Factor analysis, Simulation methods, Regression analysis
Horn’s parallel analysis (PA) is an empirical method used to decide how many components in a principal component analysis (PCA) or factors in a common factor analysis (CFA) drive the variance observed in a data set of n observations on p variables (Horn, 1965). This decision of how many components or factors to retain is critical in applications of PCA or CFA to reducing the dimensionality of data in analysis (as when compositing multiple scale items into a single score), and also in exploratory factor analysis where the different contributions of each factor to each observed variable help generate theory (Preacher & MacCallum, 2003; Velicer & Jackson, 1990). As will be shown, the development of PA was predicated upon properties of PCA. However, some have been exponents of the use of PA for CFA (Velicer, Eaton, & Fava, 2000). The correct application of PA with CFA requires modification to the original PA procedure. This paper attempts to clarify PA with respect to both PCA and CFA.
This paper should be cited as an unpublished manuscript using the URL: http://doyenne.com/Software/files/PA_for_PCA_vs_FA.pdf