The analysis concludes that to overcome the problems that have hindered the scientific study of violence among mentally disordered persons, it is necessary to enrich the predictor variables, strengthen the criterion variables, broaden the sampling strategy, and synchronize research efforts. If these actions are taken, the rejuvenated field of risk assessment may produce results quite different than those to which practitioners have become inured. If an actuarially valid array of risk markers for violence could be reliably identified, clinicians could be trained to incorporate these factors into their routine practice, and the accuracy of clinical predictions of violence among mentally disordered persons would be commensurately increased. Although questions of social policy and professional ethics would remain, relatively fewer people would be erroneously institutionalized as dangerously and relatively fewer people would be victimized by patients erroneously released or left untreated. 80 references