strong>How much missing data is too much?</strong></p><p>There is no absolute threshold. Many authors consider less than 5% low. The mechanism matters more than the rate: 3% missing under MNAR can cause more bias than 10% missing completely at random.</p><p><strong>Do I need to standardize the data (convert to z-scores) before using JASP?</strong></p><p>No. JASP works with the raw data and computes standardized coefficients when you choose <em>Standardization = All</em>.</p><p><strong>Does JASP read value labels (Agree, Disagree, ...)?</strong></p><p>Yes, JASP shows the labels next to the numeric values in the data view. Labels do not affect CFA or SEM results when the variables are Scale.</p><p><strong>Can JASP read Excel files?</strong></p><p>JASP reads <code>.csv</code>, <code>.txt</code>, <code>.sav</code>, <code>.dta</code>, <code>.ods</code> and some other formats. For Excel, save the sheet as <code>.csv</code> first. The file <code>marketplace_x_sem.csv</code> in the repository is ready to use.</p><div><hr></div><h2>References</h2><p>Byrne, B. M. (2016). <em>Structural equation modeling with AMOS: Basic concepts, applications, and programming</em> (3rd ed.). Routledge.</p><p>Enders, C. K. (2022). <em>Applied missing data analysis</em> (2nd ed.). Guilford Press.</p><p>Free Software Foundation. (2026). <em>GNU PSPP</em> (Version 2.1.2) [Computer software]. https://www.gnu.org/software/pspp/</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. <em>Journal of the American Statistical Association, 83</em>(404), 1198&#8211;1202. https://doi.org/10.1080/01621459.1988.10478722</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p>]]></content:encoded></item><item><title><![CDATA[Common Method Bias in JASP: Harman's Test and the Common Latent Factor]]></title><description><![CDATA[Check common method bias in free JASP with Harman's single-factor test and a common latent factor, and read honestly what each check can show]]></description><link>https://www.phuonglib.com/p/common-method-bias-jasp-harman-clf</link><guid isPermaLink="false">https://www.phuonglib.com/p/common-method-bias-jasp-harman-clf</guid><dc:creator><![CDATA[PHUONGLIB]]></dc:creator><pubDate>Sun, 22 Nov 2026 13:47:31 GMT</pubDate><enclosure url="https://substack-post-media.s3.amazonaws.com/public/images/8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ZXdx!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ZXdx!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!ZXdx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp" width="2560" height="1440" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1440,&quot;width&quot;:2560,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:690816,&quot;alt&quot;:&quot;Twenty glass bottles on a long table tinted by orange light from one shared window pane, a small figure inspecting them with a magnifying glass.&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/webp&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Twenty glass bottles on a long table tinted by orange light from one shared window pane, a small figure inspecting them with a magnifying glass." title="Twenty glass bottles on a long table tinted by orange light from one shared window pane, a small figure inspecting them with a magnifying glass." srcset="https://substackcdn.com/image/fetch/$s_!ZXdx!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 424w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 848w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 1272w, https://substackcdn.com/image/fetch/$s_!ZXdx!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a47a638-4158-4178-9d9d-a4ccee8608fa_2560x1440.webp 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><blockquote><p><strong>Key takeaways</strong></p><ul><li><p>Prevention through questionnaire design matters more than any test run after data collection.</p></li><li><p>Harman's test (30.6% with principal axis, 33.9% with PCA) only rules out a single dominant factor, not common method bias as such.</p></li><li><p>A one-factor CFA fits far worse than the five-factor model (CFI 0.638 against 1.000).</p></li><li><p>The common latent factor needs equal loadings to converge and does not improve the fit (<em>p</em> = 0.749); prefer the chi-square difference over the <em>z</em> test of its loading.</p></li><li><p>On data simulated without method variance the CLF still "finds" 7% to 9%, so the key question is whether the path coefficients change; here they do not.</p></li></ul></blockquote><div><hr></div><p>In the Marketplace X study, every respondent answered all 20 items on confirmation of expectations, perceived value, satisfaction, trust and repurchase intention, in the same survey, on the same 5-point Likert scale. A reviewer will ask: are the correlations between the constructs <strong>inflated</strong> simply because they were all measured in the same way? That is the question of <strong>common method variance</strong> and <strong>common method bias</strong> (CMB).</p><p>This post shows the two checks used most often in theses: <strong>Harman's single-factor test</strong> and the <strong>common latent factor</strong> (CLF). Just as important, it shows what each check <strong>can and cannot tell you</strong>, using real results on the sample data.</p><blockquote><p><strong>Sample data and scripts.