|| Checking for direct PDF access through Ovid
The purpose of this study was to assess the utility of social cognition and insight in the prediction of violence in a psychiatric inpatient sample. Violence history, demographic information, symptomatology, neuropsychological functioning, social cognition (i.e., attributional style), and insight were assessed in 29 inpatients with severe mental illness. Greater posttest violence was associated with greater pretest violence, less education, greater psychiatric distress, neuropsychological impairment, and hostile attributional and personalizing biases. Hierarchical multiple regression analyses showed that history of violence contributed the most variance to posttest violence. Hostile attributional and personalizing biases were also uniquely associated with posttest violence. Overall, this study supported the modest utility of attributional style measures in the prediction of inpatient violence. The predictive value of insight in this context appears limited.