[{"data":1,"prerenderedAt":427},["ShallowReactive",2],{"content-query-IsLT8OQ52S":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"layout":10,"body":11,"_type":421,"_id":422,"_source":423,"_file":424,"_extension":425,"sitemap":426},"/documentation/algorithms-and-techniques/validity-and-model-fit/model-comparison","validity-and-model-fit",false,"","Model Comparison in SmartPLS: Comparing PLS-SEM Models","SmartPLS's model comparison feature compares two PLS-SEM models using PLSpredict, CVPAT, BIC, and Akaike weights. Learn how to run a model comparison in SmartPLS.","algorithm-description",{"type":12,"children":13,"toc":408},"root",[14,23,29,36,41,82,87,93,100,105,111,116,122,127,133,138,144,182,188],{"type":15,"tag":16,"props":17,"children":19},"element","h1",{"id":18},"model-comparison",[20],{"type":21,"value":22},"text","Model Comparison",{"type":15,"tag":24,"props":25,"children":26},"p",{},[27],{"type":21,"value":28},"Researchers often have alternative ways of theorizing their models. The PLS-SEM model comparison feature enables the comparison of two distinct models by assessing them against model selection criteria and statistical tests, giving researchers a foundation for informed decision-making when selecting the most suitable model.",{"type":15,"tag":30,"props":31,"children":33},"h2",{"id":32},"how-to-run-a-model-comparison-in-smartpls",[34],{"type":21,"value":35},"How to Run a Model Comparison in SmartPLS",{"type":15,"tag":24,"props":37,"children":38},{},[39],{"type":21,"value":40},"Create two alternative models within a SmartPLS project (e.g., Model 1 and Model 2). Select the first model and open it in the modeling view. Under \"Calculate\" in the menu bar, you will find the option \"Model comparison\". In the \"Model comparison\" start dialog, you can select the second model with which the currently open model is to be compared. After starting the algorithm, SmartPLS generates the following model comparison results:",{"type":15,"tag":42,"props":43,"children":44},"ul",{},[45,58,69],{"type":15,"tag":46,"props":47,"children":48},"li",{},[49,56],{"type":15,"tag":50,"props":51,"children":53},"a",{"href":52},"/documentation/algorithms-and-techniques/prediction-and-segmentation/predict",[54],{"type":21,"value":55},"PLSpredict",{"type":21,"value":57}," (Shmueli et al., 2016; Shmueli et al., 2019),",{"type":15,"tag":46,"props":59,"children":60},{},[61,67],{"type":15,"tag":50,"props":62,"children":64},{"href":63},"/documentation/algorithms-and-techniques/resampling-and-inference/cvpat",[65],{"type":21,"value":66},"cross-validated predictive ability test",{"type":21,"value":68}," (CVPAT; Liengaard et al., 2022; Sharma et al., 2023), and",{"type":15,"tag":46,"props":70,"children":71},{},[72,74,80],{"type":21,"value":73},"Bayesian information criterion (BIC) for ",{"type":15,"tag":50,"props":75,"children":77},{"href":76},"/documentation/algorithms-and-techniques/validity-and-model-fit/prediction-oriented-model-selection",[78],{"type":21,"value":79},"predictive model selection",{"type":21,"value":81}," (Sharma et al., 2019; Sharma et al., 2021) and Akaike weights (Danks et al., 2020; Rigdon et al., 2023).",{"type":15,"tag":24,"props":83,"children":84},{},[85],{"type":21,"value":86},"Based on these results, researchers and practitioners can decide which of the two model alternatives is advantageous, for example with regard to predictive power.",{"type":15,"tag":30,"props":88,"children":90},{"id":89},"frequently-asked-questions",[91],{"type":21,"value":92},"Frequently Asked Questions",{"type":15,"tag":94,"props":95,"children":97},"h3",{"id":96},"what-does-the-smartpls-model-comparison-feature-do",[98],{"type":21,"value":99},"What does the SmartPLS model comparison feature do?",{"type":15,"tag":24,"props":101,"children":102},{},[103],{"type":21,"value":104},"It compares two alternative PLS-SEM models against model selection criteria and statistical tests, including PLSpredict, the cross-validated predictive ability test (CVPAT), the Bayesian information criterion (BIC) for prediction-oriented model selection, and Akaike weights.",{"type":15,"tag":94,"props":106,"children":108},{"id":107},"how-do-i-run-a-model-comparison-in-smartpls",[109],{"type":21,"value":110},"How do I run a model comparison in SmartPLS?",{"type":15,"tag":24,"props":112,"children":113},{},[114],{"type":21,"value":115},"Create two alternative models in the same SmartPLS project, open the first model in the modeling view, and choose \"Model comparison\" under \"Calculate\" in the menu bar. Then select the second model to compare it