The evaluation found the Boys Town Model was well documented and theoretically based. There were clearly delineated job responsibilities, a strong emphasis on staff training, and the number of daily interactions met or exceeded program guidelines. Despite frequent fidelity review, the sites' fidelity clustered slightly below average. Program utilization was reduced by a national shift in juvenile justice philosophy away from out-of-home placement toward community-based interventions. The results support the conclusion that the Boys Town girls may be expected to have superior delinquent and sexual behavior outcomes one year after enrollment compared with girls who received traditional probation. As the level of program exposure was increased - whether through increased staff interactions, length of stay, or both - the propensity of girls to engage in subsequent delinquency was reduced. No significant impact for substance abuse, academic commitment, and employment attitude was found. As one of the more rigorous evaluations on short-term care for female offenders, this study provides evidence that such programs can be effective in improving certain behaviors. The authors recommend altering expectations of short-term residential programs so that such placements are used to stabilize the youth and their family and to conduct assessments for recommendations on future interventions and treatment. They also suggest using the Model to develop a community-based day treatment program. The comparison group was composed of girls on standard probation. The sample consisted of 365 (treatment=235; comparison=130) participants across three sites. Program impact was assessed through a series of sequential analyses: 1) exploring the differences in means between the two groups on pretreatment characteristics; 2) performing a series of difference-of-means analyses to test for the main effects of the intervention; 3) using regression models for factors other than the intervention that may affect the outcomes, and 4) using survival analysis to predict new arrests.