Rape: Behavioral Aspects

In a recent meta-analysis of sixty-one follow-up studies of sexual offenders (N=23,393), the recidivism rate, on average, was low (13.4 percent) (Hanson and Bussiere). It would be impossible to abstract rates of recidivism specific to rapists, since the aggregated studies used highly heterogeneous samples. Hanson and Bussiere did identify, however, subgroups that recidivated at higher rates. These higher base-rate recidivists were those offenders who evidenced clear deviant sexual preferences and those offenders with known prior sex offenses. To a lesser extent, offenders with greater criminal histories had higher recidivism rates. Sex offenders who failed to complete treatment had higher recidivism rates than those who successfully completed treatment. In one of the very few recent studies that examined sexual reoffense rates for rapists separately, the rate of sexual reconviction after an average of fifty-nine months was .20 (Quinsey, Rice, and Harris). Notably, there were only twenty-eight rapists in the Quinsey et al. study. In a subsequent twenty-five-year follow-up of 136 rapists, the failure rates for sexual charges and convictions were .19 and .11 at Year 5, .26 and .16 at Year 10, .31 and .20 at Year 15, .36 and .23 at Year 20, and .39 and .24 at Year 25 (Prentky, Lee, Knight, and Cerce). Can we conclude anything? As critical as recidivism rates are for risk assessment, for evaluating treatment efficacy, for drafting rape laws, and for social policy, we have no reliable estimates. The principal problem is the extraordinary variability of procedures and methods used to calculate recidivism among the extant studies (Prentky et al.). Studies differ considerably with regard to the composition of the samples (e.g., relative proportions of rapists, extrafamiliar child molesters, and exclusive incest offenders), the criminal behavior domains considered (e.g., only sexual offenses, all violent offenses, any new offense, etc.), the legal definition of what constitutes recidivism (e.g., a new arrest, a new conviction, a parole violation, etc.), the sources of outcome data gathered to assess recidivism (e.g., court records, public safety records, parole and probation records, F.B.I. records, etc.), and the length of the follow-up period (i.e., follow-ups range from twelve months to thirty years).

In addition, the most common method of estimating recidivism is to calculate the simple percentage or proportion of individuals who reoffended during the study period. This method will underestimate the rate of recidivism, because some of those individuals who were in the community for a briefer period of time may still reoffend (e.g., not everyone in a thirty-six-month follow-up will have been in the community for thirty-six months at the time the study ends; some individuals may have been in the community only for twelve months). This problem is addressed by using survival analysis.