against.",{"type":15,"tag":94,"props":117,"children":119},{"id":118},"which-criterion-should-i-use-to-decide-between-two-competing-models",[120],{"type":21,"value":121},"Which criterion should I use to decide between two competing models?",{"type":15,"tag":24,"props":123,"children":124},{},[125],{"type":21,"value":126},"It depends on your research objective. PLSpredict and CVPAT focus on out-of-sample predictive performance, while the BIC and Akaike weights are in-sample criteria for prediction-oriented model selection. Considering several criteria together provides a more complete picture than relying on a single one.",{"type":15,"tag":94,"props":128,"children":130},{"id":129},"why-would-i-compare-two-pls-sem-models-instead-of-just-estimating-one",[131],{"type":21,"value":132},"Why would I compare two PLS-SEM models instead of just estimating one?",{"type":15,"tag":24,"props":134,"children":135},{},[136],{"type":21,"value":137},"Researchers often have alternative, theoretically plausible ways of specifying their model. Model comparison offers a systematic, statistically grounded way of deciding which of two competing model specifications is preferable, rather than relying on judgment alone.",{"type":15,"tag":30,"props":139,"children":141},{"id":140},"related-smartpls-methods",[142],{"type":21,"value":143},"Related SmartPLS Methods",{"type":15,"tag":42,"props":145,"children":146},{},[147,156,164,173],{"type":15,"tag":46,"props":148,"children":149},{},[150],{"type":15,"tag":50,"props":151,"children":153},{"href":152},"/documentation/algorithms-and-techniques/validity-and-model-fit/prediction-oriented-model-selection/",[154],{"type":21,"value":155},"Prediction-oriented model selection (BIC)",{"type":15,"tag":46,"props":157,"children":158},{},[159],{"type":15,"tag":50,"props":160,"children":162},{"href":161},"/documentation/algorithms-and-techniques/prediction-and-segmentation/predict/",[163],{"type":21,"value":55},{"type":15,"tag":46,"props":165,"children":166},{},[167],{"type":15,"tag":50,"props":168,"children":170},{"href":169},"/documentation/algorithms-and-techniques/resampling-and-inference/cvpat/",[171],{"type":21,"value":172},"Cross-validated predictive ability test (CVPAT)",{"type":15,"tag":46,"props":174,"children":175},{},[176],{"type":15,"tag":50,"props":177,"children":179},{"href":178},"/documentation/algorithms-and-techniques/validity-and-model-fit/model-fit/",[180],{"type":21,"value":181},"Model fit",{"type":15,"tag":30,"props":183,"children":185},{"id":184},"references",[186],{"type":21,"value":187},"References",{"type":15,"tag":42,"props":189,"children":190},{},[191,221,247,273,299,325,349,374,399],{"type":15,"tag":46,"props":192,"children":193},{},[194,196,204,206,212,214,219],{"type":21,"value":195},"Danks, N. P., Sharma, P. N., & Sarstedt, M. (2020). ",{"type":15,"tag":50,"props":197,"children":201},{"href":198,"rel":199},"https://doi.org/10.1016/j.jbusres.2020.03.019",[200],"nofollow",[202],{"type":21,"value":203},"Model selection uncertainty and multimodel inference in partial least squares structural equation modeling (PLS-SEM).",{"type":21,"value":205}," ",{"type":15,"tag":207,"props":208,"children":209},"em",{},[210],{"type":21,"value":211},"Journal of Business Research",{"type":21,"value":213},", ",{"type":15,"tag":207,"props":215,"children":216},{},[217],{"type":21,"value":218},"113",{"type":21,"value":220},", 13–24.",{"type":15,"tag":46,"props":222,"children":223},{},[224,226,233,234,239,240,245],{"type":21,"value":225},"Liengaard, B. D., Sharma, P. N., Hult, G. T. M., Jensen, M. B., Sarstedt, M., Hair, J. F., & Ringle, C. M. (2021). ",{"type":15,"tag":50,"props":227,"children":230},{"href":228,"rel":229},"https://onlinelibrary.wiley.com/doi/full/10.1111/deci.12445",[200],[231],{"type":21,"value":232},"Prediction: Coveted, yet forsaken? Introducing a cross-validated predictive ability test in partial least squares path modeling.",{"type":21,"value":205},{"type":15,"tag":207,"props":235,"children":236},{},[237],{"type":21,"value":238},"Decision Sciences",{"type":21,"value":213},{"type":15,"tag":207,"props":241,"children":242},{},[243],{"type":21,"value":244},"52",{"type":21,"value":246},"(2), 362–392.",{"type":15,"tag":46,"props":248,"children":249},{},[250,252,259,260,265,266,271],{"type":21,"value":251},"Rigdon, E., Sarstedt, M., & Moisescu, O.