The literature is first used to explain the technical problems associated with such predictions. This material is then used as a means of devising various steps that can be used to improve the reliability of such predictions and to reduce the error rates. The first chapter reviews the issues involved in debates on the validity and appropriateness of attempting the clinical prediction of violent behavior. Issues discussed are prediction in life and in law, definitions of dangerous and violent behavior, the criticism of prediction in law, and the moral and political issues raised by prediction. A chapter on the clinical prediction process in theory and in practice considers several concepts before it profiles the clinical prediction process: predictor and criterion variables, outcomes of positive and negative prediction, decision rules, and base rates. Another chapter reviews the research on the ability of psychiatrists and psychologists to predict violent behavior and discusses the criticisms and limitations of this research. In considering the steps that clinicians can take to improve the accuracy of their predictions of violent behavior, two chapters consider modifications of traditional clinical practice that hold promise for augmenting predictive validity. These are an increased emphasis on using statistical concepts in clinical prediction and a heightened sensitivity to environmental or contextual variables. The concluding chapter summarizes and synthesizes the previous material in a manner designed to help a mental health professional assess violence potential. It does so by presenting 14 questions for the clinician to consider in making a prediction. 230 references