Impact of Drug Treatment on Florida Drug-Involved Probationers: A Propensity Score Analysis

Kevin Strom, Research Triangle Institute
Pamela K. Lattimore, RTI International

This paper provides findings from analyses that examined the impact of drug treatment on recidivism among a population of nearly 140,000 drug-involved probationers in Florida. These analyses are part of a NIDA-funded study of the impact of drug courts. Administrative data were provided by the Florida Department of Correction for all offenders committed to supervision between FY 1996 to FY 2000 (July 1, 1995 - June 30, 2000). Outcome data, including probation revocatio and commitment to prison, for these offenders have been acquired through September 2001, as have arrest data from the Florida Department of Law Enforcement (FDLE). The study is strengthened by the inclusion of an entire State's probationer population over a 5-year period, by the use of multiple data sources to track offenders over time and provide details on treatment and supervision, and by the use of multiple indicators for recidivism outcomes (i.e., arrest, incarceration, drug use).

This paper provides findings from a propensity score analysis that examined the effectiveness of drug treatment. Propensity score models were estimated with a logit model in which appearance at treatment was the outcome and a variety of factors hypothesized to be related to treatment were included as independent variables (e.g., race, age, current offense, UA screening required, type of supervision). Subjects were then grouped by the probability of receiving treatment generated by the estimated model (the "p-hat"). With 140,000 subjects, we grouped subjects by percentile (e.g., those with a likelihood of receiving treatment of between 25% and 26%). Analyses were then conducted within these percentile groupings to identify whether those who received treatment did better than those who did not (conditioned on equal lielihood of having been assigned to treatment). Results suggest that drug treatment reduced felony arrests and that the reduction was greater for those who received more than 90 days of treatment.

(Return to Program Resources)

Updated 05/20/2006