serve as a control for other randomized treatment groups, that is, in a situation where no one would get less than the standard postrelease treatment (see, for example, National Research Council, 2001).
Recidivism rates, defined as the probability that parolees are rearrested or returned to prison, are significantly different for different groups of parolees. They are lower for women than for men; lower for older than younger parolees; lower for people with relatively short criminal records; and lower for violent offenders than for property or drug offenders (Langan and Levin, 2002; Petersilia, 2003).
Black parolees have higher recidivism rates than white parolees for violent and property crime, but not for drug crimes (Rosenfeld et al., 2005; see, also, Langan and Levin, 2002; Solomon et al., 2005). We note, however, that the race difference in recidivism is smaller than the race difference in overall arrest or imprisonment rates.
Parolees released from prison for the first time have lower recidivism rates than those who have been released in the past and then returned to prison. This finding holds even when sex, age, race, criminal record, offense type, and other characteristics of parolees are controlled (Rosenfeld et al., 2005; Tonry, 2004). The cause of this difference has not been established, however. Selection may play a major role; past failure at reentry predicts future failure. It also is possible that parole authorities and the police supervise and watch “two-time losers” more closely or are less willing to overlook any violations of their parole contracts. The finding that past imprisonment predicts future rearrest and imprisonment is consistent with the idea that the prison experience itself is criminogenic, but, recidivism does not appear to be related to the length of time an individual spends in prison (Rosenfeld et al., 2005). Another possibility is that people who have been imprisoned multiple times possess unmeasured traits or deficits that impede desistance. At present, the simple conclusion one can draw from what is known is that past recidivism predicts future recidivism.
One of the most significant findings that emerges from our work is that the peak rates for recidivism occur in the days and weeks immediately following release. Arrest rates decline over time after release from prison, especially for property and drug crimes. Moreover, death rates for new releasees—within the first days and weeks—are much higher than for matched demographic groups in the general population. This is a new research finding, the importance of which is underscored by the fact that the causes of death for parolees and inmates are different. For the state prison population, the leading causes of death are disease related: cardiovascular disease, cancer, liver diseases, and AIDS-related illnesses. For the releasee