Connecticut General Assembly

OFFICE OF LEGISLATIVE RESEARCH

December 11, 1997 97-R-1451

TO:

FROM: Lawrence K. Furbish, Assistant Director

RE: Recidivism Rates for Treated and Untreated Pedophiles

You asked if studies showed any difference in the recidivism rates of pedophiles (child molesters) who received sex offender treatment and those who do not. You also asked what percentage of sex offenders, including pedophiles, seek or receive treatment.

SUMMARY

The material we examined suggests that there is no solid scientific evidence from controlled experimental studies that clearly proves treatment programs reduce sex offender recidivism. But some studies have shown that a particular treatment program may be associated with lower recidivism. Many of the sex-offender recidivism studies appear to have methodological or design weaknesses. Investigators seem to agree that more research needs to be done on this issue.

Recidivism rates have been reported ranging from less than 10% to around 50%. Some studies have reported no statistically significant differences in recidivism rates between offenders receiving treatment and those that do not, while others have concluded that certain types of offenders who receive treatment are less likely to re-offend.

It is difficult to compare studies because of their differences in defining sex offenders and offenses, types of treatment programs, measures of recidivism, and how long offenders are tracked after release. Some critics believe that significant ethical issues exist concerning how a truly scientific offender recidivism study would have to be structured.

We were unable to find any reliable estimates of what percentage of inmates seek or receive treatment. Dr. Brett Rayford, the chief of health and mental health services for the Department of Correction, informed us that about 150 inmates currently are receiving sex offender treatment, but that many more could or should be receiving treatment. He says that some would probably accept treatment if the department's resources allowed it to offer more, but many probably would not because of fear. Sex offenders, especially child molesters, are the most despised of inmates, and many offenders fear for their safety if their crime becomes generally known in the institution. Participation in a treatment program makes their offense visible to other inmates and thus decreases their sense of security.

SELECTION OF STUDIES

A great deal of material has been published about sex offenders. We used studies cited in a brief filed in 1996 by Assistant U.S. Attorney Daniel Alter in the Second Circuit Court of Appeals in Doe v. Pataki, (Docket Nos. 96-6249(L) and 96-6269). The case involved a constitutionality challenge to New York's sex offender registration and community notification law. The cited studies appear to be a good representation of empirical research currently being done in this area.

LITERATURE REVIEWS OF RECIDIVISM STUDIES

Two of the articles, one by Becker and the other by Song and Lieb, review other sex offender recidivism studies. They not only review the work done by others, but they also reach conclusions about the general state of knowledge in this area.

Becker

In “Offenders: Characteristics and Treatment” psychology and psychiatry professor Judith Becker of the University of Arizona expresses concern about the scarcity of controlled therapy studies and methodological problems with uncontrolled studies. Nevertheless she believes that major changes in treatment programs over the years have led to a few studies which provide optimism about the effectiveness of treatment for some offenders. She expresses a personal belief that “modern treatments . . . have a significant impact on the recidivism rate for many offenders.” But she admits “because of the need to follow treated offenders for several years in order to establish recidivism rates, it will be some time before this perception can be reliably demonstrated” (Becker, p. 185).

Becker notes that a small number of offenders voluntarily seek treatment. She states that no data is available on the number of people entering treatment under the different mechanisms she lists. This includes voluntarily seeking treatment prior to contact with the criminal justice system, treatment as part of a pretrial diversion program, treatment as a condition of probation, treatment in prison, treatment as a condition of parole, or treatment under civil commitment after release from prison.

Becker classifies treatment into five classes of therapies. Biological therapies include surgical and chemical castration and various androgenic medications that reduce the sexual drive. Psychoanalysis and related therapies can involve in-depth individual counseling, peer group activities, and behavioral therapies. Family therapy, which involves both individual and group sessions, is primarily used with incest offenders. Cognitive-behavioral therapies try to teach people to recognize their inaccurate beliefs and control their inappropriate behaviors. Finally, relapse prevention programs focus on helping molesters recognize the signs of their impending abuse and teaching them self management techniques.

Methodological concerns regarding many of the existing studies, according to Becker, prevent making firm statements about the efficacy of treatment for reducing recidivism. She believes that to scientifically assess recidivism rates after treatment it is essential to compare the behavior of treated offenders with the behavior of comparable untreated offenders. Ideally this should be done by randomly assigning offenders to treatment and non-treatment groups, but she objects to this on ethical grounds. She cites other methodology problems including samples based on certain types of offenders, samples of the most serious or the least serious offenders, studies ignoring attrition among program participants, and studies with short follow through.

Becker cites a 1989 analysis published in Psychological Bulletin of 42 published and unpublished sex-offender recidivism studies. It concluded that there was “as yet no evidence that clinical treatment reduces rates of sex reoffenses in general and no appropriate data” to judge if it is effective for particular types of offenders (Becker, p. 189). She also cites several other studies that are described below by Song and Lieb.

