The performance on community supervision of shock incarceration graduates was compared to the performance on community supervision of prison parolees, probationers, and shock dropouts. Community performance was measured as the rate of rearrests, revocations for new crimes, and revocations for technical violations during a follow-up period of either 1 or 2 years of community supervision. Although the original research design stipulated a sample size of 100, final sample sizes varied. Two instruments were used, one to collect demographic information and the other to collect recidivism data. Evaluation results regarding recidivism were inconsistent across States and did not permit a conclusive answer regarding whether shock incarceration programs reduce recidivism. This finding should be expected, given that the programs differ considerably in the type of offenders participating, the selection process, program activities, and the intensity of community supervision. The evidence suggests that the selection process for shock incarceration is more likely to predict the lower recidivism rate for both graduates and dropouts compared to the other groups, rather than the program itself. Further, the components of community supervision also play a role in recidivism rates. There is no evidence that the shock incarceration experience in and of itself had an impact on behavior during community supervision. The issues raised in this study must be addressed through the use of an experimental design that involves random assignment. 11 tables and appended supplementary information