</strong> The data files and scripts for this post are in the <a href="https://github.com/phuonglib/research-data-analysis/tree/main/ch30-common-method-bias-harman-clf">research-data-analysis repository on GitHub</a> (folder <code>data/</code> and <code>scripts/</code>). Download the whole repository with <strong>Code &gt; Download ZIP</strong>.</p></blockquote><blockquote><p><strong>About the figures and numbers.</strong> The screenshots come from JASP 0.99.0 (JASP Team, 2026) on macOS, and the result tables are taken from JASP's <em>Export Results</em>. The numbers were checked with the lavaan package for R (Rosseel, 2012), using the script <code>cmb.R</code> in the repository. Where JASP did not print a number in this run (the eigenvalues of all 20 components, the standardized CLF loadings, the complete-case results, two-decimal chi-square differences), the number comes from <code>cmb.R</code> and the text says so. The syntax file <code>harman.sps</code> runs the same Harman test in SPSS or PSPP (Free Software Foundation, 2026).</p></blockquote><h2>1. What is common method variance?</h2><p>Every answer score has three parts: a part due to the construct being measured, a part due to <strong>the way it is measured</strong> and random error. When two constructs are measured in the same way, their method parts can correlate and add to the correlation between the two constructs (Podsakoff et al., 2003).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!xbqO!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!xbqO!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!xbqO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png" width="2100" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:148076,&quot;alt&quot;:&quot;Figure 1. Common method variance&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 1. Common method variance" title="Figure 1. Common method variance" srcset="https://substackcdn.com/image/fetch/$s_!xbqO!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!xbqO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F104c4c78-16af-4111-bf32-6af6a1243842_2100x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 1. Common method variance adds to the true correlation between two constructs.</figcaption></figure></div><p>Method variance <strong>does not always inflate</strong> coefficients: with several predictors it can also deflate them (Podsakoff et al., 2012). Nor is it <strong>certain to occur</strong> just because a self-report questionnaire was used. The right question is therefore: is there evidence of method variance <strong>large enough to change the conclusions</strong>?</p><p><strong>Prevention matters more than testing.</strong> The best approach is to design the questionnaire to reduce the risk from the start (Podsakoff et al., 2012):</p><ul><li><p><strong>Separate</strong> the predictors and the outcomes: different times, different sources (for example, repurchase taken from actual purchase records) or at least different parts of the questionnaire.</p></li><li><p><strong>Use</strong> different scale formats for some constructs.</p></li><li><p><strong>Guarantee</strong> anonymity and state clearly that there are no right or wrong answers, to reduce socially desirable responding.</p></li><li><p><strong>Write</strong> clear, short, non-leading items; mix in reverse-worded items such as TRU4 (<a href="https://www.phuonglib.com/p/recode-compute-composite-scores-pspp-spss">Post 10, "Recoding Variables (Recode) and Computing Composite Scores (Compute Mean) in PSPP (SPSS)"</a>).</p></li></ul><p>The tests below are checks made <strong>after</strong> the data have been collected. They complement prevention; they do not replace it.</p><h2>2. Harman's single-factor test</h2><p>The idea (Harman, 1976; Podsakoff et al., 2003): if method variance dominates, entering all items into an <strong>unrotated</strong> factor analysis will give <strong>a single factor</strong>, or one factor that explains <strong>most</strong> of the variance. The threshold often used in theses is that the first factor explains <strong>less than 50%</strong> of the variance.</p><p>There are two extraction methods, and they give different numbers:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/q0Gt6/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/98a5cd98-d075-4fd2-a4c7-fcb3f1a3c068_1220x388.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/85493f28-61e5-4279-ba12-94a131dc084b_1220x546.png&quot;,&quot;height&quot;:267,&quot;title&quot;:&quot;Table 1. Share of variance of the first factor under two extraction methods&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/q0Gt6/1/" width="730" height="267" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>Report which method you used. Here both lead to the same conclusion.</p><h3>Steps in JASP</h3><ol><li><p><strong>Open</strong> <code>marketplace_x_sem.sav</code> and choose <em>Factor &gt; Exploratory Factor Analysis</em>.</p></li><li><p><strong>Move</strong> the 20 model items (CONF1 to RI4, with TRU4R instead of TRU4) into the <em>Variables</em> box.</p></li><li><p><strong>Choose</strong> <em>Number of Factors &gt; Manual</em> and set <em>Number of factors</em> = <strong>1</strong> (Figure 2, box 1).</p></li><li><p><strong>Choose</strong> <em>Rotation Method &gt; Orthogonal</em> and <strong>none</strong> (box 2). JASP's default is <em>Oblique / promax</em>.</p></li><li><p><strong>Choose</strong> <em>Factoring Method = Principal axis factoring</em> (box 3).