Of 2,039 documents published in 5 languages, 69 studies with 80 independent comparisons between treated and untreated offenders fulfilled stepwise eligibility criteria (N=22,182). Most of the studies were done in North America, and nearly three-fourths of the studies had been published since 1990, although actual program implementation began much earlier. Nearly one-half of the comparisons involved cognitive-behavioral treatment programs. Fourteen comparisons addressed physical therapy, 8 of which dealt with surgical castration. Most treatments were designed specifically for sex offenders. Despite a wide range of positive and negative effects, the majority of evaluations confirmed the benefits of treatment. Treated offenders showed 37 percent less sexual recidivism than controls. Effects for violent and general recidivism were similar. Physical therapy (surgical castration and hormonal medication) showed larger effects than psychosocial interventions, although this difference was partially confounded with methodological and offender variables. Among psychological programs, cognitive-behavioral approaches showed the most robust effect. Nonbehavioral treatments did not have a significant impact. There was no outcome difference between randomized and other designs; however, group equivalence was linked with slightly larger effects. Various other moderating factors had a stronger impact on effect size, including small sample size, quality of outcome reporting, program completion compared with dropouts, age homogeneity, outpatient treatment, and authors' affiliation with the program. The meta-analysis concluded that more differentiated, high-quality evaluations are required to determine what works for whom and under what circumstances. 4 tables, 1 figure, and appended list of studies included in the meta-analysis