Treatment of Sexual Offenders: Research, Best Practices, and Emerging Models

Treatment of sexual offenders has evolved substantially over the years; various theoretical and practice models of treatment been developed, modified, refined, and proposed over time. The predominant current recommended approach, supported by research, adheres to specific principles of effective correctional intervention, follows a cognitive-behavioral, skills-based orientation, and explicitly targets risk factors empirically associated with sexual offending and with recidivism, such that risk of re-offending may be reduced. Cognitive-behavioral treatment focuses on changing behavior, cognition, and affect, using a skills-based approach, with the aim of reducing risk of recidivism, and includes specific characteristics and methods on the part of treatment providers. Treatment is also guided by specific principles of intervention to maximize effectiveness. New models of treatment have been proposed with the aim of replacing and/or augmenting existing models. This article discusses existing and emerging models, their research basis, and recommendations for best practice in the treatment of sexual offenders.

Sexual offending has long been recognized as a serious problem with significant impacts on victims, their families, and society at large. Coinciding with this recognition has been the development and implementation of treatment interventions designed to reduce the risk of recidivism, empirical research into treatment effectiveness, and an increase in the availability of treatment programs for sexual offenders (McGrath et al., 2010). Current best practice involves the application of cognitive-behavioral interventions that target risk and that adhere to the principles of effective correctional intervention (Andrews & Bonta, 2010;Hanson, Bourgon, Helmus, & Hodgson, 2009). In addition, meta-analytic research has found that cognitive-behavioral treatment is most effective in reducing recidivism in comparison to both other types of treatment and to criminal sanctions (Hanson et al., 2002;Lösel & Schmucker, 2005). Lastly, research indicates that effective therapists and therapeutic techniques are associated with improved outcomes (Beech & Fordham, 1997;Marshall, Anderson, & Fernandez, 1999;Marshall et al., 2002Marshall et al., , 2003Yates et al., 2000). In this article, a review of sexual offender treatment is provided, with accompanying research support for specific approaches. This is followed by a review of emerging treatment models and their potential to inform the practice of sexual offender treatment. Various models of sex offender treatment have been proposed and implemented over time, including general psychotherapy, neurosurgery, physical castration, pharmacological interventions, behavioral reconditioning, cognitive-behavioral intervention, and relapse prevention (for a review, see Laws 2003;Yates, 2002;. Early treatment approaches assumed that sexual offending was caused by a single factor, such as anger or deviant sexual arousal (Becker & Murphy, 1998;Marshall, 1996). However, over time, the multidimension-al nature of sexual offending became evident, and treatment approaches incorporated multiple elements in order to address these multiple influences on behavior and sexual offending risk (Marshall et al., 1999;Marshall, Marshall, Serran, & Fernandez, 2006;Yates, et al., 2000). Many early interventions, such as psychotherapy and neurosurgery, have been found to be ineffective (see Yates, 2002Yates, , 2003, for a review), while others, including pharmacological interventions, are potentially promising in some specific cases (Bradford, 1990;Grossman, Martis, & Fichtner, 1999;Meyer, Cole, & Emory, 1992). Still others, including relapse prevention, continue to be used in spite of an absence of research support for their effectiveness. Treatment models that have been shown to be effective, and emerging promising models, are the focus of this review.

 Principles of Effective Correctional Intervention
In general, in correctional intervention with offenders, specific principles have been found to be essential in interventions designed to reduce recidivism, and specifically, the principles of risk, need, and responsivity (RNR model; Andrews & Bonta, 2010). While originally intended to be applied predominantly to criminal justice sanctions (i.e., sentencing, diversion, and supervision), in practice this model has additionally been applied to treatment, and perhaps more so to treatment than to sanctions.