</p></li><li><p><strong>Open</strong> <em>Output Options</em>: drag <em>Display loadings above</em> down to <strong>0</strong> so that every loading is shown, tick <em>Scree plot</em> and <em>Parallel analysis results</em> and choose <em>Missing Values &gt; Exclude cases listwise</em> (Figure 3).</p></li></ol><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!CvrF!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!CvrF!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 424w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 848w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 1272w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!CvrF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png" width="1400" height="784" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:784,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:245702,&quot;alt&quot;:&quot;Figure 2. Harman's test in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 2. Harman's test in JASP" title="Figure 2. Harman's test in JASP" srcset="https://substackcdn.com/image/fetch/$s_!CvrF!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 424w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 848w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 1272w, https://substackcdn.com/image/fetch/$s_!CvrF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45c19f4a-1c66-4dbf-833c-4a615f5ced31_1400x784.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 2. (1) One factor forced, (2) no rotation, (3) principal axis factoring, (4) proportion of variance of the factor: 0.3058, that is 30.6%.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!FBPl!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!FBPl!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 424w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 848w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 1272w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!FBPl!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png" width="760" height="865" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:865,&quot;width&quot;:760,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:144680,&quot;alt&quot;:&quot;Figure 3. Output options of the EFA&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 3. Output options of the EFA" title="Figure 3. Output options of the EFA" srcset="https://substackcdn.com/image/fetch/$s_!FBPl!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 424w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 848w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 1272w, https://substackcdn.com/image/fetch/$s_!FBPl!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5433d455-6420-4eab-841b-1d19bb2be0ba_760x865.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 3. (1) Display loadings above 0, (2) scree plot with parallel analysis, (3) listwise exclusion, so that the 254 respondents with complete data are used.</figcaption></figure></div><p>The <em>Factor Loadings</em> table shows that all 20 items load on the single factor (0.319 for TRU3 to 0.687 for CONF4); the four trust items have the lowest loadings. The <em>Chi-Squared Test</em> table (896.138, df = 170, <em>p</em> &lt; .001) already rejects the one-factor model, but Harman's test is read from the proportion of variance.</p><p>In this run JASP did not print a table with the eigenvalues of all components, as SPSS does. The scree plot (Figure 4) shows a first eigenvalue of about 6.8, that is 6.8 / 20 = 33.9% (the PCA figure). For the full table, use <em>Factor &gt; Principal Component Analysis</em>, the syntax file <code>harman.sps</code> in SPSS or PSPP or the script <code>cmb.R</code>; Figure 5 is drawn from the eigenvalues computed by <code>cmb.R</code>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!bt5c!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!bt5c!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 424w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 848w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 1272w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!bt5c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png" width="620" height="467" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:467,&quot;width&quot;:620,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:53090,&quot;alt&quot;:&quot;Figure 4. Scree plot&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 4. Scree plot" title="Figure 4. Scree plot" srcset="https://substackcdn.com/image/fetch/$s_!bt5c!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 424w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 848w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 1272w, https://substackcdn.com/image/fetch/$s_!bt5c!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F65fdad80-861b-4e57-a9d7-ffcc0160e271_620x467.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 4. (1) The first eigenvalue is about 6.8; (2) the 4th and 5th eigenvalues are above 1 but just below the line of simulated data from parallel analysis.</figcaption></figure></div><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!pYJ8!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!pYJ8!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!pYJ8!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png" width="2100" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:132401,&quot;alt&quot;:&quot;Figure 5. Result of Harman's test&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 5. Result of Harman's test" title="Figure 5. Result of Harman's test" srcset="https://substackcdn.com/image/fetch/$s_!pYJ8!