Programs aimed at changing criminal conduct can be grouped into four categories: control, insight, experience, and opportunity. Although evaluation research that has tested these program types has been deficient in quality and quantity, it nevertheless indicates that each program type provides some guidance for reducing recidivism rates, but only for a limited segment of offenders. Deterrence of offenders by threat of punishment (the traditional type of control theory) is effective only with offenders with strong conditioning toward social normalization and little involvement in a criminal subculture. The most effective punishment for such persons does not impair their bonds with conventional society or increase their ties with other offenders. Where such offenders are incarcerated, visits and furloughs that facilitate contact with law-abiders tend to reduce recidivism. Insight by purely verbal methods was shown to work only with unadvanced offenders. Only a small proportion of such programs, however, are effective even with the unadvanced offender. Research indicates that most young felony offenders have failed to make the transition from childhood to economic self-sufficiency in the legitimate job market. Studies show that the recidivism rate with such offenders is most likely to be reduced through educational and vocational programs that increase their ablity to enter legitimate occupations. Aid to relieve desperate economic and social needs is also important. The corrections system, far from exhausting the aforementioned rehabilitation approaches to the extent that they can be reliably deemed fruitless, has yet to commit its resources thoroughly to the matching of proven programs with the types of offenders who have benefited from them in the past. Approximately 70 references are provided. (Author abstract modified)