Problems in the application of structural equation modeling: Comment on Randhawa, Beamer, and Lundberg (1993).

Abstract

The purpose of this article is to critique methodologicaI aspects of the B. S. Randhawa, J. E. Beamer, and I. Lundberg's (1993) structural equation modeling (SEM) of relations between self-efficacy, attitudes, and achievement in mathematics. Major findings of their study were that self-efficacy mediated the effects of attitudes on achievement and that self-efficacy and attitudes were more strongly related to achievement for boys than for girls. Our reanalysis of their data does not substantiate these conclusions. Rather, it identifies inappropriate applications of SEM and the consequent misinterpretations of results