Predicting multiple criteria of criminal behavior with HEXACO domains and facets

HEXACO model, a novel lexical model of personality, can be particularly useful for determining the personality dispositions toward criminal behavior. HEXACO-PI-R inventory was used to predict multiple criteria of criminal behavior (number of offenses, number of verdicts and criminal-legal recidivism) in the sample of male convicts from two penal institutions in Serbia ( N  = 181; mean age 36.3; SD = 9.7). Data analysis was done using the multiple linear regressions, on the domains level and on the level of facets, separately. On the domains level traits, only one of the obtained regression functions is statistically significant, with criminal-legal recidivism as criteria; Agreeableness and Conscientiousness are most important predictors. On facets level traits, all three regression functions are significant. Fairness, Gentleness, Organization and Creativity, all with negative beta ponders, have the most important role in this regression models. Results suggest that analyses on facets-level, rather than on domains-level, are more useful for understanding criminal behavior.