Practice with national-level exam (FACT, FACT Plus, NET, CUET, etc.) mocks, learn from structured notes, and get your doubts solved in one place.
The two dominant frameworks for forensic-psychological treatment of sex offenders: the Andrews-Bonta Risk-Need-Responsivity (RNR) model 1994 + 2010 (match treatment intensity to risk level, target criminogenic needs, deliver in a responsive manner); the Tony Ward Good Lives Model (GLM) 2002 + 2014 strengths-based framework; the UK Sex Offender Treatment Programme (SOTP) and its 2017 Ministry of Justice meta-analysis showing null-to-negative results that triggered SOTP withdrawal and replacement with the Horizon and Becoming New Me Plus programmes; the ATSA practice standards 2014 + the Marshall + Marshall outcome literature.
Last updated:
Whether a person who has committed a sexual offence can be treated, and by what means, sits at the intersection of clinical psychology, criminology, and human-rights law. Two theoretical frameworks have dominated that debate for three decades. The Risk-Need-Responsivity (RNR) model, developed by Donald Andrews and James Bonta at Carleton University in the late 1980s, tells practitioners to match the intensity of treatment to the offender's risk level, to target the psychological needs that causally drive reoffending, and to deliver intervention in ways the individual can actually engage with. The Good Lives Model (GLM), first articulated by Tony Ward and colleagues in New Zealand in 2002, accepts RNR's empirical base but reframes the therapeutic goal: rather than building a risk-avoidance skill set, help the person construct a meaningful, prosocial life in which offending becomes instrumentally unnecessary.
Both frameworks have now run through enough randomised and quasi-experimental outcome studies to be evaluated critically. The UK's Sex Offender Treatment Programme (SOTP), which ran in the prison estate for over two decades and cost hundreds of millions of pounds, was withdrawn in 2017 after a Ministry of Justice commissioned meta-analysis by Mews, Di Bella, and Purver found null-to-adverse outcomes for reconviction rates. That finding does not end the argument for sex-offender treatment, but it imposes a burden of methodological seriousness that any subsequent programme must meet.
This topic reviews the theoretical and empirical architecture of RNR and GLM, the SOTP outcome data and its aftermath, the current service landscape in England, the United States, Canada, Australia, and India, and the outstanding unresolved debates in the field.
*Three principles sound simple. Three decades of replication testing show they are necessary but not sufficient.*
Donald Andrews and James Bonta published the foundational RNR paper in 1990 in Criminology, after a decade of meta-analytic work on the Canadian correctional literature. The model rests on three core principles, each supported by a distinct empirical stream.
The Risk Principle holds that treatment intensity should be matched to the individual's reoffending risk level. High-risk offenders require intensive intervention, typically defined as 200 or more treatment hours in structured programming. Low-risk offenders benefit from minimal or no formal intervention, and providing intensive treatment to low-risk individuals can paradoxically increase their reconviction rate by disrupting the informal prosocial networks that already buffer their risk. The meta-analytic evidence for this principle, originally assembled by Andrews and colleagues and subsequently replicated by Hanson, Bourgon, Helmus, and Hodgson (2009) in a 23-study sample, is robust across North American and European samples.
The Need Principle distinguishes between criminogenic and non-criminogenic needs. Criminogenic needs are dynamic risk factors: attitudes supportive of sexual offending, deviant sexual interests, emotional identification with children, lack of victim empathy, intimacy deficits, and impaired self-regulation are the six most replicated criminogenic need domains in sex-offender treatment literature. Non-criminogenic needs include self-esteem, housing stress, and general anxiety: addressing these may improve quality of life but does not predict reoffending outcomes and should not dominate programme time. The ATSA (Association for the Treatment of Sexual Abusers) practice standards 2014 embed this distinction throughout their assessment and programme-design guidance.
The Responsivity Principle has two components. General responsivity holds that cognitive-behavioural methods work best across offender populations, a finding replicated from the general criminal-justice treatment literature. Specific responsivity requires adapting delivery to the individual's cognitive capacity, learning style, cultural context, and motivation to change. The Stages of Change model (Prochaska and DiClemente) is frequently integrated here to assess whether an offender is in pre-contemplation, contemplation, or preparation, because motivation-building work is needed before skills-based CBT can take hold.
In the sex-offender-specific literature, Karl Hanson and colleagues at Corrections Canada produced the most influential meta-analytic database. The 2002 review of 43 studies found a weighted mean recidivism reduction of 37 percent for treated versus untreated sex offenders. The 2009 review, incorporating 23 well-controlled studies, found that programmes addressing criminogenic needs reduced sexual recidivism from 19.2 percent (controls) to 10.9 percent (treatment), a clinically substantial effect. However, the quality of the control condition matters enormously: studies with no-treatment controls show larger effects than studies with wait-list or treatment-as-usual controls, suggesting some of the apparent treatment benefit reflects selection effects or regression to the mean.
