The Wakefield Psychology Department's inmate assessment procedure combines an examination of past offending behavior with the monitoring of prison behaviors. The behavioral risk-assessment process has two underlying assumptions: that the best predictor of future conduct is generally how the individual has behaved in the past and that individuals behave with consistency across various situations. The process involves the identification of the behaviors, characteristics, and situational aspects that have contributed to the original offense and the monitoring of an inmate's behavior throughout sentence to determine if behaviors similar to those associated with the offense reoccur. The system has been used on one wing at Wakefield for almost a year. The hypothesis tested in this study is that in a sample of sex offenders, predictions of risk would be found more often than at a chance level. Two independent raters examined wing reports and identified actual behavior reported. Two additional independent raters judged whether the predicted and actual behaviors were the same. The study found that for all the sample, 62 percent of the predicted behaviors occurred. There was, however, significant variation between individual cases ranging from 19 percent to 90 percent. An analysis was conducted across subjects to examine what types of behavior were best predicted. It was found that 41 out of 65 types of behavior were predicted better than chance. It was also possible to describe what particular elements of offending behavior related to particular types of prison behavior. Thus far, the research demonstrates the utility of using behavioral indicators as part of a multimodal approach to the assessment of sex offenders. 2 tables