Implicit measurement of violence-related cognitions

OBJECTIVE: Violence-related cognitions may underpin many acts of violence, but explicit self-report measures of these cognitions may be inadequate to assess them fully due to their unconscious nature or due to deliberate dissimulation. We designed three versions of the implicit association test (IAT) that separately examined violence-related associations to valence (good vs bad), hedonic-value (enjoy vs dislike) and arousal (exciting vs boring) and examined if these were associated with greater levels of past criminal activity and self-reported violence. METHOD: The three IATs were administered to an offender sample (N = 108) with a history of serious offending and a diagnosis of Personality Disorder. Violence was assessed via the Proactive and Reactive Aggression Questionnaire and by examination of previous convictions. RESULTS: The IATs showed good to moderate reliability. The valence-IAT did not show any significant correlations to the measures of previous violence. The hedonic value-IAT showed positive relationships with official records of convictions, especially among participants without a conviction for homicide. The arousal-IAT was positively related to self-reported aggression in those without a conviction for homicide. CONCLUSIONS: The results show some promise that indirect techniques may be able to measure violence-related cognitions