The model evaluation instrument is based on a quasi-experimental design with nonequivalent comparison groups and statistical controls. A survey containing questions on family background, education, and work experience will be administered to nearly every inmate upon intake, and a similar survey will be administered to nearly every inmate upon exit. Attributes and attitudes of inmates will be captured. Intake and exit surveys incorporated in the instrument are intended to assure the instrument's validity. Steps involved in developing the instrument are detailed, including the development of recidivism measures and the tracking of released inmates. Measurable objectives of the instrument are delineated, as well as variables and information collection procedures. The role of case studies in validating the instrument is examined, and summaries of recidivism studies are appended. 8 references