The risk instruments used to apply the Risk Principle in sex-offender treatment allocation are discussed in detail in Module 4 (Static-99R, Stable-2007, Acute-2007). To avoid reproducing that material here, it is sufficient to note that RNR-consistent programmes formally integrate the Static-99R risk score into the treatment-dosage decision: offenders scoring in the low-average or below-average risk band may be directed to community-based maintenance rather than custody-based intensive programming.
*Every offence is a maladaptive attempt to meet a legitimate human need, says Ward. That reframe changes what treatment targets.*
Tony Ward, then at Victoria University of Wellington, introduced the Good Lives Model in a 2002 paper in Clinical Psychology Review co-authored with Claire Stewart. The GLM draws on positive psychology (Martin Seligman), eudaimonic wellbeing theory (Aristotle as filtered through contemporary values psychology), and the self-determination theory of Deci and Ryan. Its central empirical claim is not that RNR is wrong but that RNR's risk-avoidance framing is incomplete: an intervention that only teaches an offender what not to do leaves a motivational vacuum that the therapeutic process does not fill.
Primary goods. GLM identifies a set of basic human goods that people universally pursue: life (including physical health and functioning), knowledge, excellence in play and work, autonomy, inner peace, friendship and intimacy, community membership, spirituality (understood broadly as a sense of purpose and meaning), happiness, and creativity. Ward argues that sexual offending typically reflects a distorted or blocked attempt to acquire one of these goods. An offender who sexually abused a child was often attempting to meet goods of intimacy or knowledge through a harmful route. Effective treatment, on this account, must identify which goods the offender's life plan fails to provide through prosocial means, then build the skills and opportunities to pursue those goods adaptively.
Implications for programme design. GLM-influenced programmes place explicit emphasis on developing a personally meaningful, detailed, and realistic Good Lives Plan alongside the risk-management work. The plan specifies how the person intends to meet each primary good post-release without harming others. Therapeutic work includes developing occupational and relational skills, addressing identity and sense of purpose, and working with the environment (housing, employment supports, community) as well as the individual. The GLM has been particularly influential in New Zealand's Kia Marama and Te Piriti programmes, which integrated Maori cultural good-lives concepts alongside CBT offence-mapping work.
Empirical evaluation. The published outcome literature on GLM-specific programmes is thinner than the RNR base, partly because most programmes blend GLM motivational framing with RNR-targeted CBT content. Barnett, Mandeville-Norden, and Rakestrow (2014) evaluated the Thames Valley Sex Offender Groupwork Programme (TV-SOGP) in England, which incorporates GLM goal-setting, and found pre-post improvements on dynamic risk factors. Willis, Ward, and Levenson (2014) reviewed five GLM-informed programmes and concluded that the evidence of superior reconviction outcomes over standard RNR-CBT is not yet established. The critique from strict empiricists is that GLM adds therapeutic narrative and motivational benefit but has not demonstrated additive reconviction reduction over RNR-alone programmes. Ward's response is that reconviction is an inadequate sole outcome measure: wellbeing, stable prosocial reintegration, and reduced indirect harms to victims and families also matter.
*The Ministry of Justice's own data showed the programme did not reduce reoffending. That was not a minor finding.*
The UK Sex Offender Treatment Programme was the world's largest structured sex-offender treatment service. Running in the prison estate from 1991, it delivered a rolling programme of group-based CBT targeting the same criminogenic need domains identified by Andrews and Bonta. At its peak, SOTP enrolled thousands of prisoners annually across England and Wales. The Core Programme (90-100 sessions) and the Rolling Programme (for lower-functioning participants) were the main tracks, supplemented by the Better Lives Booster and the Adapted Sex Offender Treatment Programme for intellectually disabled offenders.
Mews, Di Bella, and Purver (2017), commissioned by the Ministry of Justice, conducted the outcome evaluation using a matched-sample design comparing 2,562 prisoners who completed SOTP between 2000 and 2012 against a matched control group of untreated sex offenders. The primary outcome was reconviction for a sexual offence within a follow-up period of up to 8 years. The headline finding was striking: SOTP completers were actually more likely to be reconvicted of a sexual offence than controls, though the absolute risk difference was small (10.0 percent versus 8.0 percent reconviction). The difference was not statistically significant at conventional thresholds, but the direction was adverse.
