Confirmatory factor analysis and the validity of the measurement scales within a causal modelling framework

Abstract

This research discusses the role of structural equation models and, specifically, confirmatory factor analysis, regarding the validity of the measurement scales when a casual model is proposed. The viewpoints of Hayduk (1996) and Borsboom, Mellenbergh and van Heerden (2004) are discussed, in order to provide a framework upon which researchers can decide how to proceed in applied research. The divergences of these per-spectives with regard to the widespread use of the two-step procedure (Anderson and Gerbing, 1988) should encourage thinking about the suit-ability of the confirmatory factor analysis, and the adequacy of the empiri-cal procedures for studying criterion, convergent and discriminant validity.