-I. (2023). ",{"type":15,"tag":50,"props":253,"children":256},{"href":254,"rel":255},"https://doi.org/10.1111/ijcs.12906",[200],[257],{"type":21,"value":258},"Quantifying model selection uncertainty via bootstrapping and Akaike weights.",{"type":21,"value":205},{"type":15,"tag":207,"props":261,"children":262},{},[263],{"type":21,"value":264},"International Journal of Consumer Studies",{"type":21,"value":213},{"type":15,"tag":207,"props":267,"children":268},{},[269],{"type":21,"value":270},"47",{"type":21,"value":272},"(4), 1596–1608.",{"type":15,"tag":46,"props":274,"children":275},{},[276,278,285,286,291,292,297],{"type":21,"value":277},"Sharma, P. N., Liengaard, B. D., Hair, J. F., Sarstedt, M., & Ringle, C. M. (2023). ",{"type":15,"tag":50,"props":279,"children":282},{"href":280,"rel":281},"https://www.emerald.com/insight/content/doi/10.1108/EJM-08-2020-0636/",[200],[283],{"type":21,"value":284},"Predictive model assessment and selection in composite-based modeling using PLS-SEM: Extensions and guidelines for using CVPAT.",{"type":21,"value":205},{"type":15,"tag":207,"props":287,"children":288},{},[289],{"type":21,"value":290},"European Journal of Marketing",{"type":21,"value":213},{"type":15,"tag":207,"props":293,"children":294},{},[295],{"type":21,"value":296},"57",{"type":21,"value":298},"(6), 1662–1677.",{"type":15,"tag":46,"props":300,"children":301},{},[302,304,311,312,317,318,323],{"type":21,"value":303},"Sharma, P. N., Sarstedt, M., Shmueli, G., Kim, K. H., & Thiele, K. O. (2019). ",{"type":15,"tag":50,"props":305,"children":308},{"href":306,"rel":307},"https://aisel.aisnet.org/jais/vol20/iss4/4/",[200],[309],{"type":21,"value":310},"PLS-based model selection: The role of alternative explanations in information systems research.",{"type":21,"value":205},{"type":15,"tag":207,"props":313,"children":314},{},[315],{"type":21,"value":316},"Journal of the Association for Information Systems",{"type":21,"value":213},{"type":15,"tag":207,"props":319,"children":320},{},[321],{"type":21,"value":322},"20",{"type":21,"value":324},"(4), 346–397.",{"type":15,"tag":46,"props":326,"children":327},{},[328,330,337,338,342,343,347],{"type":21,"value":329},"Sharma, P. N., Shmueli, G., Sarstedt, M., Danks, N., & Ray, S. (2021). ",{"type":15,"tag":50,"props":331,"children":334},{"href":332,"rel":333},"https://onlinelibrary.wiley.com/doi/abs/10.1111/deci.12329",[200],[335],{"type":21,"value":336},"Prediction-oriented model selection in partial least squares path modeling.",{"type":21,"value":205},{"type":15,"tag":207,"props":339,"children":340},{},[341],{"type":21,"value":238},{"type":21,"value":213},{"type":15,"tag":207,"props":344,"children":345},{},[346],{"type":21,"value":244},{"type":21,"value":348},"(3), 567–607.",{"type":15,"tag":46,"props":350,"children":351},{},[352,354,361,362,366,367,372],{"type":21,"value":353},"Shmueli, G., Ray, S., Estrada, J. M. V., & Chatla, S. B. (2016). ",{"type":15,"tag":50,"props":355,"children":358},{"href":356,"rel":357},"https://www.sciencedirect.com/science/article/abs/pii/S0148296316301217",[200],[359],{"type":21,"value":360},"The elephant in the room: Predictive performance of PLS models.",{"type":21,"value":205},{"type":15,"tag":207,"props":363,"children":364},{},[365],{"type":21,"value":211},{"type":21,"value":213},{"type":15,"tag":207,"props":368,"children":369},{},[370],{"type":21,"value":371},"69",{"type":21,"value":373},"(10), 4552–4564.",{"type":15,"tag":46,"props":375,"children":376},{},[377,379,386,387,391,392,397],{"type":21,"value":378},"Shmueli, G., Sarstedt, M., Hair, J. F., Cheah, J. H., Ting, H., Vaithilingam, S., & Ringle, C. M. (2019). ",{"type":15,"tag":50,"props":380,"children":383},{"href":381,"rel":382},"https://www.emerald.com/insight/content/doi/10.1108/EJM-02-2019-0189/",[200],[384],{"type":21,"value":385},"Predictive model assessment in PLS-SEM: Guidelines for using PLSpredict.",{"type":21,"value":205},{"type":15,"tag":207,"props":388,"children":389},{},[390],{"type":21,"value":290},{"type":21,"value":213},{"type":15,"tag":207,"props":393,"children":394},{},[395],{"type":21,"value":396},"53",{"type":21,"value":398},"(11), 2322–2347.",{"type":15,"tag":46,"props":400,"children":401},{},[402],{"type":15,"tag":50,"props":403,"children":405},{"href":404},"/documentation",[406],{"type":21,"value":407},"More literature ...",{"title":7,"searchDepth":409,"depth":409,"links":410},2,[411,412,419,420],{"id":32,"depth":409,"text":35},{"id":89,"depth":409,"text":92,"children":413},[414,416,417,418],{"id":96,"depth":415,"text":99},3,{"id":107,"depth":415,"text":110},{"id":118,"depth":415,"text":121},{"id":129,"depth":415,"text":132},{"id":140,"depth":409,"text":143},{"id":184,"depth":409,"text":187},"markdown","content:documentation:algorithms-and-techniques:validity-and-model-fit:model-comparison:index.md","content","documentation/algorithms-and-techniques/validity-and-model-fit/model-comparison/index.md","md",{"loc":4},1784805337653]