For this program evaluation, which utilized a randomized field experiment, two separate substudies were conducted: one investigated the "front door" Intensive Community Supervision (ICS) program that diverted prisoners into the community at the beginning of their prison terms, and the other studied the "back door" Intensive Supervised Release (ISR) program that provided enhanced supervision services for offenders who were just finishing their terms of confinement and had a residential mandate upon release from prison. The random assignment procedure began in October 1990 and continued until June 1992. Prison caseworkers in the Office of Adult Release (OAR) within the Minnesota Department of Corrections screened offenders for participation in the prison diversion program (ICS) according to established criteria. The RAND coordinator assigned offenders to the experimental program or to the control program (prison) by consulting a predetermined random list of assignments. For the ISR program, institutional caseworkers reviewed the treatment plans for offenders who were scheduled to be released from prison within the next six months. The same procedure for random assignment was used as in the ICS study. The final sample sizes were 124 in the ICS program and 176 in the ISR program. Parts 1 and 9, Background Data, include demographic information such as sex, race, education, marital status, number of dependents, and living arrangement at time of most recent arrest. Also included is information on the offender's prior employment history, drug use prior to drug treatment, status after random assignment, various probation/parole/release conditions ordered, and criminal record information for prior arrests, for the governing offense and for the offense immediately prior to the current prison admission. Each offender was also rated on various items relating to risk of recidivism and need for treatment. The 6-month, 12-month, and 13-month review data (Parts 2-4 and 10-12) record the same information for each month. Variables provide information on the current status of the offender, days under regular supervision, intensive community supervision, special services, electronic surveillance, detention or incarceration (jail or prison), and days on other status. Information was also recorded for each month during the review regarding number and type of face-to-face contacts, number and type of phone contacts, number of drug tests taken, number and type of monitoring checks performed, number and type of sessions in counseling, number of days job hunting or in training, hours of community service, number of days employed and amount of earnings, amount of fines and court costs paid, amount of victim restitution paid, and amount of probation fees paid. Because a large percentage of the ICS control offenders were expected to remain in prison during a 12-month follow-up (resulting in premature recidivism outcomes), recidivism data for all ICS offenders were collected for a period of 24 months after assignment to the study (Part 5). Part 5 contains up to three status codes and number of days at each status for months 1-25 for the ICS cases only. Also included is information on work release, violations of supervision, absconding, returns to jail, returns to prison, and other releases. Parts 6 and 13 provide drug violation data, including first and second type of drug, action taken, and number of days since random assignment. Parts 7 and 14 provide technical violation data, including technical violation, first, second, and third action taken, days from assignment to each action, and most serious action taken. Finally, Parts 8 and 15 provide arrest data, including arrest code, age at arrest, if convicted, conviction code, type of sentence, and age at disposition. Dates were converted by RAND to time-lapse variables for the public release files for purposes of time-at-risk analysis.