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!pYJ8!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F792e24fe-5e36-479a-be63-6a3f15a210c2_2100x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 5. Percentage of variance of the 20 principal components, unrotated (eigenvalues computed in R).</figcaption></figure></div><p><strong>Reading the result.</strong> The first factor explains 33.9% (PCA) or 30.6% (principal axis), below the 50% threshold. Five components have an eigenvalue above 1, exactly the number of constructs in the model (parallel analysis in Figure 4 is stricter and keeps only the first three). There is no sign of a single common factor dominating the data.</p><p><strong>Limitations.</strong> Podsakoff et al. (2003), the very authors usually cited for this test, call it a <strong>weak test</strong>:</p><ul><li><p>It detects method variance only when it is <strong>extremely large</strong>, large enough to pull the 20 items into one factor. Moderate method variance can still bias coefficients without being detected.</p></li><li><p>The 50% threshold is a <strong>convention</strong> without a firm theoretical basis. Fuller et al. (2016) showed that method variance biases coefficients noticeably only when it is very large, and at that level Harman's test usually detects it. The test is therefore not useless, but it catches only <strong>severe</strong> cases.</p></li><li><p>A large first factor can also arise because the constructs are <strong>genuinely strongly correlated</strong> (here CONFIRM, VALUE and SATISF correlate 0.58 to 0.64), not because of the method.</p></li></ul><p>A measured conclusion is "Harman's test does not indicate that a single common factor dominates", <strong>not</strong> "the data have no CMB".</p><h2>3. Harman as a CFA: the one-factor model</h2><p>A stricter version compares two CFA models: all 20 items loading on <strong>one</strong> factor, against the <strong>five</strong>-factor model of <a href="https://www.phuonglib.com/p/confirmatory-factor-analysis-cfa-jasp">Post 22, "Confirmatory Factor Analysis (CFA) in JASP"</a>. If method variance dominates, the one-factor model will fit almost as well as the five-factor model.</p><p>In JASP, open <em>SEM &gt; Structural Equation Modeling</em> and use several model tabs (<em>Model 1</em>, <em>Model 2</em>, <em>Model 3</em>, added with the green plus button). JASP compares the models automatically in the same <em>Model Fit</em> table.</p><ul><li><p><strong>Model 1</strong> (five factors): the CFA syntax of Post 22, plus the line <code>CONFIRM + VALUE + SATISF + TRUST + INTENT ~ 0*1</code>.</p></li><li><p><strong>Model 2</strong> (one factor):</p></li></ul><pre><code>GENERAL =~ CONF1 + CONF2 + CONF3 + CONF4 + PV1 + PV2 + PV3 + PV4 + SAT1 + SAT2 + SAT3 + SAT4 + TRU1 + TRU2 + TRU3 + TRU4R + RI1 + RI2 + RI3 + RI4
GENERAL ~ 0*1</code></pre><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/XdFjt/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/af8fa9d1-7eac-4c7a-b91b-3debf93a2e53_1220x456.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/69c31712-d569-4ff9-bb9d-bc0a19713ab5_1220x592.png&quot;,&quot;height&quot;:272,&quot;title&quot;:&quot;Table 2. Fit of the one-factor and five-factor CFA models&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/XdFjt/1/" width="730" height="272" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The one-factor model fits <strong>very poorly</strong> (CFI about 0.64) and significantly worse than the five-factor model. A single common factor cannot explain the data. This version has the same limitation as Harman's test: it only rules out the case where method variance accounts for <strong>everything</strong>.</p><h2>4. The common latent factor (CLF)</h2><p>The CLF goes one step further: it keeps the five-factor model and <strong>adds</strong> one factor that loads on all 20 items (Podsakoff et al., 2003; Williams et al., 1989). This factor stands for the variance that all items share beyond the five constructs.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!qKAl!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!qKAl!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 424w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 848w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 1272w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!qKAl!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png" width="2100" height="960" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:960,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:236450,&quot;alt&quot;:&quot;Figure 6. Model with a common latent factor&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 6. Model with a common latent factor" title="Figure 6. Model with a common latent factor" srcset="https://substackcdn.com/image/fetch/$s_!qKAl!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 424w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 848w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 1272w, https://substackcdn.com/image/fetch/$s_!qKAl!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96e99407-b3a0-46cd-8632-06de94a6b113_2100x960.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 6. Five-factor CFA model with an added common latent factor (CLF).</figcaption></figure></div><p>Three constraints make the model estimable:</p><ol><li><p><strong>The CLF is uncorrelated</strong> with the five factors (covariances fixed at 0). If it were allowed to correlate, the model could not tell which part is the construct and which part is the method.</p></li><li><p><strong>The variance of the CLF is 1</strong>, to set its scale.</p></li><li><p><strong>The loadings on the CLF are equal</strong> for all 20 items (same label <em>m</em>). The version with 20 free loadings <strong>does not converge</strong> on the Marketplace X data: 20 method loadings plus 20 construct loadings for the items are too many for the data to separate. The equality constraint is the usual remedy, and it assumes that the method affects every item equally.</p></li></ol><h3>Steps in JASP</h3><p><strong>Model 3</strong> (five factors plus the CLF):</p><pre><code>CONFIRM =~ CONF1 + CONF2 + CONF3 + CONF4
