Comparing Out-of-Sample Predictive Ability of PLS, Covariance, and Regression Models

Partial Least Squares Path Modelling (PLSPM) is a popular technique for estimating structural equation models in the social sciences, and is frequently presented as an alternative to covariance-based analysis as being especially suited for predictive modeling. While existing research on PLSPM has focused on its use in causalexplanatory modeling, this paper follows two recent papers at ICIS 2012 and 2013 in examining how PLSPM performs when used for predictive purposes. Additionally, as a predictive technique, we compare PLSPM to traditional regression methods that are widely used for predictive modelling in other disciplines. Specifically, we employ out-ofsample k-fold cross-validation to compare PLSPM to covariance-SEM and a range of atheoretical regression techniques in a simulation study. Our results show that PLSPM offers advantages over covariance-SEM and other prediction methods.

[1]  T. Dijkstra.  Latent Variables and Indices: Herman Wold’s Basic Design and Partial Least Squares , 2010 .

[2]  Marko Sarstedt,et al.  An assessment of the use of partial least squares structural equation modeling in marketing research , 2012 .

[3]  Humam Elagha,et al.  The Use of Partial Least Squares Path Modeling in IT Governance Discipline , 2014, 2014 11th International Conference on Information Technology: New Generations.

[4]  R. P. McDonald,et al.  Path Analysis with Composite Variables. , 1996, Multivariate behavioral research.

[5]  Philippe Jacquart,et al.  On making causal claims: A review and recommendations , 2010 .

[6]  Geoffrey S. Hubona,et al.  A Scientific Basis for Rigor in Information Systems Research , 2009, MIS Q..

[7]  T. Næs,et al.  Canonical partial least squares—a unified PLS approach to classification and regression problems , 2009 .

[8]  Thomas H. Davenport.  Keep Up with Your Quants , 2013 .

[9]  S. D. Jong.  SIMPLS: an alternative approach to partial least squares regression , 1993 .

[10]  Herman Wold,et al.  Systems under indirect observation : causality, structure, prediction , 1982 .

[11]  Bart Baesens,et al.  An empirical evaluation of the comprehensibility of decision table, tree and rule based predictive models , 2011, Decis. Support Syst..

[12]  Manendra Mohan,et al.  International Marketing , 1978, International Business Strategy.

[13]  Arun Rai,et al.  Predictive Validity and Formative Measurement in Structural Equation Modeling: Embracing Practical Relevance , 2013, ICIS.

[14]  Marko Sarstedt,et al.  PLS-SEM: Indeed a Silver Bullet , 2011 .

[15]  Mary Tate,et al.  Comparing the Predictive Ability of PLS and Covariance Models , 2012, ICIS.

[16]  William Lewis,et al.  Research Note - Statistical Power in Analyzing Interaction Effects: Questioning the Advantage of PLS with Product Indicators , 2007, Inf. Syst. Res..

[17]  Mikko Rönkkö,et al.  A Critical Examination of Common Beliefs About Partial Least Squares Path Modeling , 2013 .

[18]  Rudolf R. Sinkovics,et al.  The Use of Partial Least Squares Path Modeling in International Marketing , 2009 .

[19]  Galit Shmueli,et al.  Predictive Analytics in Information Systems Research , 2010, MIS Q..

[20]  Tormod Næs,et al.  Multivariate calibration. I. Concepts and distinctions , 1984 .

[21]  Herman Wold,et al.  Soft modelling: The Basic Design and Some Extensions , 1982 .

[22]  William Lewis,et al.  Comparing PLS to Regression and LISREL: A Response to Marcoulides, Chin, and Saunders , 2012, MIS Q..

[23]  H. Wold.  Models for Knowledge , 1982 .

[24]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[25]  Mary Tate,et al.  Testing Models or Fitting Models? Identifying Model Misspecification in PLS , 2010, ICIS.

[26]  D. Straub,et al.  Editor's comments: a critical look at the use of PLS-SEM in MIS quarterly , 2012 .

[27]  T. Davenport.  Competing on analytics. , 2006, Harvard business review.

[28]  S. Wold,et al.  The kernel algorithm for PLS , 1993 .

[29]  Kalle Lyytinen,et al.  Data matters in IS theory building , 2009, J. Assoc. Inf. Syst..

[30]  Jan-Bernd Lohmöller,et al.  Latent Variable Path Modeling with Partial Least Squares , 1989 .

[31]  Shirley Gregor,et al.  The Nature of Theory in Information Systems , 2006, MIS Q..

[32]  T. Dijkstra.  Some comments on maximum likelihood and partial least squares methods , 1983 .

[33]  Alex Alves Freitas,et al.  Comprehensible classification models: a position paper , 2014, SKDD.