Risk Principle. According to the risk principle, the intensity of correctional interventions must be matched to the level of risk posed by the offender. Treatment, as well as supervision, should be longer in duration, applied more frequently, and include more contact hours as assessed risk to reoffend increases (Andrews & Bonta, 2010;Bourgon & Armstrong, 2005;Hanson & Yates, 2013;Lowenkamp, & Latessa, 2002;Lowenkamp, Latessa, & Holsing-er, 2006). Thus, the most intensive levels of service should be reserved for higher risk offenders, while lower levels of intervention (or no intervention) should be applied to lower risk offenders. In fact, low risk offenders likely do not require specialized treatment at all, and will benefit from routine supervision (Andrews & Bonta, 2010;Hanson & Yates, 2013). Adherence to the risk principle, in addition to being the best use of limited resources, demonstrates that treatment is most effective when its level of intensity is matched to risk (Andrews & Bonta, 2010;, 1997Gendreau, Little, & Goggin, 1996;Gordon & Nicholaichuk, 1996;Hanson et al., 2009;Nicholaichuk, 1996). That is, when higher risk offenders receive higher intensity treatment, and moderate risk offenders receive intervention at more moderate levels of intensity, the impact on reduced recidivism is greatest. Furthermore, research indicates that, when risk and treatment intensity are not appropriately matched, recidivism can increase as a function of treatment, as in the case of lower risk offenders who receive treatment at an intensity that is greater than required to address their needs (Andrews & Bonta, 2010;Lowenkamp, & Latessa, 2002;Lowenkamp, et al., 2006). Among sexual offenders, specific static and dynamic risk factors have been associated with increased risk of recidivism. Static risk factors -those that cannot be changed through intervention -include younger age, previous sexual offenses, the commission of non-contact sexual offenses and non-sexual violent offenses, and offending against male victims, unrelated victims, and strangers (Hanson & Thornton, 1999). Dynamic risk factors are discussed below.
When considering treatment intensity, little research has been conducted regarding the most appropriate length of intervention, and practice varies substantially across jurisdictions (McGrath, et al., 2010). Some programs recommend between 80 (Beech & Mann, 2002) and 120 contact hours (e.g., Marshall, et al., 2006), while others recommend between 160 to 195 contact hours for moderate risk sexual offenders and approximately 300 hours of treatment contact for high risk offenders (Correctional Service Canada, 2000). In a comprehensive evaluation, Bourgon and Armstrong (2005) examined treatment intensity as a function of both risk and criminogenic needs (see below). They found that 100 contact hours was sufficient to reduce recidivism for general offenders presenting with moderate risk and few criminogenic needs, 200 hours was more effective when offenders were either high risk or had multiple criminogenic needs, and that 300 contact hours or more was required to reduce recidivism among offenders who were both higher risk and who had multiple criminogenic needs. Based on research pertaining to general offenders, as well as results from accredited sexual offender programs, Hanson & Yates (2013)  This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
offenders, and a minimum of 300 hours for sexual offenders presenting with high risk and high needs (the top 10% to 20% of the risk distribution; Hanson et al., 2012).
Need principle. The second principle of effective correctional intervention, the need principle, states that treatment and interventions such as supervision should explicitly target the criminogenic needs of offenders -that is, the specific risk factors that can be changed through intervention (i.e., dynamic risk factors) and that are empirically associated with recidivism risk (Andrews & Bonta, 2010). Targeting these risk factors for change leads to reduced re-offending.
Research indicates that, among sexual offenders, criminogenic needs include such risk factors as sexual deviance and antisocial lifestyle, the two strongest predictors of recidivism among sexual offenders (Hanson & Morton-Bourgon, 2004. It is important to note here that research has consistently found that sexual offenders are more likely to reoffend with offenses that are non-sexual in nature than to commit new sexual offenses (Hanson & Bussière, 1998;Hanson & Morton-Bourgon, 2005;Nicholaichuk et al., 2000). In addition, research indicates that the predictors of recidivism are different for different types of re-offending. That is, while the strongest predictors of sexual recidivism among sexual offenders include deviant sexual interest and antisocial orientation/lifestyle (antisocial personality, antisocial traits, a history of rule violation, and self-regulation problems such as impulsivity, lifestyle instability, and a history of non-sexual criminal offending), sexual deviance has been found to be unrelated to violent non-sexual offending (Hanson & Morton-Bourgon, 2005). Therefore, when determining criminogenic needs to be targeted during treatment, it is important to attend to the type of recidivism that is likely to occur, and to tailor treatment accordingly.
