Parental sexual offending: Managing risk through diversion

Foreword | Public policy initiatives to manage parental child sexual offending have been hindered by the absence of risk instruments sensitive to unique factors associated with the distinctive profile of this core group of offenders. Using an Australian sample of 172 male parental offenders referred to community-based treatment designed for low-risk offenders, this study retrospectively compared risk levels and reoffence rates of offenders accepted into treatment (46%) with those who returned to court for standard criminal prosecution (54%). Overall, the results indicated that offenders with low risk scores were significantly more likely to be accepted into treatment, spend longer in treatment and complete treatment than offenders with high risk scores. Low-risk offenders who experienced standard criminal prosecution and sentencing reoffended 11 times faster than their counterparts who attended the community-based program (after controlling for pre-treatment dynamic risk levels). This diversion program was an effective preventive intervention that increased offender desistance and reduced threats to the safety and welfare young children and their families.

Adam Tomison
Director

Parental child sexual abuse has enduring and potentially devastating effects for victims and their families. These offenders are typically determined to be low risk on static risk assessment instruments and are thus often excluded from sexual offender treatment programs. The paucity of risk instruments sensitive to dynamic factors in this subtype of sexual offenders has impaired policy development to manage risk and address their treatment needs. Few established risk scales have been validated on Australian offenders. The current study reports on static and dynamic risk factors in an Australian sample of parental child sex offenders and the predictive strength of those factors regarding reoffending with or without a treatment intervention.

Limited success of standard criminal justice responses

Evidence of successful criminal justice responses or interventions with parental intrafamilial sex offenders is scant (Cossins 2010). Case attrition within the criminal justice process distinguishes parental offending, beginning with the reporting of the crime and persists through investigative, adjudicative and sentencing proceedings (Fitzgerald 2006).

Disincentives to victims to report a parent include reluctance to seek legal redress for a family problem, the detriment to relationships between the victim, the non-offending parent and siblings, and alienation from members of their extended family. In most cases, the offender is employed and is the main family breadwinner; thus, victims and non-offending parents often face financial hardship by pursuing legal action. Standard criminal prosecution encourages denial by parental offenders (Pratley & Goodman-Delahunty 2011). Testifying against a parent who denies the allegations is exceptionally challenging for young children, who can be re-traumatised by having to repeat their account of the events in court.

Few victims report parental sexual abuse to police (Priebe & Svedin 2008). Cases that proceed through standard criminal prosecution yield conviction rates substantially lower than those for other general criminal offences and for extrafamilial child sexual offences (BOCSAR 2011). Short sentences often preclude eligibility for custodial treatment. Additionally, many incarcerated intrafamilial child sex offenders are housed in protective custody to prevent assault from other inmates, limiting custodial treatment options. For the same reason, it is counterproductive to include these offenders in group therapy. Even if they are accepted into rehabilitation programs, group treatment aimed at extrafamilial sex offenders does not aim to address the unique criminogenic profile and needs of parental sex offenders (Goodman-Delahunty 2014). Moreover, exposure in custodial settings to higher risk sex offenders can exacerbate reoffending in lower risk participants (Wakeling, Mann & Carter 2012).

Publication of offenders’ names on a sex offender register (Child Protection [Offenders Registration] Act 2000) is largely symbolic (La Fond 2005), aimed at extrafamilial offenders whose victims are strangers. These registers have no positive impact on parental offenders, but may exacerbate their social stigmatisation and the loss of protective factors that reduce risks of sexual recidivism, such as social support, employment and housing (Tewksbury 2005). This practice can further inhibit disclosure of offending behaviours and discourage the reporting and prosecution of parental offending (La Fond 2005). Thus, traditional sentencing, deterrent and rehabilitation models may inadequately address the needs of parental offenders and therefore fail to reduce their risk of reoffending.

In sum, traditional criminal justice policies and procedures often obfuscate information about parental child sex offending rather than reduce its incidence. A dearth of reliable information about parental child sex offenders and their amenability to treatment has impeded the development of sound evidence-based public policies to decrease the number of sex crimes within the family. The foregoing deficits establish the need to assess viable alternatives to standard criminal prosecution to manage risks of parental sexual reoffending.

Treatment for intrafamilial sexual offenders

Despite research suggesting that intrafamilial sexual offenders–a group that includes parental sex offenders–are amenable to treatment (Gelb 2007), public attention and resources have focused almost exclusively on high-risk extrafamilial offenders such as adult rapists and extrafamilial child molesters. Comparatively little is known about how parental offenders respond to treatment and the type of program best suited to address their criminogenic needs. It is, therefore, unsurprising that studies of the effectiveness of custodial and community treatment for intrafamilial sex offenders have shown little or no measurable impact in reducing recidivism (Villetaz, Killias & Zoder 2006).

