THE EFFECTS OF UNDEREXTRACTION IN FACTOR AND COMPONENT ANALYSES

The consequences of underextracting factors and components within and between the methods of maximum likelihood factor analysis (MLFA) and p rincipal components analysis (PCA) were examined. Computer-simulated d ata sets represented a range of pattern structures. Manipulated condit ions included component (factor) structure coefficients (a(ij) = .8, . 6, and .4), sample size (N = 75, 150, 225, and 450), and variable-to-c omponent (factor) ratio (p:m = 6:1 and 4:1). The principal components score and the Anderson and Rubin factor score estimate were calculated for both the correct patterns and the incorrect (underextracted) patt erns. In Study 1, underextraction led to substantial degradation of sc ores within both methods, but the component score degraded less rapidl y. Score degradation was related to the number of original components (factors). In Study 2, between-method comparisons indicated very high similarity for baseline score patterns, but dissimilarity occurred wit h underextraction.