Methods: The sample (N = 228) was composed of African-American (70.5%) and Caucasian youth (26.0%) with a mean age of 16.26 (SD = 1.16). TC participants included boys (65.3%) and girls (34.7%) who primarily reported a history of marijuana (92.9%) use. Psychosocial data were collected at intake and monthly TC service data were collected throughout treatment. Post-release data were retrieved from official records. For multivariate analyses, correlates included measures of demographic and preincarceration characteristics, drug use and offense history, and TC treatment and institutional factors. Hierarchical multiple regression was used to examine predictors of treatment services, and binary logistic regression (LR) was used to determine variables that best predicted post-release supervision. For both multivariate models, predictor variables were entered in five blocks.
Results: Youth received, on average, six months (M = 5.85, SD = 4.21) of TC treatment and over half (60.8%) were placed under formal supervision upon release. Predictors of TC treatment services received included race, gender, marijuana use, offense type, psychosocial risk factors, disciplinary infractions, and number of visits. Program completion was included as an additional predictor of post-release supervision. Regression results showed that the independent variables were significant contributors to TC treatment services received (R = .45(2, 94), p<.05). Approximately 21 % of the variance in treatment services was explained by the inclusion of all predictor variables (R Square = .205). Among predictors, gender (t=-3.09,p<.01)) and disciplinary infractions (t=2.13,p<.05) explained a significant proportion of the variance, with girls receiving less TC treatment, and those receiving disciplinary tickets receiving more. The LR model was statistically reliable in distinguishing between youth who were and were not placed under supervision (-2 Log Likelihood=78.658;Wald X Square = 29.550,df=10,p<.01). Inclusion of all predictors explain3d approximately 40% of the variance in whether a TC participant was placed under post-release supervision (Nagelkerke R X Square =.400,p<.01). The model correctly classified 80.0% of cases. Among predictors, only the number of disciplinary tickets predicted whether a youth was placed under supervision (Wald Statistic = 7.6351, p<.01). As the number of tickets increased by 1, the odds of being placed under supervision decreased by .80, indicating little likelihood of change.
Conclusions: Results showed that girls received less treatment than boys, indicating a need to ensure equal access to TC treatment within juvenile institutions. Youth with more disciplinary tickets received more TC treatment, suggesting flexibility in TC programming within juvenile institutions. Recidivism and drug use outcomes should be examined among supervised youth with extensive drug use histories.
