Currently, there is an apparent consensus among criminologists that the actuarial method (based on what has been scientifically and empirically proven) and clinical method (based on subjective observation and informal decisionmaking by experienced behavioral scientists), despite their inherent flaws, both contribute to the evaluation of violent behavior. Regarding recommended ways for communicating a person's risk of violence, the literature recommends two methods: making use of frequency scores and probability scores and using categories that vary from "low risk" to "high risk." Actuarial measuring instruments are particularly effective in predicting imminent, but less serious, violent reoffending, as well as violent behavior over the long term; however, the actuarial method of prediction is less accurate in cases of repetitive and serious sex offenses, and it cannot predict intrafamilial reoffending and the termination of reoffending. Clinical evaluations, on the other hand, are valuable in predicting behavior over the short term. A discussion of the evaluation process for predicting violent behavior addresses guidelines for the evaluation of potential perpetrators of violence and the effective communication of the results of such an evaluation. Six guidelines for evaluation pertain to information collection, the offender's insight regarding the occurrence and dynamics of his/her violent behavior, variables possibly linked to risky behavior, consultation with other professionals, development of a preventive/treatment plan, and the availability of sufficient time for a thorough evaluation. Some common errors in such evaluations are also identified. 96 references