Sex Offender Treatment Efficacy
This paper discusses a prospective replication study by Seager, Jellicoe, and Dhaliwal taken in part from Debra Jellicoe’s Bachelor of Arts Thesis completed at the University of Alberta, Canada in 2000. Prior to this study, there was concern among researchers regarding the validity of results indicating sex offender treatment efficacy due to treatment dropouts and refusers not being part of control group calculations. Seager, Jellicoe, and Dhaliwal’s study attempt to address this gap in the research by designing a program that took into consideration dropouts and refusers in the control group. This essay begins with a summary of the scientific methods and findings of the study followed by a review of some of the research’s strengths and limitations. The practical, followed by the theory and policy implications of this study are described. Finally, the study will be situated within the broader field of literature of the time looking at two prior and one subsequent study. Throughout this essay prisoners that have participated in the sex offender treatment, the program is referred to as subjects. This study ultimately found that there was no significant reduction in future offending following sex offender treatment.
Seager, Jellicoe, and Dhaliwal were concerned with the validity of research that purported the efficacy of sex offender treatments to reduce recidivism. To investigate this issue they developed a sex offender treatment program manual based on the relevant theories from the 1990s and implemented their program at a medium-security Canadian prison at which 65% of prisoners reported being of Aboriginal descent. One hundred and seventy-seven of the adult male prisoners were offered the opportunity to complete the program. At the time of follow-up, 109 men had completed the treatment, and 37 did not, either refusing or dropping out of the program. Treatments were delivered by two professionals to groups of 10 male subjects. The program consisted of five distinct yet connected parts: disclosure, victim empathy, issues of masculinity/femininity, sexuality, and relapse prevention. Treatments were delivered over half-day periods, five days per week from 1995 until 1999 for a total of 275 hours. The risk of recidivism was assessed in 2000 using the Static 99 (a 10-item actuarial risk assessment scale used to assess the risk of sexual/violent recidivism). Recidivism was defined as a conviction or charge for sexual and violent crime in the two years following a subject’s release from prison. Pearsons Chi-square and Fishers Exact tests (1-sided) were used to compare later charges and convictions with rates of predicted reoffending based on offenders that successfully completed, unsuccessfully completed, dropped out, refused to participate, terminated, completed, and didn’t complete the program and in relation to the five areas of treatment.
Researchers hypothesis the following four outcomes in relation to recidivism: i) positive evaluations in certain parts of the program would forecast lower rates of recidivism; ii) passing of “a final treatment evaluation” would forecast lower re-offending; iii) successful completion of the program would forecast lower rates of re-offending, and iv) treatment completers would re-offend less than those that did not complete or participate. Contrary to the first hypothesis, positive evaluations gained in certain parts of the program by subjects that completed the program did not forecast lower rates of reoffending. Further, contrary to the second and third hypotheses It was found that whether subjects successfully or unsuccessfully completed treatment recidivism was not reduced. Finally, results supported the fourth hypothesis, those that did not complete or were terminated from the program reoffended 6% more than those that completed the program. Once dropouts and refusers were accounted for in the control group, results indicated a lack of any significant treatment efficacy.
A particular strength of this study is that it successfully addresses issues raised by previous researchers that potentially undermine the validity of previous research outcomes. For example, the unconsidered effects of refusers and dropouts in control groups and the lack of use of actuarial measures to match more than three risk factors between control and treatment groups. This study successfully contributes to a better understanding of the efficacy of sex offender treatments by addressing these failings of previous research.
Several limitations are also found in this study. Firstly, there is little ecological validity, as was noted in the article the small sample size was insufficient to accurately demonstrate treatment efficacy or predict recidivism. This limitation should be addressed in future research by using a significantly larger sample size. Secondly, sample bias occurred as the majority of those treated were males of aboriginal descent from Canada, thus as this population is not representative of the general population results are not generalizable. Future researchers should repeat this study using sample diversity to more accurately reflect the general population using randomized control trials for participant selection as well as individuals from different geographic locations. Thirdly, there is no indication of the subject’s motivations for treatment. It is possible that subjects that participated were motivated by a desire to reform and thus would be less likely to re-offend regardless of attending treatment. Future studies could address this issue by gathering and evaluating information from both participants and non-participants regarding their motivation for treatment. Finally, this study had a follow-up period of two years. It is possible that with a longer follow-up time a more accurate picture of rates of recidivism would become apparent. It is therefore advisable that future research repeats this study with a longer follow up time-frame.
