As part of ongoing program evaluation, reoffending was tracked through the Canadian Police Information Centre to determine sexual recidivism. The mean length of time from discharge from the inpatient program to follow-up was 106.7 months. Since both recidivism and testosterone levels among sexual offenders are related to age, the analysis controlled for this variable in the part of the study that involved recidivism. To investigate the influence of treatment efficacy on the link between testosterone and recidivism, the sample was subsequently re-examined on the basis of treatment completion or noncompletion. The findings show that men with higher testosterone levels tended to have committed the most invasive sexual crimes. Also, a positive partial correlation (controlling for age) between testosterone and sexual recidivism was found. When the sample was distinguished by completion and noncompletion of treatment, an important ameliorating treatment effect was found. Serum testosterone remained significantly predictive of sexual recidivism for those who did not complete treatment; whereas, for those who completed treatment, testosterone was no longer predictive of sexual recidivism. Apparently, effective therapy can disrupt the link between higher serum testosterone and the greater likelihood of further sexual violence. This suggests that the completion of intensive treatment should be accorded significance in the design of actuarially based risk-prediction instruments. 2 tables and 43 references