Recent research has additionally demonstrated specific dynamic risk factors that are associated with recidivism among sexual offenders. These include deviant sexual preferences, a lack of positive social influences, intimacy deficits, problems with sexual self-regulation, problems with general self-regulation, attitudes supportive of sexual assault, and problems with cooperation with supervision (Hanson, Harris, Scott, & Helmus, 2007). In treatment, it is recommended that these be assessed a priori, and included as appropriate in individualised treatment plans (Yates, Prescott, & Ward, 2010), along with assessment of static risk in order to determine treatment intensity by these factors in combination (Hanson, et al., 2007;Yates et al., 2010).
In addition to ensuring that factors empirically-related to risk of recidivism are addressed, the need principle also specifies that treatment should not focus on non-criminogenic needs --factors not found to be associated with recidivism --as this expends resources on addressing factors that are unlikely to result in reduced re-offending (Andrews & Bonta, 2010). Non-criminogenic factors include such areas as self-esteem, personal distress, victim empathy, and denial (Hanson & Bussière, 1998;Hanson & Morton-Bourgon, 2005;Yates, 2009a), none of which has been found to be reliably linked to recidivism in research. While it is common practice in treatment to address such factors, these are not empirically supported and are unlikely to be the best use of limited resources that aim to reduce reoffending.
Responsivity Principle. The third principle of effective correctional intervention, the responsivity principle, concerns the interaction between the individual and treatment. Specifically, this principle indicates that treatment, in addition to being cognitive-behavioral in orientation (see Andrews & Bonta, 2010), should be delivered in a manner that is responsive to various characteristics of the individual, such as language, culture, personality style, intelligence, anxiety levels, learning styles, and cognitive abilities, in order to increase their engagement and participation in treatment to ensure maximal effectiveness (Andrews & Bonta, 2010). These factors can affect clients' engagement with treatment, their motivation, their ability to understand and apply information presented in treatment to their own personal circumstances, and their manner of processing information presented in treatment. Therefore, treatment implementation should be varied and adapted to individual styles and abilities in order to maximize effectiveness, which involves significant skill on the part of clinicians.
Research support is strong for the application of the RNR model and its principles, and indicates that treatment that complies with these principles is superior to treatment that does not adhere to these principles and to criminal sanctions alone. Specifically, meta-analytic research clearly indicates that adherence to this model is effective for intervention with offenders in general, young offenders, violent offenders, and female offenders (Andrews, et al., 1990;Dowden & Andrews, 1999a, 1999b. Importantly, adherence to these principles also applies to the treatment of sexual offenders. Specifically, meta-analytic research indicates that, when treatment adheres to these principles, it is associated with reduced sexual re-offending. The most significant treatment effect has been found among treatment programs that adhered to all three principles (Hanson et al., 2009), and treatment effectiveness increases as a function of adherence to principles (odds ratios of 1.17, .64, .63, and .21, respectively for adherence to all three principles, only two, only one, and no adherence). The odds ratio is the likelihood of an event occurring or not occurring, and in this case indicates that treatment was most effective when it adhered to all three principles, and decreased progressively in effectiveness when treatment adhered to fewer principles.

 Cognitive-Behavioral Treatment
Although there is some debate regarding whether treatment with sexual offenders is effective, cognitive-behavioral treatment remains the most widely accepted and empirically supported model of sexual offender treatment with respect to reducing recidivism (e.g., Hanson et al., 2002;Lösel, & Schmucker, 2005). Based on behavioral, cognitive, and social learning theory and models (e.g., Bandura, 1986;Beck, 1964Beck, , 1967Beck, , 1976Yates et al., 2000Yates et al., , 2010, sexual offending is conceptualized as behavioral and cognitive patterns that are developed and maintained as a result of modeling, observational learning, and reinforcement of behavior, attitudes, and cognition. The focus of treatment is on altering patterns of behavioral, cognitive, and affective responding associated with sexual offending, such that such problematic, deviant, and/or criminal behavioral patterns and responses are replaced with adaptive, non-deviant, pro-social responding.
In doing so, treatment targets such responses as these are related to the specific dynamic risk factors known to be linked to risk for re-offending, as described above.
