Exploratory Factor Analysis

This chapter demonstrates the method of exploratory common factor analysis in SPSS. Exploratory factor analysis is quite different from components analysis. In the exploratory factor analysis, the user can exercise more modeling flexibility in terms of which parameters to fix and which to free for estimation. The major critique of exploratory factor analysis is that the loadings obtained in the procedure are not unique. The chapter demonstrates how factor loadings are in part a function of the number of factors extracted. Factor analysis generates potential factors due to correlation among observed variables. If there is no correlation among variables, then there is essentially nothing to factor analyze. The chapter concludes with a brief discussion of cluster analysis, which shares a conceptual link to factor analysis in that instead of attempting to group variables, cluster analysis attempts to group cases through a consideration of relative distances between objects.