Reducing recidivism is a central goal of treatment programs for offenders. Preliminary evidence suggests that cognitive‐behavioral group interventions based on the National Institute of Corrections curriculum (Bush, Glick, & Taymans, 1997) may be effective in reducing recidivism rates among adult probationers. We evaluated the effectiveness of a program based on this curriculum among 167 high‐ and medium‐risk probationers assigned to this program and a comparison group of 120 high‐ and medium‐risk probationers matched on age and number of prior criminal charges. Improvements over prior studies included use of survival analytic methods and propensity score matching, a longer follow‐up interval, and examination of treatment effectiveness within ethnic groups. Relative to the comparison group, treatment group probationers were more likely to complete probation satisfactorily and survive longer before rearrest. Moreover, supplementary analyses suggested that ethnicity was associated with differences in intervention effectiveness. Treatment was predictive of lower recidivism rates among European Americans and African Americans but was less effective among Latino American probationers.