Clinical assessments were based on a 1-page instrument and medical records and court letters; subjects were followed for 24 months after assessment. Inter-rater agreement on dangerousness was good among professional groups, although social workers gave the highest mean dangerousness ratings and psychiatrists the lowest. Analysis of followup data for 408 subjects showing later criminal charges, misconducts during incarceration, and psychiatric hospital contacts indicates that the best background predictors of dangerousness outcome scores (DOS') were attitude/behavior, alcohol use, prior incarceration, and age. The best possible weighing of actuarial factors resulted in the correct classification of 63.7 percent of cases, with a ratio of false to true positives of about 3:2. A comparison of clinical assessments with DOS' indicates that overall, psychiatrists' assessments were most accurate. However, the size of correlations varied by individual clinician and type of subject, as well as by discipline. While many of the correlations were statistically significant, they tended to be below conventional standards. 7 tables, 2 figures, 8 notes, and 87 references.