In practical application, cognitive-behavioral treatment involves changing attitudes, challenging cognitive distortions, addressing general self-regulation skills such as problem-solving, improving sexual, intimate, and social relationships, managing affective states, developing adaptive cognitive processes, and addressing sexual self-regulation, such as reducing deviant sexual arousal (Barbaree & Marshall, 1998;Marshall et al., 1999Marshall et al., , 2006Yates, 2002Yates, , 2003Yates et al., 2000Yates et al., , 2010. Appropriately applied, treatment should, therefore, explicitly target the development of client skills matched to these dynamic risk factors, in addition to risk factors for offending that is are not sexual in nature, such as general criminal attitudes. Cognitive-behavioral treatment also includes extensive rehearsal and practice of the adaptive and self-regulatory skills that are being learned by the client, as such skills require repetition in order to become well-entrenched in the individual's behavioral repertoire (Hanson, 1999;Hanson & Yates, 2004).
Common components of cognitive-behavioral intervention include general and sexual self-regulation, addressing relationship and intimacy deficits, developing empathy for victims of offending, challenging cognitive distortions, delineating the offense process and circumstances that trigger offending, inculcating responsibility for behavior in the offender, and developing relapse prevention plans. Some of these targets, such as victim empathy and taking responsibility, have received little research support for their contribution to reducing recidivism (e.g., Hanson & Morton-Bourgon, 2005;Yates, 2009a), and so are questionable treatment targets. In addition, much has been written in recent years regarding targeting cognitive distortions versus addressing cognitive schema in treatment. Targeting cognitive distortions (Abel, Becker, & Cunningham-Rathner, 1984;Barbaree, 1991) has historically been a common component This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
of sexual offender treatment. However, it has come to be recognized that cognitive schema represent individuals' underlying views and attitudes, while cognitive distortions are the products of these underlying schema (Mann & Beech, 2003). In cognitive theory (e.g., Beck, Freeman, & Davis, 2004), schema are cognitive structures that function to process, organize, and evaluate incoming information, direct cognitive activity, and influence information processing. Schema are based on individuals' previous experiences, contain attitudes, beliefs, and assumptions about the self, the world, and others, and provoke affective and behavioral responses. Schema have specific content and are activated by situational cues, particularly in ambiguous or threatening situations . Among sexual offenders, specific schema, such as sexual entitlement, a general view that the world is a hostile place, or the belief that children can consent to sexual activity, have been found to be implicated in sexual offending (e.g., Mann & Beech, 2003;Ward & Keenan, 1999). It is suggested, therefore, that treatment should focus more on identifying and altering schema, rather than focusing solely on cognitive distortions (Gannon, 2009). Lastly, Fernandez, Shingler, & Marshall (2006) and others (e.g., Yates et al., 2000Yates et al., , 2010 have observed that treatment for sexual offenders has, in recent years, over focused on cognitive aspects, with insufficient reliance on the rehearsal and practice that is essential for behavioral change, and recommend that treatment approaches explicitly place greater emphasis on skills development and practice, including in situ.

 Relapse Prevention
The relapse prevention (RP) approach has long been the predominant approach to sexual offender treatment (e.g., Laws, 1989Laws, , 2003Pithers, 1990;Pithers, Kashima, Cumming, & Beal, 1988;Pithers, Marques, Gibat, & Marlatt, 1983) and this continues to be the case (McGrath et al., 2010), in spite of a lack of evidence supporting its effectiveness with sexual offenders (Laws, 2003;Laws & Ward, 2006;Yates, 2005Yates, , 2007. In some respects, RP has become synonymous with cognitive-behavioral treatment; however, this is an inaccurate conceptionalization. Relapse prevention was initially developed as a post-treatment follow-up intervention for motivated alcoholic patients who had successfully ceased alcohol use but who demonstrated difficulty maintaining abstinence following treatment (Marlatt, 1982(Marlatt, , 1985. RP had an intuitive appeal to clinicians delivering sexual offender treatment and was applied to the treatment of this group following revisions to adapt the model to this population (Laws, 1989;Marlatt & Gordon, 1985;Marques, Day, & Nelson, 1992;Pithers, 1990;Pithers et al., 1988). It is noted that, at the time, there was a dearth of research pertaining to static and dynamic risk and the processes of sexual offending generally.
