Theoretical versus mathematical approach to modeling psychological and physiological data

Abstract

Variable selection for predictive modeling has traditionally relied on theory in the psychological domain. Given the recent advancements in computing technology and availability, researchers are able to utilize more sophisticated mathematical modeling techniques with greater ease. The challenge becomes evaluating whether theory or mathematics should be relied upon for model development. The presented analyses compared the use of hierarchical and stepwise variable selection methods during a predictive modeling task using linear regression. The results show that the stepwise variable selection method is able to obtain a more efficient model than the hierarchical variable selection method. Implications and recommendations for researchers are further discussed.