Item Response Theory Analysis of the Triarchic Psychopathy Measure

This study investigated item and scale functioning in the triarchic psychopathy measure (TriPM) using an item response theory (IRT) analysis. TriPM data collected from 937 (410 men and 527 women) college students and community members were analyzed. A graded response model was utilized to analyze the items comprising the TriPM's three scales. Exploratory factor analyses (EFAs) supported the unidimensionality of the Disinhibition and Meanness scales, whereas Boldness was best represented by two domains. Each TriPM scale was found to include some items that had limited ability to differentiate between respondents possessing varying levels of the trait being measured. Across all scales, 34 items (61%) yielded evidence of significantly different responding between men and women possessing similar levels of the underlying trait, or differential item functioning (DIF). The scoring format (i.e., directly scored vs. reverse scored) also influenced item functioning, especially when the given scale included a large majority of one scoring type. Overall, the IRT and DIF results suggest that the TriPM effectively identifies disinhibition and meanness in individuals possessing high levels of those traits, and boldness at lower levels, but does not seem to identify these traits equally well in men and women. Possible revisions to the Boldness scale as well as ideas for addressing female and male DIF were provided.