VALUE   =~ PV1 + PV2 + PV3 + PV4
SATISF  =~ SAT1 + SAT2 + SAT3 + SAT4
TRUST   =~ TRU1 + TRU2 + TRU3 + TRU4R
INTENT  =~ RI1 + RI2 + RI3 + RI4
CLF =~ NA*CONF1 + m*CONF1 + m*CONF2 + m*CONF3 + m*CONF4 + m*PV1 + m*PV2 + m*PV3 + m*PV4 + m*SAT1 + m*SAT2 + m*SAT3 + m*SAT4 + m*TRU1 + m*TRU2 + m*TRU3 + m*TRU4R + m*RI1 + m*RI2 + m*RI3 + m*RI4
CLF ~~ 1*CLF
CLF ~~ 0*CONFIRM + 0*VALUE + 0*SATISF + 0*TRUST + 0*INTENT
CONFIRM + VALUE + SATISF + TRUST + INTENT + CLF ~ 0*1</code></pre><p><code>NA*CONF1</code> switches off the default rule that fixes the first loading at 1, so that all 20 loadings share the label <code>m</code> and are estimated freely; the scale of the CLF is set by <code>CLF ~~ 1*CLF</code>.</p><div class="captioned-image-container"><figure><div class="image-link image2" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!wDde!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 424w, https://substackcdn.com/image/fetch/$s_!wDde!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 848w, https://substackcdn.com/image/fetch/$s_!wDde!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 1272w, https://substackcdn.com/image/fetch/$s_!wDde!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!wDde!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png" width="1400" height="231" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:231,&quot;width&quot;:1400,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:152528,&quot;alt&quot;:&quot;Figure 7. CLF syntax&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 7. CLF syntax" title="Figure 7. CLF syntax" srcset="https://substackcdn.com/image/fetch/$s_!wDde!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 424w, https://substackcdn.com/image/fetch/$s_!wDde!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 848w, https://substackcdn.com/image/fetch/$s_!wDde!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 1272w, https://substackcdn.com/image/fetch/$s_!wDde!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5202d955-473a-43c2-bd91-ebd3305ba569_1400x231.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div></div></div></div><figcaption class="image-caption">Figure 7. (1) Three model tabs, (2) the CLF loads on the 20 items with the same label m (the line continues beyond the right edge of the editor; the full syntax is above), (3) the variance of the CLF is 1 and it is uncorrelated with the five factors.</figcaption></figure></div><p>Press <strong>Cmd+Enter</strong>. The <em>Model Fit</em> table orders the models by degrees of freedom and compares each model with the one listed above it (Figure 8). Model 3 has df = 159, one fewer than Model 1: this confirms that the 20 CLF loadings were held equal (one extra parameter, <em>m</em>).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!I5HM!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!I5HM!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 424w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 848w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 1272w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!I5HM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png" width="1000" height="269" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:269,&quot;width&quot;:1000,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:76161,&quot;alt&quot;:&quot;Figure 8. Comparing the three models in JASP&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 8. Comparing the three models in JASP" title="Figure 8. Comparing the three models in JASP" srcset="https://substackcdn.com/image/fetch/$s_!I5HM!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 424w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 848w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 1272w, https://substackcdn.com/image/fetch/$s_!I5HM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F027a8d10-7291-4e4d-afe9-0add4fc44e45_1000x269.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 8. (1) Model 1 (five factors) against Model 3 (with CLF): chi-square difference 0.1024 with 1 df, p = .749; (2) Model 2 (one factor) against Model 1: 800.8 with 10 df, p &lt; .001; (3) the n(Parameters) columns do not match the rows (see the note below).</figcaption></figure></div><p><strong>A display problem in JASP 0.99.0.</strong> In the <em>n(Parameters)</em> columns of this table, Model 3 shows 70 / 70, Model 1 shows 60 / 60 and Model 2 shows 90 / 71. These do not belong to the rows they are printed in. The <em>Additional Fit Measures</em> table of the same run (Figure 9, box 4) gives the correct number of free parameters: 70 for Model 1, 60 for Model 2 and 71 for Model 3, which is also what lavaan reports and what the degrees of freedom imply (230 sample moments, means included, minus 70, 60 and 71 gives df 160, 170 and 159). The <em>Free</em> column therefore seems to list the models in tab order (1, 2, 3) while the rows are sorted by df. The <em>Total</em> values (70, 60, 90) could not be reproduced with lavaan. The chi-squares, df and difference tests in the table are correct; only read the parameter counts from <em>Additional Fit Measures</em>.