Every day in courts across the country, judges look to psychiatrists and psychologists to try to determine the potential dangerousness of mentally disordered persons. But recent studies have shown that the ability to predict violence is severely limited - and experts tend to be wrong more often than they are right. This comprehensive review of research from 1970 to 1993 offers new empirical and theoretical work on mental disorder that may pave the way for more accurate predictions of violent behaviour. The contributors consider why there has been little progress in the field, citing the narrowness of previous research and the lack of a theory of mental disorder. They assess the predictor variables in four domains: an individual's dispositional tendencies, clinical and psychopathological factors, case histories, and contextual issues. And they show how to develop better instruments for evaluating the risk of violent behaviour.