The goal of RP as initially conceptualized was to assist patients to identify, anticipate, and prevent high risk situations that could lead to lapses, defined in the original RP model as a temporary return to the alcohol use, as well as to avoid relapse, defined as a return to chronic alcohol abuse (Marlatt, 1982). In doing so, treatment involved teaching patients to cope with problems and high risk situations when these arose, and to address skill deficits in patients' abilities to do so. Applied to sexual offenders, this model did not fit and required adaptation (Laws & Ward, 2006;Marques et al., 1992;Pithers, 1990). For example, RP as applied to sexual offenders represents a "one size fits all" approach and does not adequately address the multiple treatment needs with which offenders present or the pathways to offending they follow, and it incorrectly regards sexual offending behavior as an addictive process, and presents such conceptual difficulties as defining what constitutes a lapse (Laws, 2003;Laws & Ward, 2006;Yates, 2005;Yates & Kingston, 2005;. For example, although illegal and potentially signaling an increase in risk, a single use of child pornography would be considered as a simple lapse within the RP framework. As such, many core constructs of the RP approach are not applicable to the sexual offending process.
The RP model also presents a narrow view of sexual offending behavior, not acknowledging multiple pathways to offending, and assumes that sexual offending results from negative affective states and a lack of coping skills. As such, this approach ignores processes of gratification and does not fit with offenders who explicitly plan their offending behavior (Laws, 2003;Laws & Ward, 2006;Yates, 2005;Yates & Kingston, 2005;.
Nonetheless, this model was adopted in the treatment of sexual offenders, and continues to be an accepted approach to treatment, in spite of a lack of empirical research supporting its application to intervention with sexual offenders (Hanson, 1996(Hanson, , 2000Laws, 2003;Laws, Hudson, & Ward, 2000;Laws & Ward, 2006;Yates, 2003Yates, , 2005Yates & Kingston, 2005;.

 Importance of Therapeutic Process in Treatment
While not a model of sexual offender treatment per se, the characteristics of therapists and the approaches they use in treatment, have been found in research to be associated with improved treatment outcomes (Beech & Fordham, 1997;Fernandez et al., 2006;Hanson et al., 2009;Marshall et al., 1999Marshall et al., , 2002Yates, 2002;Yates et al., 2000). For example, research indicates that establishing a positive therapeutic relationship with the client accounts for a significant proportion of the variance in treatment outcome Fernandez et al., 2006;Hanson, 2009;Witte, Gu , Nicholaichuck, & Wong, 2001;Mann, Webster, Schofield, & Marshall, 2004;Marshall et al., 1999Marshall et al., , 2003.
Specific therapist characteristics that have been shown to maximize treatment gains include demonstrating empathy, respect, warmth, friendliness, sincerity, genuineness, directness, confidence, and interest in the client. In addition, being a pro-social model, communicating clearly, listening actively, being "firm but fair, " reinforcing and encouraging clients without being collusive, creating opportunities for success, dealing appropriately with frustration and other client difficulties, being appropriately challenging without being aggressively confrontational, and creating a secure treatment atmosphere, all contribute to treatment outcome (Fernandez, 2006;Marshall et al., 1999Marshall et al., , 2002. Relatedly, using specific techniques of motivational enhancement is also viewed as essential to sexual offender treatment (Prescott, 2009). Importantly, creating a positive treatment environment leads to improved cooperation and compliance with treatment, treatment progress, enhanced motivation, and prevents termination or dropout from treatment (Beech & Fordham, 1997;Kear-Colwell & Pollack, 1997;Marshall et al., 1999;Miller, 1995).
As research clearly indicates that offenders who do not complete treatment re-offend at significantly higher rates than offenders who complete treatment (Hanson & Bussière, 1998;Hanson et al., 2002), it is essential that treatment is delivered in a positive manner that is motivating to clients.

