A total of 231 probationed sex offenders (assaulters/rapists, pedophiles, and exhibitionists) were randomly assigned to treatment (group therapy) or probation only. The recidivism history of each subject was examined 10 years after project participation began. Although 57 percent of the subjects were rearrested, only 11.3 percent were rearrested on a sex charge. More sex arrests were found in the exhibitionist's criminal history than were found in the subjects convicted of assault and pedophile behavior. The assaulters had many arrests for non-sex offenses. The perception of the sex offender as a person continually driven to perform aberrant sexual behavior is not supported by the findings. Of the 26 sex offender recidivators found in the study, 7 committed their first sex offense 4 years or more after group therapy or probation. Seven committed their first sex offense in less than 1 year following treatment. Criminal history was found to be the best predictor of recidivism, as those with a long history of sex offenses had a higher probability of recidivating regardless of the type of intervention. Offenders who were in a homogeneous therapy group (e.g., all rapists) and who attended regularly had a much higher rate of recidivism than any other treatment group. Recommendations include (1) the use of mixed or heterogenous groups in treatment, (2) the use of future research to formulate typologies of sex offenders, and (3) the need to match treatments to improved understanding of the etiology of sexual deviance and to implement and evaluate new treatment strategies. Tabular data and about 130 bibliographic entries are provided. (Author summary modified)