This study sought to evaluate the effectiveness of the structured substance abuse treatment program at Barrett Juvenile Correction Center in Virginia by comparing the outcomes of youth admitted to Barrett with the outcomes of youth who were eligible for admittance to Barrett but were detained at one of the traditional juvenile correctional centers in Virginia. The effectiveness of Barrett's program was also assessed by comparing the outcomes of youth who were admitted to Barrett but who differed according to how many of the four phases of treatment, focused on modifying negative attitudes and behaviors, they completed. Barrett differs from the six other juvenile correctional centers in Virginia in that it provides a highly structured substance abuse treatment program to all admitted youth. Youth are considered for admission to Barrett if they are male, aged 11 to 18, have a sentence of six to 18 months, and have a recommended or mandatory need for substance abuse treatment as determined by the Reception and Diagnostic Center (RDC), which assesses youths' needs prior to sentencing. Barrett's treatment program takes a therapeutic community approach, which emphasizes altering negative attitudes and behaviors through the completion of four sequential phases of treatment. In contrast, the goal of the traditional institutions was to achieve public safety while meeting the disciplinary, medical, recreational, and treatment needs of the youth. These facilities offered some treatment programs but only on an "as needed" basis. The sample for this study consists of all 412 youth released from Barrett Juvenile Correctional Center from July 1, 1998, to June 30, 2000, and a matched sample of 406 youth released from other juvenile correctional centers in Virginia during the same period. The treatment staff at Barrett submitted information on youths' treatment progress at the time of discharge. The RDC provided demographic, criminal history, and assessment information for all youths. The Virginia Department of Juvenile Justice provided information concerning actual time served and recidivism at the juvenile level. The Virginia State Police supplied additional recidivism data, including information on adult recidivism. Parole officers also provided data on recidivism and on progress toward meeting the conditions of parole. Demographic variables included in the dataset are race of the offender and his age at commitment. Clinical variables for Barrett youth only are Substance Abuse Subtle Screening Inventory (SASSI) and Intelligence Quotient (IQ) scores, total number of categories for which the youth scored yes on the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV), the length of the sentence, whether the youth had a recommended or mandatory need for substance abuse treatment, and the highest phase of treatment completed. Parole officers supplied data at three, six, and 12 months after release on whether they judged youths to be currently using a substance and whether youths were meeting the conditions of parole. These conditions included curfew, counseling services, educational programs, the employment requirement, and the electronic monitoring requirement. Also included are arrests and substance-related charges as reported by the Virginia Department of Juvenile Justice, the Virginia State Police, and parole officers. A variable for total reconvictions is included as well.