Confirmatory Factor Extraction Versus Confirmatory Factor Rotation

Abstract

Confirmatory factor analysis is a true statistical technique in the sense that it can test hypothesized relationships among variables. Unlike exploratory factor analysis, confirmatory factor analysis can be used to disconfirm or support a priori expectations about the target population based on sample data. Confirmatory factor analysis can be combined with exploratory techniques in theory building. Such is the case in confirmatory factor extraction with subsequent relaxation of parameter constraints in stepwise fashion to determine the best model. Initial exploratory factor extraction can also be followed by confirmatory factor rotation. Confirmatory factor extraction differs from confirmatory factor rotation with respect to the nature of the hypothesized relationships and the point at which the hypotheses are tested. Results of both confirmatory factor extraction and confirmatory factor rotation must be interpreted with caution. Assumptions of the techniques should not be violated, and magnitude of residuals should be examined as an estimate of the extent to which the model is capitalizing on error variance. A hypothetical data set (scores of 50 subjects on nine variables) is used to empirically demonstrate these methods. (JAZ)