| A quasi-experimental design is utilized to examine three-year recidivism outcomes for two groups of nonviolent, repeat felony offenders arrested for low-level drug sales: experimental subjects diverted from prison to a community-based, residential drug treatment alternative to prison program, and comparison subjects (matched offenders, sentenced to prison terms equivalent to successful treatment completion length). Subjects were tracked for three years from treatment admission (experimentals) or release from prison (comparisons), using official records. Multiple measures compare the impact of residential drug treatment on recidivism patterns over time, including rearrest, roconviction, reinicarceration, probation and parole violations, and survival time to first rearrest. Analyses include 1) descriptive comparisons of prevalence rates for various time frames, 2) multivariate analyses of prevalence rates using logistic regression, and arrest rates using OLS regression, and 3) Cox proportional hazard models of time to rearrest. Control variables are included from the following domains- demographics, family, social, educational, employment, medical, psychological, criminal, sexual behavior, and drug use and treatment histories, The data indicated, a significant overall impact of treatment on recidivism, although the effect decays somewhat over time. Time in treatment is significantly related to post-program recidivism. Other factors affecting recidivism for the experimental and comparison groups are presented. |
Updated 05/20/2006