Confirmatory factor analysis and the factor structure of expagg in context: A reply to forrest et al., 2002

摘要

It has been suggested that confirmatory factor analysis (CFA) can be used to investigate the construct validity of psychometric scales and Forrest et al. [2000] specifically query the factor structure of Expagg using this technique. In this paper we report unsuccessful attempts to confirm the factor structure of three widely used scales using CFA criteria. In the fourth study, a two-factor model of Expagg, which has been derived from previous studies, is tested for fit on new data. The results suggest that from a CFA point of view, Expagg is best considered as two scales measuring expressivity and instrumentality with five items on each scale. This model satisfies four of the five fit criteria (CFI = 90, GFI = .94, RMSEA = .08, ECVI = .44), failing only on the chi square test, a benchmark that has attracted criticism from statisticians. Other concerns are raised about the meaning of CFA results and their importance.