The Ministry of Justice withdrew SOTP in September 2017 and replaced it with two programmes developed by Hossack and colleagues: Horizon (for moderate-to-high-risk adult offenders) and Becoming New Me Plus (BNM+) for lower-risk and learning-disabled offenders. Both programmes draw on GLM framing alongside RNR need-targeting.
Methodological debates. The interpretation of the Mews et al. finding is genuinely contested. Critics note that the matched-control design cannot fully eliminate selection bias (those who enrolled in SOTP may have differed from controls in motivation, which itself predicts different outcomes). Seto and Barbaree had already observed in 2004 that showing high programme engagement sometimes predicts worse outcomes, possibly because psychopathic or pro-criminal participants learn to simulate therapeutic responsiveness while remaining unchanged. Other criminologists note that the reconviction outcome measure misses offences that were committed but not detected, and that SOTP-treated offenders may have committed offences that were qualitatively less severe even if slightly more numerous in official records.
The honest take is that the SOTP evidence does not demonstrate that treatment can never work, but it does demonstrate that SOTP as designed and delivered did not reduce sexual reconviction. The comparison benchmark matters: reconviction rates for sexual offenders in English prisons are already low relative to general-violence offenders, which constrains the statistical power to detect any treatment effect.
In Australia, the Queensland Corrections Sex Offender Programme (QCSOP) ran a quasi-experimental evaluation and found modest positive reconviction effects for high-risk participants, consistent with the RNR prediction that treatment benefit concentrates in the highest-risk tier. The Andrews-Bonta Corrections Learning and Adaptation model, used in the Correctional Service of Canada, applies a more tightly controlled RNR delivery system and has maintained generally positive meta-analytic outcomes.
*California built a hospital; New Zealand built a cultural programme; India built a referral chain. Each reflects a distinct theory of what treatment means.*
United States: Atascadero and SVP civil commitment. Atascadero State Hospital in California operates the largest forensic sex-offender treatment programme in the US, serving persons committed under California's Sexually Violent Predator Act 1996. The SVP scheme, mirrored in 20 US states, allows indefinite civil commitment of sex offenders assessed as likely to reoffend after completing their criminal sentence. The treatment protocol at Atascadero uses an RNR-structured, CBT-based relapse-prevention model. The ATSA 2014 practice standards for the treatment of adults who have sexually abused set a benchmark that state programmes are expected to meet. Controversially, the US civil-commitment system has attracted constitutional challenge under the substantive-due-process analysis of Kansas v. Hendricks (1997) and Kansas v. Crane (2002), with the US Supreme Court upholding the scheme subject to the requirement that the committed person's mental abnormality or personality disorder must make it difficult for them to control their behaviour.
Canada: Corrections Canada and the Hare-RNR integration. Canada's Correctional Service (CSC) operates one of the most systematically RNR-consistent sex-offender treatment systems in the world. The National Sex Offender Programme (NSOP) and the High Intensity Sex Offender Treatment Programme (HISOTP) are differentiated by Static-99R risk score. Canada produced some of the foundational meta-analytic evidence for sex-offender treatment through the work of Karl Hanson, Andrew Harris, and Scott Phenix at Public Safety Canada. The Dynamic Risk Assessment for Offender Re-Entry (DRAOR) bridges institutional and community supervision.
Australia: Queensland, New South Wales and ANZAPPL. Each Australian state runs its own sex-offender treatment programme within corrections. The Serious Sex Offenders (Detention and Supervision) Act 2009 in Victoria and the Dangerous Prisoners (Sexual Offenders) Act 2003 in Queensland allow post-sentence extended supervision orders. The Australian and New Zealand Association of Psychiatry, Psychology and Law (ANZAPPL) published guidelines for the treatment of sexual offenders that align with ATSA practice standards while acknowledging the limited local-population validation data for risk instruments.
India: POCSO referral chain and NIMHANS. India does not yet have a national structured sex-offender treatment programme equivalent to SOTP or Atascadero. The Protection of Children from Sexual Offences Act 2012 (POCSO) mandates sensitivity in the investigation and prosecution of child sexual abuse but does not create a treatment framework for convicted offenders. NIMHANS Bangalore operates a forensic psychology service that provides court-ordered psychological assessment and limited therapeutic intervention for offenders referred by sessions courts. The Institute of Human Behaviour and Allied Sciences (IHBAS) Delhi offers a forensic mental health programme that accepts sex-offender referrals. The Directorate of Forensic Science Services (DFSS) in several Indian states has begun developing referral protocols for psychosexual assessment following POCSO convictions. The gap between assessment and structured treatment remains wide: most convicted sex offenders in India receive no psychologically informed intervention beyond incarceration.
