# Self-report psychopathy-III facet scores predict sexual crimes, sexual preferences, and sexual deviance index validity more precisely than total scores

## TL;DR

Using detailed scores from a psychopathy assessment tool predicts sexual crimes and deviant preferences better than overall scores.

## Contribution

The study shows that analyzing individual facets of the SRP-III improves prediction accuracy over total scores.

## Key findings

- Facet scores of SRP-III predict sexual crimes and deviant preferences more precisely than total scores.
- Factor scores perform better than total scores but worse than facet scores in prediction accuracy.
- Only facet scores show significant correlations with dependent variables.

## Abstract

Understanding the profiles of sexual offenders, such as the presence of psychopathic traits, is key to preventing future sexual crimes. The self-report psychopathy-III (SRP-III) is a tool used to assess the characteristics of psychopathy, but improvements on its interpretation are required to maximize its precision. The SRP-III can be interpreted by examining the scores on each of the four facets (interpersonal manipulation, callous affect, erratic lifestyle, antisocial behavior), on each of two factors (factor 1, factor 2), or by examining the total score. Here, we investigate the interpretation of the results from the SRP-III using these three approaches of analysis of the data for predicting types of sexual crimes, sexually deviant preferences (measured via PPG), and the validity of the sexual deviance results. Logistic regressions were carried out using either the four facets, two factors, or the total score of the SRP-III. Data were previously obtained from 198 Canadian men who were convicted of, or who admitted to committing, at least one sexual crime, or who reported experiencing sexually deviant fantasies. We also examined the point-biserial correlations between each of the methods of interpreting the SRP-III results and each of the dependent variables. We find that SRP-III facet scores most precisely predict types of sexual crimes, sexually deviant preferences, and sexual deviance index validity, followed by SRP-III factor scores, and lastly SRP-III total scores. Additionally, significant correlations are only found between SRP-III scores and one dependent variable. Potential reasons for this are discussed. Based on these findings, we recommend that future studies consider facet and factor scores in addition to the standard practice of examining total scores.