Figure 1 Program acceptance and treatment completion by VRS-SO risk level (%)

Treating offenders in accord with the risk principle (Andrews & Bonta 2010), requires that the dosage and intensity of interventions be matched to an offender’s risk of reoffending and that the most intensive services be reserved for higher risk offenders. Intrafamilial offenders often appear as low risk on measures of historical risk factors as they rarely have prior convictions, especially for sexual offences. Low-risk offenders are typically assumed to require minimal or no intervention. While ‘risk’ is a relative term, low risk of sexual reoffending ‘does not necessarily imply no risk’ (Wakeling, Mann & Carter 2012: 294).

To better assess risk of reoffending, psychologists distinguish static or historical risk features from those that are situational or dynamic. Static factors are associated with long-term risk and include features such as age, marital status, offence type, age at first conviction and number of previous convictions (Hanson & Thornton 2000). Dynamic factors are features that change, such as stress, anger, attitudes towards offending and substance abuse (Beggs & Grace 2010). Research indicates that some sexual offenders who score as low risk on static risk assessment instruments may nonetheless have high dynamic needs (Beech et al. 2002). Thus, they may require more treatment than is offered to ‘typical’ low-risk sexual offenders (Wakeling, Mann & Carter 2012)

Few empirical studies have demonstrated how to translate the risk principle into practice to determine how much treatment low compared with high-risk offenders require. The limited evidence that is available suggests that between 80 (Beech & Mann 2002) and 120 hours (Marshall et al. 2006) of treatment is sufficient for most low-risk offenders, unless they present with significant or entrenched dynamic needs (Beech, Fisher & Beckett 1998; Bourgon & Armstrong 2005). By contrast, high-risk offenders are presumed to require at least 300 hours of intervention to adequately reduce their risk of recidivism (Hanson & Yates 2013).

Diversion programs for parental child sex offenders

To break ‘costly and ineffective cycles of arrest, incarceration, release, and re-arrest that has often characterised the criminal justice system’s response’ to offenders perceived as low risk, renewed interest has emerged in community-based programs that divert these offenders from standard prosecution (Heilbrun & DeMatteo 2012: 349). However, few studies have investigated the effectiveness of diversion programs in reducing recidivism among low-risk sexual offenders and if so, for whom this form of treatment is most effective; that is, matching intensity of treatment with risk.

The NSW pre-trial diversion program

In Australia, a small number of diversion programs have been available for parental sex offenders, but most have not been formally evaluated. One exception is the New South Wales Pre-trial Diversion Program for Child Sex Offenders, known as Cedar Cottage, established in 1989. Administered by the NSW Department of Health, Cedar Cottage uniquely focused on sparing child victims the burdens of the traditional criminal justice processes, provided therapeutic services to all affected family members (particularly the child victims, their siblings and the non-offending parent) and in addition, offered community-based treatment to parental sex offenders. Treatment took a holistic approach, supporting fathers in the program to make positive changes in all aspects of their lives (Pratley & Goodman-Delahunty 2011).

To be eligible for diversion, applicants in a parental role had to indicate their willingness to plead guilty to charges of sexually abusing a child in their care, have no record of violence during the sexual abuse or prior convictions for sexual assault. Offenders meeting these statutory eligibility criteria were referred by police, prosecutors and the local courts to the Program Director for a clinical assessment to determine their suitability for the program. During the clinical assessment process, lasting eight weeks, legal proceedings were adjourned. To enter the program, offenders had to demonstrate acceptance of responsibility for their offending behaviour, awareness of the impact of the crime on their victim and family, and sufficient communication skills to participate in treatment. Offenders who were declined entry to the program because they did not meet the legal or clinical eligibility criteria, returned to court to resume traditional prosecution. No formal risk assessment tools were used in screening eligible referrals, in part because the program and criteria were established in 1985–88, before risk assessment tools were developed.

Offenders accepted into the program pled guilty to their offences and were diverted from the criminal justice system to attend two to three years of treatment in the community. The Cedar Cottage treatment program applied cognitive-behavioural therapy, narrative therapy and invitational practice to address criminogenic needs and support offenders to change the behaviours that led to their offending. Treatment involved attendance at fortnightly individual and group therapy sessions (1.25 and 2.5 hours respectively), on alternating weeks. A minimum of two years’ participation, totalling 180 hours of treatment, was required before an offender was determined to have successfully completed treatment. At the discretion of the Director, offenders could be required to continue treatment for an additional year. Completion of the program precluded sentencing for the index offence.

