In November 2005, a Police Officer’s wife obtained information that a sexual offender was due to be released into a beachside community in Christchurch, New Zealand. Consequently, she initiated a pamphlet-drop around the community warning residents of the offender’s imminent release, urging them to keep their neighborhood safe and asking them what they would do to “stop him from coming”. Not long after this, another sexual offender was released into a small community on the west coast of the South Island of New Zealand but, within days, residents forced him away through consistent protesting. These events were catalysts leading me to research the effects of poor reintegration planning on sexual recidivism.
There is a widespread belief among clinicians and other professionals working in correctional settings that poor planning for the transition to living in the community after prison release can increase the likelihood of recidivism for sexual offenders. The difficulties such offenders face in finding suitable housing and employment have been reflected in news items across the world and elsewhere in the popular media, including the critically well-received film, The Woodsman. Yet, despite a growing body of research on risk assessment (Beech, Fisher, & Thornton, 2003), there has been no systematic investigation to date of the effect of community reintegration planning on sexual offender recidivism.
Acute dynamic risk factors (see Hanson & Harris, 2000) are those that can potentially change rapidly and precipitate recidivism (e.g., negative mood, substance abuse, anger/hostility, and victim access). Hanson and Harris (2002) conducted interviews with community supervision officers for matched control groups of recidivists and non-recidivists and found that recidivists’ mood decreased significantly in the month immediately prior to their reoffending. They also found that anger, substance abuse, and access to victims significantly increased for recidivists in that month prior to recidivism. This leads us to question whether acute dynamic risk factors might be triggered by specific environmental contexts, for example, living near a park where children frequently play (victim access) or in a neighborhood in which drugs are widely available (substance abuse). The importance of environmental context in ameliorating such risk factors suggests that effective planning for the reintegration of offenders back into the community may be crucial. Successful planning would facilitate the offender’s re-entry to environments in which the impact of acute dynamic risk factors would be minimized.
What are the needs of released sexual offenders that warrant consideration in planning a sexual offender’s re-entry to the community? Although research investigating the needs of released offenders is limited (Taxman, Young, & Byrne, 2002), the following domains have been consistently identified as potential barriers to successful re-entry to the community: (a) Individual needs (including physical and mental health needs, and offense-specific treatment needs), (b) social needs, (c) accommodation needs (i.e., finding a place to live), and (d) employment needs (Graffam, Shinkfield, Lavelle, & McPherson, 2004; Petersilia, 2003). Sexual offenders likely have needs above and beyond these. The Good Lives Model (GLM―Ward & Stewart, 2003) provides a useful theoretical framework for sexual offender reintegration planning. The GLM is a contemporary model of offender rehabilitation that differs, to a degree, from the prominent Risk-Need-Responsivity (RNR―Andrews & Bonta, 2003) and Relapse Prevention (RP―e.g., Laws, 1989) models through its emphasis on the promotion of primary goods as treatment targets. According to the GLM, primary goods are basic human needs that, once met, enhance psychological well-being. Examples include forming pro-social relationships, excellence in play and work, and inner peace (see Ward & Brown, 2004). The weightings or priorities given to specific primary goods reflect an offender’s life values. Instrumental or secondary goods provide concrete means of securing primary goods, and take the form of approach goals (Ward, Vess, Collie, & Gannon, 2006). Various problems have been noted in sexual offenders’ pursuit of primary goods, including the use of inappropriate secondary goods to obtain primary goods. For example, sexual offending may be an attempt to form an intimate relationship (Ward & Gannon, 2006). Incorporating GLM principles into reintegration planning likely enhances an offender’s motivation for change, because the focus reflects the offender’s priorities in life (Ward, Day, & Casey, 2006).
The goal of the present study was to systematically investigate the quality of community reintegration planning for sexual offenders, and to determine the extent to which poor planning might represent a risk factor for sexual recidivism. We studied a sample of sexual offenders against children released from Kia Marama, a prison-based treatment program for child molesters in New Zealand (Hudson, Wales, & Ward, 1998). A coding protocol was developed for reintegration planning that included items relating to accommodation, social support, individual needs, employment, and GLM secondary goods. The protocol was then applied retrospectively to the cases in the sample. We hypothesized that recidivists would have poorer release planning than non-recidivists. In addition, we had measures of IQ and stable dynamic risk factors based on a psychometric battery (Allan, Grace, Rutherford, & Hudson, 2007). Provided that reintegration planning was better overall for the non-recidivists, we were interested to know whether any differences remained significant after controlling for these other measures.
The sample for this study was drawn from males who completed the 32-week prison-based treatment program at the Kia Marama Special Treatment Unit between 1990 and 2000. Our sample included 39 recidivists (men who had been reconvicted of a sexual offense as of February 2001) and a control group of 42 non-recidivists. Non-recidivists were individually matched with the recidivists for static risk level and time at risk.
