LAUSR

OBJECTIVE Anger rumination and imagined violence, in the context of anger/aggression proclivity, are examined for their direct and conjoint associations with violent behavior by psychiatric patients. METHOD A secondary analysis of data from the MacArthur Violence Risk Study was conducted with 1136 acute civil commitment patients, assessed during hospitalization and after hospital discharge. Anger/aggression proclivity was assessed with the Brief Psychiatric Rating Scale hostility subscale, anger rumination was indexed using items from the Novaco Anger Scale, and imagined violence was measured with Grisso's Schedule of Imagined Violence. Violence, prehospitalization and posthospitalization, was indexed by the MacArthur project measure. RESULTS Correlational analyses, mediation analyses, and moderated mediation analyses were conducted. Anger rumination significantly predicted pre- and posthospitalization violence, when controlling for age, sex, race, childhood physical abuse, and anger/aggression proclivity; and it partially mediated the relation between anger/aggression proclivity and violence. Imagined violence and anger rumination were highly inter-related. When imagined violence was added to the regression model, it was a significant predictor of prehospitalization violence; however, it did not moderate the association of anger rumination with pre- or posthospitalization violence. CONCLUSION Anger rumination may be a mechanism through which anger activates violent behavior, which has important implications for psychotherapeutic intervention targeting. Future research should investigate the association between anger rumination and imagined violence, with attention given to revenge planning as a link.