A previous evaluation established that the program effectively reduced sexual recidivism rates by 52 percent (Butler, Goodman-Delahunty & Lulham 2012) and that biological and non-biological fathers benefited equally from the intervention (Titcomb, Goodman-Delahunty & Waubert de Puiseau 2012). No previous study has investigated whether acceptance into the Cedar Cottage program was equally effective for lower compared with higher risk offenders (ie how appropriate matching of offenders based on risk level might influence outcomes). In part, this is because the most commonly used risk assessment instrument for predicting sexual recidivism, the Static-99 (Hanson & Morton-Bourgon 2009; Smallbone & Ransley 2005), yielded floor effects and no useful diagnostic information (Butler, Goodman-Delahunty & Lulham 2012). The need to better distinguish risk among parental offenders by exploring differences in dynamic risk levels was identified as a priority for future research (Titcomb, Goodman-Delahunty & Waubert de Puiseau 2012).

Aims of the research

The current study examined two related aims:

• whether the level of dynamic risk posed by offenders diverted from traditional criminal prosecution to the Cedar Cottage community-based treatment program was similar to that of declined offenders who received standard criminal prosecution; and
• the influence of community-based treatment on recidivism rates by level of assessed risk.

Method

Participants

Participants were 213 male parental intrafamilial offenders referred to Cedar Cottage from 1989–2003 (Goodman-Delahunty 2009). Of this group, records for 41 men were excluded, as the information in their clinical files was insufficient to code dynamic risk, leaving a study sample of 172 eligible offenders. Of this group, half (54.1%; n=93) were accepted into the Cedar Cottage program and thus diverted from the criminal justice system, while 45.9 percent (n=79) were declined. The declined group consisted of offenders who did not meet the clinical screening criteria noted above and accepted offenders who chose traditional criminal justice processes instead. At the time of assessment, participants ranged in age from 24–58 years (M=39.8; SD=7.1). Most participants were legally married (64%, n=111) or in a de facto relationship (20%, n=34). Participants were more likely to be non-biological (55%), than biological fathers (45%) of the victim; however, this difference was not statistically significant (for further information about characteristics of this sample of parental sex offenders referred to the program, see Goodman-Delahunty 2014).

Procedures and data sources

The retrospective study entailed reviewing and manually coding information in clinical assessment and treatment files maintained by the NSW Pre-Trial Diversion of Offenders Program (data was collected between 2007 and 2012). A team of 11 postgraduate research assistants manually audited offender files to code details about acceptance into the program, participation in treatment (treatment duration, completion status), personal and criminal history, and the index sexual offence. Inter-rater reliability scores were good (mean Fleiss free-marginal Kappa 0.76, mean Intraclass Correlation Coefficient 0.84). These data yielded static scores on the Violence Risk Scale: Sexual Offender Version (VRS-SO; Wong et al. 2003). Subsequently and independently, two postgraduate research assistants performed a second manual audit of offender assessment and treatment files to code information required for the dynamic scales of the VRS-SO, yielding very good inter-rater reliability, with an average ICC for pre-treatment dynamic scores of 0.95.

Information gathered throughout the eight week assessment process (referral information, clinical notes and records from assessment interviews) was the source to code pre-treatment VRS-SO dynamic items for all 172 eligible offenders. As might be expected, less information was available about certain declined offenders; for example, if they withdrew early from the assessment process. Although there were no missing data for static VRS-SO items, in a number of cases, data were missing for some VRS-SO dynamic items. Accordingly, missing values were estimated by means of a stepwise regression procedure (Goodman-Delahunty & O’Brien 2014). Only pre-treatment VRS-SO scores were utilised in the current study.

The violence risk scale: Sexual offender version

The VRS-SO (Wong et al. 2003) is a 24 item clinical rating scale, designed to assess static and dynamic risk factors for sexual recidivism. Seven static risk items provide information about historical risk factors, whereas 17 dynamic risk items provide information about criminogenic needs that can be addressed in treatment. Total VRS-SO scores range from 0–72. All items on the VRS-SO are ‘empirically or conceptually linked’ to sexual offending (Olver et al. 2007: 318). Research on the validity and reliability of the VRS-SO has demonstrated that it is an effective tool for assessing reoffence risk and treatment change in both adult (Canales, Olver & Wong 2009; Olver et al. 2007) and child sexual offender populations (Beggs & Grace 2011, 2010; Olver et al. 2013).

