A Monté Carlo Comparison of Factor Analytic Methods

Abstract

Three factor analytic procedures, principal components analysis, image analysis, and alpha factor analysis, are compared by means of a Monté Carlo simulation. The simulation allows systematic tests of the ability of each method to recover prespecified factor patterns. Image and alpha analysis are found to be more accurate than principal components analysis, especially for the type of problem to which factor analysis is applied in marketing. Recommendations for applied research and suggestions for future methodological work are made on the basis of the study findings.