Psychopathy Checklist: Screening Version: A Bifactor Structure for Forensic and Community Samples

We evaluated competing measurement models for the Psychopathy Checklist: Screening Version (PCL:SV), one of the most widely used assessment instruments of psychopathy, which applies to clinical, forensic, and community samples. With four all-male community and forensic samples collected in Germany and the United States (nTotal = 853), we evaluated several measurement models of the PCL:SV, including correlated factor models, higher-order models, and bifactor models, on numerous psychometric criteria including model fit, factor saturation, and predictive utility. Across the samples, we found that a bifactor model had better fit to the data and explained comparable or more variance in measures of fluid intelligence, crystallized intelligence, and characteristics of antisocial personality disorder relative to the four correlated factors model. The best bifactor structure model differed between forensic and community samples. With both populations, in addition to a g factor which loaded on all items, a nested factor representing the lifestyle and antisocial facets was also modeled. With the forensic sample, an additional orthogonal nested factor representing the interpersonal facet was needed. We conclude that there is strong communality across the 12 PCL:SV items, which is best captured by a bifactor structure. However, the specific nesting structure differs due to whether or not the sample includes a forensic or community population. Public Significance Statement We evaluated the appropriateness of current scoring practices of Psychopathy Checklist: Screening Version, a popular measure for assessing psychopathy. We found that the current scoring approach may not be optimal and that scores may not be comparable between forensic and non-forensic samples.