Recidivism data

Recidivism rates based on official criminal records tend to underestimate the true rate of reoffending (Greenberg et al. 2000). This is especially true for sexual crimes, as relatively few offences come to the attention of the criminal justice system. To counteract this problem, some researchers have relied on offenders’ self-reports of their reoffending. In the current study, no access to the offenders was feasible. Accordingly, a broader range of official reports of reoffending was used, in addition to reconviction data, to better estimate reoffence rates. Recidivism was defined as any new police report/arrest, charge and conviction for offences committed after the date of last contact with Cedar Cottage. Three categories of recidivism were distinguished—sexual, violent (excluding sexual offences) and overall recidivism (a combination of all sexual, violent and nonsexual/nonviolent offences).

Recidivism data were derived from official criminal records maintained in the NSW Police Computerized Operational Policing System and the NSW Bureau of Crime Statistics and Research Reoffending Database. Data regarding length of incarceration for the index offence (for declined offenders and program non-completers who returned to court for traditional prosecution) was collected from the Department of Corrective Services, to ensure that only periods when offenders were free in the community during the follow-up period were taken into account. On average, offenders in this study were followed for 9.1 years (follow-up period range=4–18 years).

Statistical analyses

A series of independent sample t-tests examined differences in VRS-SO static and dynamic risk scores for offenders accepted into treatment (ie based on intent to treat) and offenders declined entry to the program. Next, offenders were divided into two groups based on their VRS-SO scores (low vs high risk), to examine whether risk level interacted with participation in treatment to influence recidivism outcomes. A series of chi-square analyses and a Cox regression survival analysis (to control for differences in follow-up and time at risk) addressed the second aim. In addition, relative reductions in sexual reoffending in the two groups were explored to examine treatment effects.

Results

Overall, a substantial proportion of the parental offender sample had prior criminal records (47.1%, n=81), but few had prior convictions for sexual offences (5.3%, n=9) or violent offences (12.8%, n=22). Consistent with these data, the mean VRS-SO static scale score for the sample was relatively low (M=2.7, SD=2.6, range 0–13). However, the mean VRS-SO dynamic scale scores for this sample of parental sex offenders was higher than anticipated compared with other intrafamilial offender samples (Beggs & Grace 2010, M=25.9; Olver et al. 2007, M=26.1). The overall mean pre-treatment VRS-SO dynamic total score was 36.3 (SD=8.0, range 18.0–48.8) and therefore, the mean total pre-treatment VRS-SO score was 39.0 (SD=9.0, range 20–59).

Level of risk for accepted versus declined offenders

Offenders who were accepted into the program had significantly lower VRS-SO dynamic scale scores (M=31.3, SD=6.9) than offenders declined entry to the program (M=42.2, SD=4.4; t(170)=12.48, p<.001). However, the two groups of offenders were undifferentiated based on their pre-treatment VRS-SO static risk scale scores—accepted offenders M=2.7 (SD=2.5), declined offenders (M=2.8, SD=2.8; t(170)=0.41, p=.685).

To further explore differences between high and low-risk offenders, two groups were created by splitting the sample at the VRS-SO total scale score median of 39.0—low-risk offenders (n=81, VRS-SO scores 0–38) and high-risk offenders (n=91, VRS-SO scores 39–72). Overall, 80.2 percent (n=65) of low-risk offenders were accepted into the treatment program and diverted from prosecution, whereas only 30.8 percent (n=28) of high-risk offenders were accepted into treatment (see Figure 1); this difference was significant (χ2 (1)=42.2, p>.001, φ=0.50).

Figure 2 Sexual, violent and overall recidivism rates by VRS-SO risk level and program acceptance (%)

Note: Accepted=offenders accepted into the Cedar Cottage program; Declined=offenders declined entry to the Cedar Cottage program; Low risk=VRS-SO scores 0–38; High risk=VRS-SO scores 39–72

Program completion and risk

Overall, more than half of the parental offenders (57%, n=53) who were accepted into the program and diverted from standard prosecution successfully completed treatment; 43 percent of the sample (n=40) were non-completers. Offenders dropped out of the program prematurely for one of two reasons—involuntary withdrawal due to a breach of terms of the treatment agreement (80.0%, n=32; for instance, if an offender did not attend all treatment sessions or had contact with children under the age of 16 years) and voluntary withdrawal from treatment (20.0%, n=8). On average, offenders remained in the program for 24 months. As expected, offenders who completed treatment spent significantly longer in the program (M=31.5 months, SD=4.9), compared with non-completers (M=14.9 months, SD=8.5; t(91)=-11.78, p<.001, d=2.5).