The Indian Supreme Court in State of Karnataka v. Madesha (2007) affirmed the relevance of psychological assessment in sentencing mitigation for sexually violent offences, which provides a judicial hook for treatment integration. A 2019 Law Commission of India report on sex-offender registration and management acknowledged the absence of any national treatment infrastructure and recommended pilot programmes modelled on UK and Canadian experience, pending evaluation.
*The relationship between therapist and client in sex-offender treatment predicts outcome at least as reliably as the treatment technology.*
William Marshall and Liam Marshall, at Queen's University and the Rockwood Psychological Services programme in Kingston, Ontario, contributed a distinctive strand to the sex-offender treatment literature that is often underweighted in discussions dominated by RNR and GLM theory. The Marshall model, developed through clinical programme work at Bath Institution and subsequently at Rockwood, emphasises the quality of the therapeutic alliance as a treatment mediator. The 2007 Psychology, Crime and Law paper by Marshall, Serran, Fernandez, Mulloy, Mann, and Thornton found that therapist empathy, warmth, directiveness, and reinforcement of disclosures accounted for more variance in treatment outcome measures than specific CBT technique delivery.
This finding is congruent with the general psychotherapy outcome literature, where the therapeutic alliance is a consistent predictor of outcome across modalities. In the specific sex-offender treatment context it matters because group CBT programmes are often delivered in a confrontational, challenge-oriented style that may actually increase defensiveness in participants, reducing engagement and potentially reproducing the adverse outcomes the Mews et al. evaluation detected in SOTP. The replacement UK programmes (Horizon and BNM+) incorporate explicit therapist-training components on alliance-building, partly in response to the Marshall critique.
The moral-reconation therapy (MRT) literature represents a different strand. MRT, developed by Greg Little and Kenneth Robinson in the US in the 1980s, uses a structured workbook series targeting moral-reasoning development. Its application in sex-offender programmes has been mixed: some institutional studies report reduced recidivism, but the methodological quality is generally weak, with no randomisation and limited follow-up periods. The ATSA practice standards treat MRT as a supplemental rather than a standalone intervention.
*Reconviction is a poor outcome measure. So is every available alternative.*
Sex-offender treatment research faces a structural problem: the most tractable outcome measure (reconviction for a new sexual offence) is also the most misleading. Sexual offences are among the most underreported crimes in every jurisdiction where victim survey data exist. In the UK Crime Survey for England and Wales, only around 1 in 6 adult rapes results in a police report. When a treated sex offender is reconvicted, researchers measure only the thin visible tip of an iceberg of potentially resumed offending. A treatment effect that reduces offence severity or victim harm without reducing official reconviction rates would be invisible to the standard evaluation design.
The field has moved toward intermediate outcome measures: changes in dynamic risk factors at treatment completion (measured using validated instruments such as Stable-2007, Thornton's Sexual Violence Risk-20, and the SONAR), reductions in self-reported sexual deviance measured by phallometry or penile plethysmography, and improvements on psychometric measures of victim empathy and emotional congruence with children. These intermediate measures have reasonable concurrent validity with later reconviction in prospective studies, but they are not substitutes for reconviction data and cannot be used to claim treatment success when reconviction reduction remains undemonstrated.
The responsivity debate also remains open. Sex offenders with high PCL-R psychopathy scores show the adverse treatment effect noted by Seto and Barbaree: intensive group CBT appears to increase rather than decrease their reconviction rates, possibly by providing social influence skills without changing underlying orientation. Whether this subgroup should be excluded from group treatment, diverted to individualised intervention, or simply managed through supervision rather than treatment is unresolved. The ATSA practice standards recommend explicit responsivity assessment and programme-matching but do not resolve the psychopathy-treatment question empirically.
Finally, the boundary between treatment and punishment in civil-commitment schemes raises ongoing human-rights challenges. The European Court of Human Rights in M. v. Germany (2009, Application No. 19359/04) found that indefinite preventive detention for sex offenders following completion of a criminal sentence violated Article 5 ECHR (right to liberty) when the detention was not genuinely therapeutic in character. The German Sicherungsverwahrung scheme was substantially reformed as a result. The UK's MAPPA (Multi-Agency Public Protection Arrangements) framework manages high-risk sex offenders in the community after release, with treatment delivered through the Probation Service's Horizon programme, avoiding the constitutional problems of the US SVP civil-commitment system.
A male prisoner convicted of rape scores in the high-risk band on Static-99R. Under strict RNR principles, the most appropriate treatment allocation is:
Test yourself on Forensic Psychology with free, timed mocks.
Practice Forensic Psychology questions