</p><p>Tick <em>Output Options &gt; Additional fit measures</em> to get CFI, RMSEA and SRMR for all three models (Figure 9), and <em>Standardized estimates</em> to get the standardized loadings (Table below).</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!EG9Q!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!EG9Q!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 424w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 848w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 1272w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!EG9Q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png" width="620" height="616" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e732643c-4c0c-4e79-901c-0119aa461d49_620x616.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:616,&quot;width&quot;:620,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:105481,&quot;alt&quot;:&quot;Figure 9. Fit indices of the three models&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 9. Fit indices of the three models" title="Figure 9. Fit indices of the three models" srcset="https://substackcdn.com/image/fetch/$s_!EG9Q!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 424w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 848w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 1272w, https://substackcdn.com/image/fetch/$s_!EG9Q!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe732643c-4c0c-4e79-901c-0119aa461d49_620x616.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 9. (1) CFI, (2) RMSEA, (3) SRMR and (4) number of free parameters of Model 1 (five factors), Model 2 (one factor) and Model 3 (with CLF); the rows in between are left out.</figcaption></figure></div><p>The standardized loadings of Model 3 were not captured from JASP for this post. The table below gives them from lavaan (<code>cmb.R</code>, FIML, N = 267), the same estimator that JASP's SEM module uses:</p><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/rvS5D/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/81a2fe76-6520-45cf-b0dc-f18bd10cab79_1220x1692.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/5b0b634f-cfea-47b4-9483-57ebdf11abb6_1220x1878.png&quot;,&quot;height&quot;:942,&quot;title&quot;:&quot;Table 3. Standardized loadings with and without the common latent factor&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/rvS5D/1/" width="730" height="942" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><h3>In AMOS</h3><p>Open a copy of the CFA file of Post 22 with the data file <code>marketplace_x_sem_complete.sav</code>. Draw an ellipse above the diagram, name it <strong>CLF</strong> and draw 20 single-headed arrows from it to the 20 items. Give the 20 arrows the common label <strong>m</strong> (<em>Object Properties &gt; Parameters &gt; Regression weight</em>) and type <strong>1</strong> in the <em>Variance</em> box of CLF. Do not draw two-headed arrows between CLF and the five factors: in AMOS, a covariance that is not drawn is fixed at 0. Click <em>Calculate Estimates</em> and read <em>Standardized Regression Weights</em> and <em>Model Fit</em>; to test the CLF, compare the chi-square with that of the five-factor CFA on the same data, or add a model with the constraint <code>m = 0</code> in <em>Manage Models</em>. Because AMOS uses the 254 complete respondents, its results correspond to the complete-case column of the next table. The one-factor model of Section 3 is drawn the same way: in a copy of the CFA file, replace the five ellipses with one ellipse GENERAL that points to all 20 items, with its first loading fixed at 1.</p><h2>5. Reading the CLF results</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/D0j7T/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/33c0706e-a92c-4c15-9260-39d60438a2f9_1220x826.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3c8f7451-0a28-4751-b024-dbffb3c38ac0_1220x1012.png&quot;,&quot;height&quot;:410,&quot;title&quot;:&quot;Table 4. Results of the common latent factor on two data sets&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/D0j7T/1/" width="730" height="410" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>There are three common ways to read these results, and they do not all point the same way:</p><ul><li><p><strong>Chi-square difference</strong>: adding the CLF <strong>does not improve</strong> the model (<em>p</em> = 0.749 and 0.676). The data do not need a method factor.</p></li><li><p><strong>Method variance</strong>: about 7% to 9%, far below the roughly 25% that Williams et al. (1989) reported for studies using self-report measures.</p></li><li><p><strong>Change in the construct loadings</strong> when the CLF is added: at most 0.063 (FIML) and 0.092 (complete cases). A common rule of thumb says that changes above 0.2 are a concern; these are much smaller. The rule has no clear statistical basis, so treat it as a rough guide only.</p></li></ul><p><strong>One point needs saying plainly: on the complete-case data, the z test of m gives p = 0.031.