Seager, Jellicoe, and Dhaliwal do not discuss policy or the theory guiding their treatment program, therefore, policy or theoretical implications are not discussed in this essay. However, the findings from this study revealed the following practical. Firstly, it would be valuable if replication studies used this same program with more culturally diverse populations to discover if the findings are generalizable. Although results indicated a lack of sex offender treatment efficacy, the population was predominantly of Canadian aboriginal descent, and results may differ with different populations. Secondly, as it was found that subjects that were terminated from or did not complete treatment reoffended more than those that completed treatment it would be valuable if future research investigated the motivations of both groups to better understand what leads those that opt to participate in treatment and to understand what motivates those that do not participate to re-offending. Finally, this study was one of the first to use an actuarial measure to match treatment and control groups on more than three risk factors. This study paved the way for future researchers to continue using actuarial assessment tools for risk assessment such as the Static 99 or the Sex Offender Risk Appraisal Guide (SORAG) to match treatment and control groups for risk and other factors to expand knowledge regarding sex offending risk prediction.
Concerns raised in previous studies led Seager, Jellicoe, and Dhaliwal to complete this study. For example, Hanson and Bussieres found that success in treatment was not reducing later recidivism. In their study, Hanson and Bussiere performed a meta-analysis of sexual offender recidivism studies to discover what factors were most strongly related to sexual re-offending using data from 61 studies from countries around the world from between 1943 to 1999. Results indicated that while on the whole recidivism by sexual offenders is low, there is a subgroup of high recidivists for which deviant sexual interest is the best measure of possible sexual offense recidivism. Further, while Hanson and Bussiere found that treatment completers reoffended less than those that did not complete treatment they were concerned that researchers were not comparing treatment and comparison groups based on their risk factors. Seager, Jellicoe, and Dhaliwal attempted to address this issue in their study by using the static 99 which allowed them to match the treatment and the control group on 10 risk factors. Secondly, a study by Seto and Barbaree similarly found that successful behaviors in treatment did not necessarily lead to a reduction in later sexual re-offending. In this study, the treatment behavior of 283 sex offenders treated at the Warkworth Sexual Behaviour Clinic between 1989 and 1996 was examined. Subjects ranged from 20 to 68 years. Successful behaviors were considered “appropriate behavior in group sessions, good homework assignments, and positive ratings of motivation and overall change”.
Later, Colton, Roberts, and Vanstone while they were similarly concerned with the efficacy of treatment programs at reducing recidivism, they took their research in a different direction. This study considered the opinions of 35 convicted of child sex offenders regarding their ‘awareness/empathy, understanding of why they offended, relapse prevention, awareness of their behavior, and accepting responsibility’. For this qualitative study Colton, Roberts, and Vanstone interviewed subjects at a prison in the United Kingdom that participated in or had completed sex offender treatment programs asking subjects for their opinions regarding their experience of treatment. Subjects all Caucasian European males ranged in age from 22 to 58 years. Victim empathy was found to be the most important element to address in treatment to reduce reoffending outcomes. As with Seager, Jellicoe, and Dhaliwal’s study, there were methodological limitations such as small sample size and a non-random sample, the researchers advised that further research in relation to offenders’ perspectives of their treatment experience would add to existing knowledge bases.
In conclusion, this essay discussed Seager, Jellicoe, and Dhaliwal’s study which found that sex offender treatment programs do not reduce sexual and or violent offending. Firstly, the method and results of their study were summarised. Secondly, the strength, limitations, and how they might be overcome were discussed. Thirdly, the practical implications of the study were described followed by a placing of their study in the broader field of literature looking at two prior and one later study. In conclusion, due to the limitations of Seager, Jellicoe, and Dhaliwal’s study, further replication studies would be advisable to discover if sex offender treatments are not effective in reducing future offending.