Pre-treatment VRS-SO risk categories significantly predicted treatment completion (Χ² (1)=40.6, p<.001, φ=0.66). Low-risk offenders were significantly more likely to complete treatment than were high-risk offenders (see Figure 1). Consistent with these findings, VRS-SO total scores were significantly negatively correlated with time in treatment (r=-.53, p<.001, n=93); that is, lower risk offenders were more likely to stay in the program longer.

Recidivism outcomes

On average, offenders were followed up for 9.1 years after their last contact with Cedar Cottage. In this period, 32.0 percent reoffended overall (ie received a police report, charge or conviction for a new offence). Approximately equivalent proportions of the sample reoffended sexually (11.6%) or committed a new violent offence (9.9%). Offenders who were accepted into the program were less likely to reoffend sexually (7.5%, n=7) than offenders declined entry to the program (16.5%, n=13); this comparison approached significance: Χ²(1)=3.31, p=.069, φ=.14. Interestingly, the sexual recidivism rate was equivalent among program completers (7.5%, n=4) and non-completers (7.5%, n=3), although none of the reoffences by program completers were against children.

The sexual, violent and overall recidivism rates for this sample of parental sex offenders are displayed in Figure 2 by VRS-SO risk category and program acceptance group. Across the sample, significantly more high-risk offenders (44%, n=40) reoffended overall, compared with low-risk offenders (18.5%, n=15; Χ²(1)=12.75, p<.001, φ=.27). Similarly, more high-risk offenders reoffended sexually and violently (14.3% and 13.2% respectively) compared with low-risk offenders (8.6% and 6.2% respectively); however, these differences were non-significant. There was no significant difference in the violent and overall recidivism rates of low versus high-risk offenders, by program acceptance.

However, from Figure 2, it appeared that acceptance into treatment may interact with risk level to influence sexual recidivism outcomes. That is, low-risk offenders diverted to the community-based program (6.2%) were less likely to reoffend sexually than low-risk offenders who were declined treatment (18.8%, although this effect was not statistically significant (Χ²(1)=2.58, p=.108, φ=.18). Sexual recidivism rates observed in high-risk offenders did not differ based on whether they were accepted into the program or declined (10.7% vs 15.9%, Χ²(1)=0.42, p=.52, φ=.07). Overall, however, the relative reduction in sexual offending observed in accepted versus declined offenders was greater among low-risk than high-risk offenders (67% vs 33% reduction in reoffending).

To explore whether program participation interacted with risk of sexual recidivism, four groups were created as a function of acceptance (intention to treat) and VRS-SO risk scores—low-risk declined (n=16), low-risk accepted (n=65), high-risk declined (n=63) and high-risk accepted (n=28). A Cox regression survival analysis examined sexual recidivism outcomes, controlling for time at risk. Furthermore, VRS-SO dynamic risk scale scores were entered into the first step of the model to control for pre-existing differences between declined and accepted offenders. The results presented in Figure 3 revealed that low-risk accepted offenders had the lowest sexual recidivism failure rate, followed by the low-risk declined and high-risk accepted groups, which were very similar. Not surprisingly, high-risk declined offenders had the highest sexual recidivism rates (after controlling for time at risk and pre-existing differences in dynamic risk levels).

On average, low-risk offenders who were declined entry to treatment reoffended sexually faster than low-risk offenders who were accepted into the program (β=2.40, SE=1.07, df=1, p=.025, Exp(β)=10.98). Furthermore, high-risk accepted offenders and high-risk declined offenders also had significantly higher sexual recidivism rates than those of the low-risk accepted offenders (β=2.47, SE=1.27, df=1, p=.052, Exp(β)=11.87 and β=3.34, SE=1.43, df=1, p=.019, Exp(β)=28.24 respectively).

Discussion

Few community-based diversionary programs for sexual offenders exist and fewer have been empirically evaluated. By comparing reoffence rates in a sample of parental offenders who experienced standard court processes and incarceration if convicted with those treated in the community, the benefits of diversion and appropriate matching of offenders to treatment intensity were revealed. Most importantly, the findings indicate that compared with standard criminal prosecution, treatment at the Cedar Cottage program reduced sexual recidivism rates in low-risk offenders by 67 percent, whereas low-risk parental offenders who underwent standard criminal prosecution reoffended faster and at a higher rate. Notably, no offender in the study sample who completed treatment after 1993, when the Cedar Cottage program was refined, was convicted for a sexual offence against a child in the follow-up period that averaged nine years. Declined offenders may have been treated in available prison programs for sex offenders, such as CUBIT or CORE (NSW Corrective Services nd). Although not a statistically significant reduction, acceptance into the treatment program reduced sexual reoffending among high-risk offenders by 33 percent.