</strong> Looking only at that number, one might write "the CLF is significant, there is CMB". But the chi-square difference test on <strong>the same data</strong> gives <em>p</em> = 0.676. The two tests disagree because the z test relies on the standard error, and the standard error is unreliable when the likelihood surface is nearly flat (Figure 10). When the two tests disagree, <strong>use the chi-square difference test</strong>.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!oun1!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!oun1!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!oun1!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!oun1!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!oun1!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!oun1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png" width="2100" height="920" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:920,&quot;width&quot;:2100,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:202987,&quot;alt&quot;:&quot;Figure 10. Likelihood over the CLF loading&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="Figure 10. Likelihood over the CLF loading" title="Figure 10. Likelihood over the CLF loading" srcset="https://substackcdn.com/image/fetch/$s_!oun1!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 424w, https://substackcdn.com/image/fetch/$s_!oun1!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 848w, https://substackcdn.com/image/fetch/$s_!oun1!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 1272w, https://substackcdn.com/image/fetch/$s_!oun1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1c176aa2-16c0-4e84-96b7-a18e159a6a0f_2100x920.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-tqxFUX" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-tqxFUX" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Figure 10. Chi-square when the CLF loading is fixed at different values (FIML, N = 267).</figcaption></figure></div><p>Figure 10 shows why. Fixing the CLF loading at any value from 0 to about 0.4 changes the chi-square of the CFA by less than 1 unit. The data <strong>cannot distinguish</strong> "no method variance" from "about 10% method variance". The estimate of 0.26 is just the lowest point of an almost flat region, so it should not be taken as evidence of CMB.</p><p><strong>The sample data were generated without method variance.</strong> The Marketplace X data set is simulated: the items were generated from five constructs and independent errors, with no method factor at all. Yet the CLF still "finds" 7% to 9% method variance. This is a direct illustration of the point made by Richardson et al. (2009): the CLF technique easily mistakes variance shared <strong>between constructs</strong> for method variance and should not be used on its own to reach a conclusion.</p><p>The evidence is consistent with that: when the CLF is added, the correlations between the constructs <strong>drop</strong> (for example VALUE with TRUST from 0.13 to 0.03, CONFIRM with SATISF from 0.63 to 0.58). The CLF has absorbed part of the true correlation between the constructs.</p><h2>6. Does CMB change the conclusions?</h2><p>The most important question is not "is there CMB?" but "if there is, do the conclusions change?" (Podsakoff et al., 2012). Check this by adding the CLF to the SEM of <a href="https://www.phuonglib.com/p/structural-equation-model-sem-jasp">Post 26, "Structural Equation Model (SEM) in JASP"</a> and comparing the seven path coefficients.</p><p>Result (from <code>cmb.R</code>): in the SEM, the best-fitting CLF loading is <strong>0</strong> (the SEM curve in Figure 10), so the seven path coefficients, their <em>p</em>-values and the R&#178; values stay <strong>the same</strong> to the third decimal (FIML: H1 = 0.577; H4 = 0.091 with <em>p</em> = 0.375; H5 = 0.233 with <em>p</em> = 0.015; H7 = 0.337). The same holds for the complete-case data. The reason: the structural model already explains the correlations between the constructs through its paths, so there is no shared correlation left for the CLF to absorb.</p><p>Conclusion: <strong>there is no evidence that method variance changes the results of the hypothesis tests</strong>.</p><h2>7. Other approaches</h2><div id="datawrapper-iframe" class="datawrapper-wrap outer" data-attrs="{&quot;url&quot;:&quot;https://datawrapper.dwcdn.net/G8BPh/1/&quot;,&quot;thumbnail_url&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/e100f1fe-416e-4cec-8877-495d30de0b9c_1220x718.png&quot;,&quot;thumbnail_url_full&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3df84221-584e-4e0a-b65e-dd89076895ef_1220x788.png&quot;,&quot;height&quot;:342,&quot;title&quot;:&quot;Table 5. Other approaches to common method variance&quot;,&quot;description&quot;:&quot;&quot;,&quot;belowTheFold&quot;:true}" data-component-name="DatawrapperToDOM"><iframe id="iframe-datawrapper" class="datawrapper-iframe" src="https://datawrapper.dwcdn.net/G8BPh/1/" width="730" height="342" frameborder="0" scrolling="no" loading="lazy"></iframe><script type="text/javascript">!function(){"use strict";window.addEventListener("message",(function(e){if(void 0!==e.data["datawrapper-height"]){var t=document.querySelectorAll("iframe");for(var a in e.data["datawrapper-height"])for(var r=0;r<t.length;r++){if(t[r].contentWindow===e.source)t[r].style.height=e.data["datawrapper-height"][a]+"px"}}}))}();</script></div><p>The Marketplace X data have no marker variable, so this post does not demonstrate that approach. If you are designing a questionnaire, add a marker item: it costs very little and gives a much stronger test than Harman's test or the CLF.