The study focuses on all males 50 years old or younger who served at least 6 months in the North Carolina prison system and were released in the first half of 1980. Dependent variables were postrelease recidivism and employment. Independent variables included the prison programs of work release, education and vocational training, acquisition of a general education diploma, and participation in prison work activity. Other independent variables were inmate characteristics generally predictive of rehabilitative success, measures of the interaction between inmate characteristics and prison programs, and control variables. Control variables were race, whether or not the subject was a property offender, whether or not released under supervision, and two indices of the subject's home environment. Linear, OLS regressions were used. Although none of the programs had a significant overall effect on new arrests or incarcerations, significant effects were achieved for certain offender types in certain programs. Only participation in prison work activity had an overall favorable impact on employment or reported earnings, but certain offender types had favorable employment and earnings results in the other programs. The findings suggest that improved offender-program matching can enhance the overall effectiveness of prison programs. 2 tables and 6 references.