An examination of Psychopathy Checklist-Revised latent factor structure via exploratory structural equation modeling.

The Psychopathy Checklist-Revised (PCL-R; Hare, 1991, 2003) is widely used in clinical and forensic practice. Its latent factor structure has been subject to debate and controversy for almost 40 years. Various principal component and factor models have been proposed. It is argued that the use of independent cluster models has been questionable. The use of models that assume that symptoms will have zero loadings on all but one factor fails to reflect the clinical reality that symptoms of personality disorder are likely to be the consequence of multiple latent traits. This misspecification of models will lead to excessive latent factor correlations, which, in turn, have a negative impact on the nomological network of relations among latent traits and other constructs. In this study, we applied exploratory structural equation modeling (ESEM) to PCL-R data derived from a multisite U.K. sample. First, we tested models derived from the 3- and 4-factor perspectives using confirmatory factor analysis. Second, we tested ESEM models based on these two perspectives. The 3-factor model displayed stability across both analytic methods, with the latent factor correlations being smaller in the ESEM analysis. The 4-factor model displayed less stability across analytic methods, with a major crime factor accounting for a large proportion of variance in numerous items. Comprehensive Assessment of Psychopathic Personality domain ratings were used as external criteria. The concurrent validity of the 3-factor ESEM model was conceptually clearer than that for the equivalent 4-factor model. The application of ESEM may provide greater clarity to the debate regarding the latent structure of the PCL-R. (PsycINFO Database Record (c) 2018 APA, all rights reserved).