Static risk level. The Automated Sexual Recidivism Scale (ASRS; Skelton, Riley, Wales, & Vess, 2006) was used to measure static risk. The ASRS is based on the Static-99 (Hanson & Thornton, 2000), which is the most widely-used and validated measure of static risk for sexual offenders (Ducro & Pham, 2006; Hanson & Thornton, 2000; Looman, 2006). The ASRS is computer-scored based on information stored in a database maintained by the New Zealand Department of Corrections, and includes seven of the 10 items from the Static-99 (excluded were items #6―any unrelated victim; #7―any stranger victim; and #10―single/ever lived with a lover for at least two years). The overall ASRS score is divided into four risk bands, which correspond closely to those associated with the Static-99. Skelton et al. showed that the ASRS had comparable predictive validity to the Static-99 for sexual recidivism in a sample of male sexual offenders (N = 1133), with the receiver operating characteristic areas under curve (AUC) ranging from .70 to .78.
Time at risk. Time at risk was measured from the date participants were released from Kia Marama until criminal history records were obtained in February 2001.
Recidivism. Criminal history information was obtained from the computer database maintained by the New Zealand Department of Corrections as of February 1, 2001. Any convictions for sexual, violent, or general offenses that occurred post-release were noted. Sexual recidivism was defined as Category ‘A’ offenses according to the Static-99 scoring criteria (see Harris, Phenix, Hanson, & Thornton, 2003)―that is, an offense with an identifiable victim (e.g., incest, sexual assault, exhibitionism). Category ‘B’ offenses (i.e., no identifiable victim) were excluded, except for possession of child pornography. Violent recidivism was recorded when the offender had been convicted for a non-sexual offense against a person (e.g., assault, robbery, kidnapping). General recidivism was defined as an offense that was neither sexual nor violent (e.g., possession of cannabis).
Release planning. We developed a coding protocol to measure aspects of release planning, based on the available research on needs of released prisoners. Items contained in the coding protocol are described below, with the scale for each item indicated in parentheses:
1. Accommodation (0 – 2). This item measured the extent of accommodation planning. The proposed type of accommodation was recorded (e.g., hostel).
2. Social support (0 – 3). This item measured whether a social support network had been established and, if so, how many systems it comprised.
3. Idiosyncratic risk factors (0 – 3). This item assessed whether high-risk situations or warning signs were indicated and, if so, whether an attempt had been made to minimise these through reintegration planning.
4. Employment (0 – 3). This item measured the extent of employment planning.
5. GLM secondary goods (0 – 1). This item indicated whether or not secondary goods were present in an offender’s reintegration plan. GLM secondary goods were defined as concrete approach goals that had been identified by the offender (rather than suggested by the therapist), relating to one of the nine primary goods listed by Ward and Brown (2004). The best fitting primary good(s) targeted were recorded.
6. Motivation (0 – 1). This item indicated motivation to follow through with post-release plans, as stated by the therapist.
Files held by the Department of Corrections Psychological Service were accessed for each participant, and the report for each participant written by Kia Marama staff to the Community Probation Service upon release was rated using the coding protocol developed for this study. These reports typically contained details relating to the offender’s conviction, a summary of assessment findings and treatment outcomes, an indication of current risk level, a list of high risk situations and warning signs, and an outline of release plans.
Approximately 30 percent of reports were double-coded, to obtain a measure of inter-rater reliability. Data coders were blind to the recidivism outcome for each participant, and rated release plans independently of each other. All disagreements between coders were resolved by consensus.
The coding protocol demonstrated adequate inter-rater reliability, with an average Cohen’s k of .83. We conducted a two-way ANOVA to determine whether ratings on the coding protocol differed depending on recidivism status and offender type (intra- vs. extra-familial offender). There were no significant effects for offender type, or significant group by type interactions. All non-recidivists received the maximum score for accommodation. Mean scores for accommodation, employment, and GLM secondary goods, as well as the release plan total score, were significantly greater for the non-recidivist group. The difference for social support approached significance (F[1, 77] = 2.95, p < 0.10). To summarize, the overall quality of reintegration planning was substantially better for non-recidivists than recidivists, but did not differ between intra-familial and extra-familial offenders.
Recidivists had significantly lower IQ scores and higher overall deviance scores (as measured by a psychometric battery―see Allan et al., 2007) than non-recidivists. We conducted several analyses of covariance (ANCOVAs) to determine whether the difference in reintegration planning scores between recidivists and non-recidivists would remain significant after controlling for IQ and overall deviance. When IQ and overall deviance were simultaneously controlled for, the mean score for accommodation remained significantly higher for non-recidivists than recidivists (F[2, 75] = 3.19, p < .05). No other significant group differences remained.
When any recidivism (i.e., reconvictions for sexual, violent, or general offenses) was used as an outcome variable (recidivists N = 51, non-recidivists N = 30), the accommodation and GLM secondary goods items, as well as the release plan total score, were significantly correlated with recidivism. As expected, the recidivists had significantly lower IQ scores and higher overall deviance scores. When IQ and overall deviance were simultaneously held as covariates, GLM secondary goods remained significant (F[1, 76] = 4.08, p < 0.05).