Song and Lieb

In January 1994 the authors prepared a review of sex offender recidivism studies for the Washington State Institute for Public Policy. They found that few studies had been conducted with scientific precision, which meant that only limited conclusions could be drawn from them. They noted that recidivism rates varied by type of offender, with rapists having higher recidivism rates than child molesters and incest offenders the lowest rates. Sex offenders with criminal histories reoffend at higher rates than first-time offenders, and some offenders, such as child molesters, can reoffend many years after an initial sex offense.

Regarding treatment and recidivism, they conclude that while “some studies have shown that particular treatment programs are associated with lower recidivism rates in certain types of sex offenders, there is a lack of solid scientific evidence (from controlled experimental studies) that clearly proves treatment programs reduce sex offender recidivism” (Song and Lieb, Review, p.12). They believe additional research is necessary to identify more effective methods of treating and supervising sex offenders and more accurate ways of predicting recidivism.

Song and Lieb cite a summary by Marshall and Barbaree (Handbook of Sexual Assault: Issues, Theories, and Treatment of the Offenders) of research done on recidivism rates among untreated sex-offenders. It concludes that incest offenders have the lowest rates (4 to 10%), rapists from 7 to 35%, child molesters of female victims from 10 to 29%, child molesters of male victims from 13 to 40%, and exhibitionists from 41 to 71%.

Despite their concerns about methodological limitations, the authors summarize several studies attempting to examine the effects of treatment on recidivism. A 1980 study of 260 mentally disordered sex offenders treated in a California maximum security hospital and compared with a group of untreated sex offenders released from prisons showed 15% recidivism for the treatment group compared with 25% for the untreated group. The greatest difference was for child molesters offending against boys (15% for treated and 38% for untreated). For child molesters offending against girls, the difference was not statistically significant (20% for treated and 18% for untreated). The follow-up period was five years and recidivism was measured by reconviction.

Several studies done in Ontario, Canada show mixed results. A 1979 study showed an 11% reconviction rate for treated offenders and a 35% rate for the comparison group of inmates released from the same institution before the treatment program was available. A 1984 follow-up evaluation showed that treatment was more effective for child molesters than rapists.

In 1988 a study of 126 child molesters in Ontario with an 11-year follow-up found a 13% recidivism rate for treated offenders and 35% for the untreated group. The 126 offenders had all admitted their problem and sought treatment, but only 68 had received it. In this study recidivism included unofficial estimates by social service agencies and patient self-reports, as well as rearrests and reconvictions.

A 1991 Ontario study of 136 child molesters released from a Canadian maximum security psychiatric institution compared 50 who had received treatment with 86 who had not, with a follow-up over six years. It found a reconviction rate of 38% for the treatment group compared with 31% for the non-treatment group, a difference that was not statistically significant.

In 1992 a long-term study (a follow-up of from 10 to 31 years) of 197 child molesters was done in southern Ontario. The offenders consisted of 106 who received treatment between 1965 and 1973, 31 who were incarcerated in the prison before the treatment program began, and 60 who were in the institution when the treatment program was in effect but who received no treatment. The recidivism rates of the treatment and control groups were not statistically significant (44% for the treatment group, 48% for the first control group, and 33% for the second).

In February 1995 Song and Lieb also completed a recidivism study of Washington state sex offenders. One of their conclusions is that people who complete the Special Sex Offender Sentencing Alternative (SSOSA) program are less subject to re-arrest than a comparable group. SSOSA is a community treatment sentence given to certain first-time sex offenders. The comparison groups for this study were people who were eligible for SSOAS but who were instead incarcerated and people ineligible for the program. Recidivism was 11% for people completing the program, 14% for people eligible but who were not granted it, and 31% for people not eligible.

The authors also examined a sex offender treatment program operated by the Washington Twin Rivers Correction Center. The study compared the first 119 offenders to complete the program with 159 sex offenders incarcerated during the same time period who didn't receive treatment. Over a three-year follow up, they found an 11% recidivism rate for those receiving treatment and a 12% rate for those with no treatment, a difference that was not statistically significant. The authors also pointed out that because the treatment program only accepts offenders who admit their guilt and voluntarily request treatment, there may have been a significant difference between the treatment group and the comparison group that was unrelated to the effects of the treatment program.

Alaska Study

We located an executive summary of a sex-offender study done by the University of Alaska at Anchorage, and we are attempting to get a copy of the complete study. It compares 411 male sex offenders who received some level of treatment in the Hiland Mountain Sex Offender Treatment Program with 74 offenders who wanted treatment but could not get it, 100 randomly selected sex offenders who did not seek treatment, and 100 randomly selected nonsex-offenders. The Hiland Mountain program uses the relapse prevention model and houses about 85 inmates for a four-stage program using cognitive-behavioral treatment. The study concluded that treated sex offenders lasted longer in the community before reoffense than did any of the comparison groups. It also found that longer treatment resulted in longer periods before reoffending. The authors concluded that treatment can and does work, at least for some offenders, and that “offenders who are amenable to treatment and willing to actively participate learn to recognize precursors to relapse and self-manage their high risk behavior.”

LKF:lc