Figure 3 Sexual recidivism rates for low versus high-risk offenders, by acceptance into the Cedar Cottage program (controlling for VRS-SO dynamic risk scale scores)

Previous research of intrafamilial offenders provided varying estimates of the anticipated recidivism rate among these offenders. After following offenders for 15 years, Harris and Hanson (2004) reported that 13 percent of incest offenders reoffended sexually (new charges and convictions). Similarly, after following offenders for 19 years (average 10.8 years) Kingston et al. (2008) found that 9.8 percent of incest offenders received a new charge or conviction for a sexual offence. The sexual recidivism rate for low-risk accepted offenders in the current study (6.4%) fell below rates observed in previous studies using similar definitions of recidivism and comparable follow-up periods.

Overall, the outcomes in this study tend to indicate that community-based programs such as Cedar Cottage can be effective in reducing reoffending by parental sexual offenders and highlight the importance of matching risk with treatment intensity. The smaller reduction in reoffending rates observed among higher risk parental offenders suggests that offenders in that group require more intensive intervention, matched to their higher criminogenic needs (Beech, Mandeville-Norden & Goodwill 2012). Overall, the foregoing results indicate that reoffending by low-risk parental sexual offenders is more successfully prevented and managed in the community than by standard criminal prosecution. The treatment dosage of approximately 200 hours provided at Cedar Cottage was appropriate for low-risk offenders with higher dynamic criminogenic needs.

Although risks posed by parental child sex offenders are low in probability, if unaddressed, the magnitude of the harm perpetrated is extreme. Risk instruments reliant on historical or static factors are commonly used in correctional settings but are insensitive to risks of parental sex offending, misclassifying the risk by underestimation (Butler, Goodman-Delahunty & Lulham 2012; Smallbone & Ransley 2005). Given that few validated methods for recidivism prediction in community-based samples have been identified (Swinburne et al. 2012), the results achieved in this study using the VRS-SO (a comprehensive risk assessment instrument that incorporates dynamic factors) are encouraging. Dynamic risk factors are not only useful predictors of risk of reoffending but can assist clinicians in identifying treatment needs, likely responsivity to intervention and can capture changes in dynamic risk during and following treatment. The findings in this study indicate that the VRS-SO could be used to enhance the screening, selection and management of offenders eligible for diversion, by identifying their level of risk, as well as the likelihood that a particular offender will comply with and complete treatment.

Limitations

The present study had a number of limitations. Importantly, only sex offences that came to the attention of NSW Police were counted when estimating sexual recidivism. As such, they are likely to underestimate of the true rate of sexual reoffending. The study was unable to assess and control for all factors related to recidivism and therefore, it is possible, especially considering the lengthy follow-up period, that factors other than treatment contributed to the observed recidivism rates (eg post-release factors). A separate propensity analysis conducted to control for selection biases in the accepted versus declined groups confirmed that treatment (based on intent to treat) reduced the recidivism rate by 52 percent (Butler, Goodman-Delahunty & Lulham 2012). In the current study, pre-existing differences in pre-treatment dynamic risk levels were taken into account when investigating the impact of program acceptance on recidivism rates, but other differences between the samples not captured by the VRS-SO may have contributed to the positive treatment results observed (eg accepted offenders may have been more motivated to change).

Policy implications and recommendations

Policy changes implemented by the NSW Attorney-General late in 2012 resulted in the removal of diversion from prison as a sentencing option for offenders charged with child sexual assault offences. Since then, no new referrals to the Cedar Cottage program were permitted, which is unfortunate given the positive results of treatment indicated by this study. The foregoing research outcomes will be of interest to the pending NSW Joint Select Inquiry into the Sentencing of Child Sexual Assault Offenders and other states and legislatures seeking effective methods to prosecute parental sex offenders and increase protection to children and families.

Accurate identification of high versus low-risk parental sex offenders and effective treatment and management can reduce costs of incarceration, re-incarceration and sexual victimisation. Recent studies of community sanctions demonstrated that offenders recidivated significantly less after performing community service compared with imprisonment and that costs of community-based treatment were less than those of custodial sentencing (Wermink et al. 2010). The impact on offenders’ lives was reduced, thereby enhancing protective factors and supporting desistance from offending by maintaining connections within the community (Ward & Laws 2010). Community containment is also consistent with shifts towards therapeutic jurisprudence (King et al. 2009), as offenders are guided in developing stronger support systems and in recognising and avoiding placing themselves at risk of reoffending.

