This report suggests a research design for a 5-year investigation of the feasibility and of the effects of three differentiated educational programs on the rate of recidivism for prison inmates. Eligibility for the program will be based on possession of a high school diploma or equivalency certificate, desire to participate in the program, and classification as a "low risk" prisoner. Three methods of instruction will be used concurrently during the program. One allows a small number of inmates to receive their instruction on the campus of Auburn Community College (New York). In addition to the three groups receiving instruction, there will also be control groups to provide comparative data. Assignment to treatment and control groups will be based on procedures designed to eliminate bias. Data to be collected include: (1) relative program costs; (2) recidivism rates between participants and eligible non-participants, and those associated with various treatment methods; and (3) relative influence of student characteristics and the educational program on recidivism rates. A large data base will be assured by randomly assigning new arrivals at the prison, who meet program entry requirements, to either the treatment or the control groups every four months. (JO)