Rater differences in psychopathy measure scoring and predictive validity

Although field studies reveal that some forensic evaluators tend to assign higher psychopathy measure scores to sexual offenders than others, the extent to which these findings apply to psychopathy measure scoring in other contexts is unclear. And no study has examined the impact of evaluator differences in scoring on predictive validity. We used data from the MacArthur Violence Risk Assessment Study to examine whether there were rater differences in psychopathy measure scoring and predictive effects among trained raters in a rigorous research context. The proportion of variance in Psychopathy Checklist: Screening Version (Hart, Cox, & Hare, 1995) scores attributable to raters was larger for Part 1 (14%) than Part 2 (4%) scores. The association between Facet 4 scores and future violence was stronger among evaluators who assigned higher and more variable Facet 4 scores, but there were no similar effects for other PCL:SV scores. Although there was only limited evidence for an association between PCL:SV scoring tendencies and predictive validity, findings show that mean differences in scoring have implications for score interpretation, with the cut score that indicates a high level of risk being lower when it comes from a rater who assigns relatively low scores compared to a rater who assigns relatively high scores. These findings suggest that evaluators should carefully consider their own psychopathy measure scoring tendencies across cases and the extent to which these tendencies are consistent with the normative sample scores that form the basis of their psychopathy measure score interpretations.