Policymakers are advised to ensure that the justice system balances confinement and containment in the community by ensuring the response is matched to the known risk, applying more control over offenders likely to commit another crime and less when not (La Fond 2005). The evidence shows that as a risk-management strategy, supervision of low-risk parental sex offenders by multidisciplinary teams in a noncustodial setting may be more effective than incarceration and registration.

By improving the ability to predict completion of a community-based program and identification of offenders who may benefit most from treatment, this study contributes to a more informed allocation of resources in the sentencing and rehabilitation of child sex offenders. Application of these findings can improve the criminal justice response to parental offending, thereby reducing threats to the safety and welfare of young children and their families.

References

All URLs correct at June 2014

• Andrews DA & Bonta J 2010. The psychology of criminal conduct, 5th ed. New Providence, NJ: LexisNexis
• Beech A, Fisher D & Beckett R 1998. STEP 3: An evaluation of the prison sex offender treatment programme. London: HMSO
• Beech AR, Friendship C, Erikson M & Hanson RK 2002. The relationship between static and dynamic risk factors and reconviction in a sample of U.K. child abusers. Sexual Abuse: A Journal of Research and Treatment 14(2): 155–167
• Beech A & Mann R 2002. Recent developments in the assessment and treatment of sexual offenders, in McGuire J (ed.), Offender rehabilitation and treatment: Effective programmes and policies to reduce re-offending. Chichester: Wiley-Blackwell: 259–288
• Beech AR, Mandeville-Norden R & Goodwill A 2012. Comparing recidivism rates of treatment responders/nonresponders in a sample of 413 child molesters who had completed community-based sex offender treatment in the United Kingdom. International Journal of Offender Therapy and Comparative Criminology 56(1): 29–49
• Beggs S & Grace R 2011. Treatment gain for sexual offenders against children predicts reduced recidivism: A comparative validity study. Journal of Consulting and Clinical Psychology
  79 (2): 182–193
• Beggs SM & Grace RC 2010. Assessment of dynamic risk factors: An independent validation study of the Violence Risk Scale: Sexual Offender Version. Sexual Abuse: A Journal of Research and Treatment 22(2): 234–251
• Bourgon G & Armstrong B 2005. Transferring the principles of effective treatment into a ‘real world’ prison setting. Criminal Justice and Behavior 32: 3–25
• Butler L, Goodman-Delahunty J & Lulham R 2012. Effectivness of pre-trial community-based diversion in reducing reoffending by adult intrafamilial child sex offenders. Criminal Justice and Behavior 39: 493–513
• Canales DD, Olver ME & Wong SCP 2009. Construct validity of the violence risk scale—sexual offender version for measuring sexual deviance. Sexual Abuse: A Journal of Research and Treatment 21(4): 474–492
• Cossins A 2010. Alternative models for prosecuting child sex offences in Australia. National Child Sexual Assault Reform Committee. Sydney: University of New South Wales
• Fitzgerald J 2006. The attrition of sexual offences from the New South Wales criminal justice system. Contemporary Issues in Crime and Justice no. 92. Sydney: NSW Bureau of Crime Statistics and Research
• Gelb K 2007. Recidivism of sex offenders. Research report for the Sentencing Advisory Council of Australia. Melbourne: Sentencing Advisory Council of Australia
• Goodman-Delahunty J 2014. Profiling parental child sex abuse. Trends & Issues in Crime and Criminal Justice no. 465. Canberra: Australian Institute of Criminology. http://aic.gov.au/publications/current%20series/tandi/461-480/tandi465.html
• Goodman-Delahunty J 2009. The NSW pre-trial diversion of offenders (child sexual assault) program: An evaluation of treatment outcomes. Sydney: Sydney West Area Health Services
• Goodman-Delahunty J & O’Brien K 2014. Reoffence risk in intrafamilial child sexual offenders. Report to the Criminology Research Advisory Council. Grant: CRG 44/10–11. Canberra: Australian Institute of Criminology
• Greenberg D, Bradford J, Firestone P & Curry S 2000. Recidivism of child molesters: A study of victim relationship with the perpetrator. Child Abuse & Neglect 24(11): 1485–1494
• Hanson RK & Morton-Bourgon KE 2009. The accuracy of recidivism risk assessments for sexual offenders: A meta-analysis of 118 prediction studies. Psychological Assessment 21(1): 1–21