DISCUSSION AND CONCLUSIONS
The major goal of the present study was to investigate whether the quality of reintegration planning for child molesters was related to sexual recidivism. Overall, the quality of reintegration planning was poorer for the recidivists in this study, who had significantly lower scores for accommodation, employment, and GLM secondary goods, as well as for the total reintegration planning score, when compared to non-recidivists. All non-recidivists in our sample received the maximum score for the accommodation item, indicating that planning for post-release accommodation needs was necessary but not sufficient to prevent sexual recidivism. When differences in IQ and overall deviance between both groups were controlled for, accommodation planning remained significantly poorer for recidivists, suggesting this was the aspect of reintegration planning that was most strongly linked to sexual recidivism. Ironically, as a community, we make finding a place to live one of the greatest challenges for released child molesters (in regard to residency restrictions, public notification, and similar practices).
Interestingly, when any recidivism served as the outcome variable, GLM secondary goods remained significantly higher for non-recidivists when IQ and overall deviance were simultaneously controlled for. This finding suggests that the presence of GLM secondary goods could represent a protective factor against any recidivism. This is consistent with the holistic and strengths-based nature of the GLM (e.g., Ward, Day et al., 2006).
The mechanisms through which aspects of reintegration planning are linked to recidivism need to be further explored. A limitation of the current study was its exclusive focus on planning for reintegration; whether or not such plans are implemented effectively is obviously important and needs to be addressed in future research. As this study has illustrated, successful community reintegration has the potential to reduce the risk of sexual recidivism. Hence, reintegration planning should be an integral component of treatment programs for child molesters, and should also be addressed for those incarcerated child molesters who do not participate in prison-based treatment programs.
Allan, M., Grace, R., Rutherford, B., & Hudson, S. (2007). Psychometric assessment of dynamic risk factors for child molesters. Sexual Abuse: A Journal of Research and Treatment, 19, 347-367.
Andrews, D. A., & Bonta, J. (2003). The psychology of criminal conduct (3rd ed.). Cincinnati, OH: Anderson Publishing.
Beech, A. R., Fisher, D. D., & Thornton, D. (2003). Risk assessment of sexual offenders. Professional Psychology: Research and Practice, 34, 339-352.
Ducro, C., & Pham, T. (2006). Evaluation of the SORAG and the Static-99 on Belgian sexual offenders committed to a forensic facility. Sexual Abuse: A Journal of Research and Treatment, 18, 15-26.
Graffam, J., Shinkfield, A., Lavelle, B., & McPherson, W. (2004). Variables affecting successful reintegration as perceived by offenders and professionals. Journal of Offender Rehabilitation, 40, 147-171.
Hanson, R. K., & Harris, A. J. R. (2000). Where should we intervene?: Dynamic predictors of sexual assault recidivism. Criminal Justice and Behavior, 27, 6-35.
Hanson, R. K., & Thornton, D. (2000). Improving risk assessments for sexual offenders: A comparison of three actuarial scales. Law and Human Behavior, 24, 119-136.
Harris, A., Phenix, A., Hanson, R. K., & Thornton, D. (2003). STATIC-99 coding rules revised - 2003 Ottawa: Department of the Solicitor General of Canada.
Hudson, S., Wales, D., & Ward, T. (1998). Kia Marama: A treatment program for child molesters in New Zealand. In W. L. Marshall, Y. M. Fernandez, S. Hudson & T. Ward (Eds.), Sourcebook of treatment programs for sexual offenders (pp. 17-28). New York: Plenum Press.
Laws, D. R. (1989). Relapse prevention with sexual offenders. New York: Guilford Press.
Looman, J. (2006). Comparison of two risk assessment instruments for sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 18, 193-206.
Petersilia, J. (2003). When prisoners come home: Parole and prisoner reentry. New York: Oxford University Press.
Skelton, A., Riley, D., Wales, D., & Vess, J. (2006). Assessing risk for sexual offenders in New Zealand: Development and validation of a computer-scored risk measure. Journal of Sexual Aggression, 12, 277-286.
Taxman, F. S., Young, D., & Byrne, J. M. (2002). Offender's views of reentry: Implications for processes, programs and services Maryland: University of Maryland, Bureau of Governmental Research.
Ward, T., & Brown, M. (2004). The Good Lives Model and conceptual issues in offender rehabilitation. Psychology, Crime & Law, 10, 243-257.
Ward, T., Day, A., & Casey, S. (2006). Offender rehabilitation down under. Journal of Offender Rehabilitation, 43, 73-83.
Ward, T., & Gannon, T. A. (2006). Rehabilitation, etiology, and self-regulation: The comprehensive good lives model of treatment for sexual offenders. Aggression and Violent Behavior, 11, 77-94.
Ward, T., & Stewart, C. A. (2003). The treatment of sexual offenders: Risk management and good lives. Professional Psychology: Research and Practice, 34, 353-360.
Ward, T., Vess, J., Collie, R. M., & Gannon, T. A. (2006). Risk management or goods promotion: The relationship between approach and avoidance goals in treatment for sexual offenders. Aggression and Violent Behavior, 11, 378-393.
[Back to Top]