Latent variable modeling and cross-national differences in college-level academic performance

This chapter offers a critique of the latent variable modeling approach taken by Ransdell et al. Suggestions for improving the analytical strategy are described, including the use of confirmatory factor analysis in conjunction with structural equation modeling. The chapter concludes with some examples of confirmatory factor analysis and structural equation modeling from other data sets in order to demonstrate each techniqueʹs advantages and disadvantages. Finally, several directions for future research, including the use of longitudinal methods, are suggested in order to help constrain the direction of causative links between academic performance and its predictors.