This study drew upon the social disorganization perspective to examine the contribution of individual-level and neighborhood-level risk factors for violence among a sample of individuals recently discharged from a psychiatric hospital after an acute inpatient stay. The basic premise of this book is that the social and organizational characteristics of neighborhoods explain and/or condition variations in the violent behavior of discharged psychiatric patients that cannot be attributed to their individual characteristics. The analyses combined individual-level data from the MacArthur Foundation's Violence Risk Assessment Study, the largest study of risk factors for violence among discharged psychiatric patients ever conducted, with objective data on the characteristics of the neighborhoods in which patients resided after discharge. The latter data were obtained from 1990 U.S. Census Summary Tape Files.^ The kinds of violence assessed included battery that resulted in injury, sexual assaults, assaultive acts that involved the use of a weapon, or threats made with a weapon in hand. The MacArthur Violence Risk Assessment Study was a 3-site, longitudinal study of discharged psychiatric patients (n=1,136) conducted between 1992 and 1995 (Steadman et al., 1998). An extensive battery of research instruments was administered to subjects during their psychiatric hospital stays and then up to five times during the first year following their discharge from the hospital. Bivariate analyses of the data produced consistent results. The following characteristics were found to be significantly associated with all three operationalizations of violence risk: Black racial status, substance abuse disorder, anger, paranoid disorder, psychopathy, history of violent arrests, and neighborhood disadvantage.^ Discharged patients with these characteristics tended to exhibit higher rates of violence, sooner after discharge, and with greater frequency. Being male and married showed significant positive associations for violence prevalence and timing, but not for numbers of incidents. Multivariate analyses indicate that neighborhood disadvantage enhances the prediction of patient violence over and above the effects of known individual-level predictors and control variables; no evidence of concentration effects was found. Although neighborhood disadvantage was found to have a significant main effect on patient violence, the individual-level risk factors analyzed accounted for the main variation in violence based on all three operationalizations of patient violence. The book concludes with a chapter that discusses implications of the findings for theory and research in the area of violence and mental illness.