The sample used in the development of the risk scale consisted of 494 adult male sex offenders who were on probation, on parole, in community corrections (prison diversion), or in prison treatment between December 1, 1996, and November 30, 1997. The offenders had been convicted of the following types of crimes: sexual assault, sexual assault on a child, exposure, assault, kidnapping, and exploitation of a minor. Eighty percent of the sample consisted of adult male sex offenders who met the definition of Sexual Predator Crimes, i.e., convicted of one of five felony sex crimes. The intent of the study was to collect a large number of potentially predictive variables to be compared to outcome at 12 months. The most predictive set of variables would then be included in a risk tool that would subsequently be cross-validated on outcomes at 30 months. Assessed variables were in the categories of demographics; developmental history; educational background; sexual, juvenile, and adult criminal histories; and current offense characteristics. Dynamic indicators included interpersonal abilities, treatment variables, and patient behavior. The final risk scale encompassed prior juvenile felony convictions, prior adult felony convictions, failure of the first or second grade, being unemployed, victim being intoxicated, the perpetrator reporting not being sexually aroused during the crime, possession of a weapon during the crime, denial in therapy, sexual deviance in therapy, and motivation in therapy. The risk scale provided significant relative risk ratios relative to program failure at 12 and 30 months. Overall, those scoring high on the risk tool were 372 percent as likely to fail as those scoring low. 4 tables and 39 references