Confirmatory Factor Analysis

This article is the second of two on the use of confirmatory factor analysis (CFA) as a method to assess construct validity. The construct validation criteria required by the conventional MTMM approach are satisfied only by certain ideal data sets, such as those in which the method variance of measures is very low. The CFA approach to multitraitmultimethod (MTMM) data is more general, in that violations of those stringent criteria can be managed. Another limitation of the conventional MTMM approach is that only a relatively small number of indicators can be examined by bivariate analysis. The economy of the CFA approach permits the analysis of a much larger number of indicators. In this article, a data set is analyzed using the CFA approach. Results are presented that illustrate the application of this-statisticai method.