</p><h2>8. Reporting the results</h2><blockquote><p>Common method variance was examined in three ways. First, Harman's single-factor test (principal axis factoring, unrotated, 254 respondents with complete data) showed that a single factor explained only 30.6% of the variance of the 20 observed variables (33.9% with principal component analysis), below the 50% threshold. Second, a one-factor CFA model fitted clearly worse than the five-factor model (FIML, N = 267: chi-square = 935.06, df = 170, CFI = 0.638 versus chi-square = 134.29, df = 160, CFI = 1.000; chi-square difference = 800.76, df = 10, p &lt; .001). Third, adding a common latent factor with equal loadings did not improve the fit (chi-square difference = 0.10, df = 1, p = .749); the mean method variance was 6.7%, and the construct loadings changed by at most 0.063. Adding the method factor to the structural model did not change any path coefficient. These checks do not suggest that common method variance threatens the results, although they cannot rule it out completely (Podsakoff et al., 2012).</p></blockquote><h2>Five common mistakes</h2><ol><li><p><strong>Writing "the data have no CMB"</strong> on the basis of Harman's test alone. The test only rules out the extreme case.</p></li><li><p><strong>Rotating the factors</strong> in Harman's test (JASP's default is promax). The test requires no rotation.</p></li><li><p><strong>Letting the CLF correlate</strong> with the construct factors, or leaving the 20 loadings free and reporting a model that did not converge.</p></li><li><p><strong>Relying on the p-value of the CLF loading</strong> instead of the chi-square difference test.</p></li><li><p><strong>Skipping prevention</strong> when designing the questionnaire and hoping that a later test will "rescue" the study.</p></li></ol><h2>Frequently asked questions</h2><p><strong>Should Harman's test use PCA or principal axis factoring?</strong></p><p>Podsakoff et al. (2003) describe the test as an unrotated factor analysis without prescribing an extraction method. Theses often use PCA because it is the SPSS default. State which method you used; with the Marketplace X data both give the same conclusion.</p><p><strong>What if Harman's test exceeds 50%?</strong></p><p>That is a warning sign, but it can also come from very strongly correlated constructs (check discriminant validity, <a href="https://www.phuonglib.com/p/convergent-discriminant-validity-cfa-jasp">Post 25, "Convergent and Discriminant Validity in JASP"</a>). Add a CLF or a marker variable and check whether the path coefficients change.</p><p><strong>Do I have to use the model with the CLF for the final results?</strong></p><p>Usually not, when the CLF does not improve the model and does not change the conclusions. If the CLF changes the coefficients considerably, report both models and discuss them.</p><p><strong>How is CMB checked in SmartPLS?</strong></p><p>Usually with the full collinearity VIF (Kock, 2015). The SmartPLS posts cover it in detail.</p><div><hr></div><h2>References</h2><p>Free Software Foundation. (2026). <em>GNU PSPP</em> (Version 2.1.2) [Computer software]. https://www.gnu.org/software/pspp/</p><p>Fuller, C. M., Simmering, M. J., Atinc, G., Atinc, Y., &amp; Babin, B. J. (2016). Common methods variance detection in business research. <em>Journal of Business Research, 69</em>(8), 3192&#8211;3198. https://doi.org/10.1016/j.jbusres.2015.12.008</p><p>Harman, H. H. (1976). <em>Modern factor analysis</em> (3rd ed.). University of Chicago Press.</p><p>JASP Team. (2026). <em>JASP</em> (Version 0.99.0) [Computer software]. https://jasp-stats.org/</p><p>Kock, N. (2015). Common method bias in PLS-SEM: A full collinearity assessment approach. <em>International Journal of e-Collaboration, 11</em>(4), 1&#8211;10. https://doi.org/10.4018/ijec.2015100101</p><p>Lindell, M. K., &amp; Whitney, D. J. (2001). Accounting for common method variance in cross-sectional research designs. <em>Journal of Applied Psychology, 86</em>(1), 114&#8211;121. https://doi.org/10.1037/0021-9010.86.1.114</p><p>Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., &amp; Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. <em>Journal of Applied Psychology, 88</em>(5), 879&#8211;903. https://doi.org/10.1037/0021-9010.88.5.879</p><p>Podsakoff, P. M., MacKenzie, S. B., &amp; Podsakoff, N. P. (2012). Sources of method bias in social science research and recommendations on how to control it. <em>Annual Review of Psychology, 63</em>, 539&#8211;569. https://doi.org/10.1146/annurev-psych-120710-100452</p><p>Richardson, H. A., Simmering, M. J., &amp; Sturman, M. C. (2009). A tale of three perspectives: Examining post hoc statistical techniques for detection and correction of common method variance. <em>Organizational Research Methods, 12</em>(4), 762&#8211;800. https://doi.org/10.1177/1094428109332834</p><p>Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. <em>Journal of Statistical Software, 48</em>(2), 1&#8211;36. https://doi.org/10.18637/jss.v048.i02</p><p>Williams, L. J., Cote, J. A., &amp; Buckley, M. R. (1989). Lack of method variance in self-reported affect and perceptions at work: Reality or artifact? <em>Journal of Applied Psychology, 74</em>(3), 462&#8211;468. https://doi.org/10.1037/0021-9010.74.3.462</p>]]></content:encoded></item></channel></rss>