• Hanson RK & Thornton D 2000. Improving risk assessments for sex offenders: A comparison of three actuarial scales. Law and Human Behavior 24(1): 119–136
• Hanson RK & Yates PM 2013. Psychological treatment of sex offenders. Current Psychiatry Reports 15(3): 1–8
• Harris AJR & Hanson RK 2004. Sex offender recidivism: A simple question. (User Report 2004–03). Ottawa: Public Safety and Emergency Preparedness Canada
• Heilbrun K & DeMatteo D 2012. Introduction to the special issue on diversion from standard prosecution. Criminal Justice & Behavior 39: 349
• King M, Freiberg A, Bagatol B & Hyams R, 2009. Non-adversarial justice. Sydney: Federation Press
• Kingston DA, Firestone P, Wexler A & Bradford JM 2008. Factors associated with recidivism among intrafamilial child molesters. Journal of Sexual Aggression 14(1): 3–18
• La Fond JQ 2005. Preventing sexual violence: How society should cope with sex offenders. Washington, DC: American Psychological Association
• Marshall WL, Marshall LE, Serran G & Fernandez YM 2006. Treating sexual offenders: an integrated approach. New York, NY: Routledge
• NSW Bureau of Crime Statistics and Research (BOCSAR) 2011. Sentencing snapshot: Child sexual assault, 2009–2010. Sydney: BOCSAR. Revised January 2012
• NSW Corrective Services nd. Sex and violent offender therapeutic programs. http://www.correctiveservices.nsw.gov.au/programs/sex-and-violent-offender-therapeutic-programs
• Olver ME, Wong SCP, Nicholaichuk T & Gordon A 2007. The validity and reliability of the Violence Risk Scale-Sexual Offender version: Assessing sex offender risk and evaluating therapeutic change. Psychological Assessment 9(13): 318–329
• Olver ME, Nicholaichuk TP, Gu D & Wong SCP 2013. Sex offender treatment outcome, actuarial risk, and the aging sex offender in Canadian corrections: A long-term follow-up. Sexual Abuse: A Journal of Research and Treatment 25: 396–422
• Pratley J & Goodman-Delahunty J 2011. Increased self-disclosure by intrafamilial child sex offenders. Sexual Abuse in Australia and New Zealand 3(1): 44–52
• Priebe G & Svedin CG 2008. Child sexual abuse is largely hidden from the adult society. An epidemiological study of adolescents’ disclosures. Child Abuse & Neglect 32: 1095–1108
• Smallbone S & Ransley J 2005. Legal and psychological controversies in the preventative incapacitation of sexual offenders. The University of New South Wales Law Journal 28(1): 299–305
• Swinburne Romine R, Miner M, Poulin D, Dwyer S & Berg D 2012. Predicting reoffense for community based sexual offenders: An analysis of 30 years of data. Sex Abuse: A Journal of Research and Treatment 24(5): 501–514. doi: 10.1177/1079063212446514
• Tewksbury R 2005. Collateral consequences of sex offender registration. Journal of Contemporary Criminal Justice 21(1): 67–81
• Titcomb C, Goodman-Delahunty J & Waubert de Puiseau B 2012. Pretrial diversion for intrafamilial child sexual offending: does biological paternity matter? Criminal Justice and Behavior 39: 552–570
• Villetaz P, Killias M & Zoder I 2006. The effects of custodial vs non-custodial sentences on re-offending. A systematic review of the state of knowledge. http://www.campbellcollaboration.org/lib/download/108/
• Wakeling HC, Mann RE & Carter AJ 2012. Do low-risk sexual offenders need treatment? The Howard Journal of Criminal Justice 51(3): 286–299.
• Ward T & Laws DR 2010. Desistance from sex offending: Motivating change, enriching practice. International Journal of Forensic Mental Health 9(1): 11–23
• Wermink H, Blokland A, Nieuwbeerta P, Nagin D & Tollenaar N 2010. Comparing the effects of community service and short-term imprisonment on recidivism: a matched samples approach. Journal of Experimental Psychology 6: 325–349
• Wong S, Olver ME, Nicholaichuk T & Gordon A 2003. Violence risk scale: Sexual offender version VRS-SO. Saskatoon, Saskatchewan, Canada: Regional Psychiatric Centre and University of Saskatchewan

About the authors

Jane Goodman-Delahunty is a Research Professor at Charles Sturt University School of Psychology and the Australian Graduate School of Policing and Security. Kate O’Brien is Doctoral Candidate in the School of Psychology and Psychiatry at Monash University.

Acknowledgements

The authors are grateful to Stephen P Wong and Mark E Olver for their guidance, to Anne Lucas and Rhianna Shi for their research assistance, and to Dale Tolliday and Karen Parsons, former Clinical Directors of Cedar Cottage who facilitated data collection. This research was funded by a grant from the Criminology Research Council (Grant: CRC 44/10–11) and a Faculty of Arts grant from Charles Sturt University.