 Self-Regulation Model
The self-regulation model (SRM; Ward & Hudson, 1998;Ward et al., 1995) is an emerging approach to sexual offender treatment that was developed as a result of shortcomings, such as those described above, with the RP approach to treatment. Originally a nine-phase model of the offense process, the model was developed specifically for sexual offenders based on self-regulation principles of behavior (Baumeister & Heatherton, 1996;Karoly, 1993;Thompson, 1994). The SRM explicitly takes into account variability in offense-related goals and the manner by which individuals regulate their behavior in order to achieve these goals. Offense-related goals include both inhibitory or avoidance goals (i.e., directed toward avoidance of undesired states or outcomes) and appetitive or approach goals (i.e., directed toward the attainment of desired states and outcomes). Offenders with avoidant goals desire or attempt to refrain from offending, while offenders with approach goals more actively seek out opportunities to offend. Achieving goals is based on individuals' self-regulation capacity, with some offenders failing to control behavior (under-regulation/disinhibition), others attempting to actively control their behavior using strategies that are ultimately counterproductive and ineffective (mis-regulation), and others having intact self-regulation abilities and an absence of self-regulation deficits (Ward et al., 1995Yates, 2007;Yates & Kingston, 2005).
Therefore, according to the SRM, offenders may follow one of four pathways to offending, as follows: The avoidant-passive pathway is associated with the desire to refrain from sexual offending (avoidance goal), but a lack of the required awareness and skills to effectively control behavior in order to achieve this goal. Thus, although individuals following this pathway desire to avoid offending, they do not implement strategies to do so, resulting in failure to achieve the avoidance goal and, ultimately, of-This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
fending. Self-regulation is under-regulated, and when confronted with the possibility of offending, disinhibition of behavior, loss of control, impulsivity, and anxiety occur, alongside goal failure. The avoidant-active pathway is a mis-regulation pathway along which individuals actively implement strategies to cope with the desire and opportunities to offend in order to meet an avoidance goal. However, the strategies selected are ineffective and, in some instances, result in the iatrogenic effect of increasing the likelihood of offending. For example, individuals may masturbate to deviant images in an attempt to avoid committing a hands-on offense, or may use substances to regulate mood. However, such strategies may function to disinhibit the individual or to further entrench deviant arousal, thus increasing risk to offend. A key difference with this pathway is that the individual is aware that there is a problem and that action is required and actively implements strategies to prevent offending. The approach-automatic pathway is associated with approach-motivated goals with respect to offending and is characterized by under-regulation. Individuals following this pathway do not desire to prevent offending, nor do they attempt to refrain from pursuing offense-related goals. Offending occurs as a response to situational cues in the immediate environment, and cognitive schema that support offending are activated by these cues. In addition, offending may appear impulsive. Lastly, the approach-explicit pathway is associated with intact self-regulation and an approach goal with respect to offending. Sexual offenses are explicitly and overtly planned in order to achieve a desired objective, such as sexual gratification, and offending is associated with attitudes and core beliefs that support sexual aggression as an appropriate means by which to achieve these goals.
From this brief overview, it is evident that the SRM is a more comprehensive approach to the sexual offending process than other models, such as RP, which posits a single pathway to offending, and is more consistent with the risk/need/responsivity (RNR) model described above. Furthermore, the SRM allows for a more comprehensive and individualized approach to treatment that better addresses individual dynamic risk factors and motivations for sexual offending and that is tailored to offense pathway (Yates & Kingston, 2005;Ward et al., 2005Yates & Ward, 2008;Yates et al., 2010).

 The Good Lives Model
The good lives model (GLM) is another emerging approach to sexual offender treatment, and was developed as a result of shortcomings identified with the RNR approach to intervention (Ward & Brown, 2004;Ward & Gannon, 2006;Ward, Melser, & Yates, 2007;Ward & Stewart, 2003). For example, while essential, the focus of the RNR approach on risk and criminogenic needs, was criticized as insufficient for treatment effectiveness due to its focus on deficits, risk management, and avoidance goals, as well as its inability to sufficiently motivate clients to change (Mann et al., 2007;Ward & Gannon, 2006;Ward et al., 2007). This is important given that sexual offenders tend not to be particularly motivated to participate in treatment (Thornton, 1997), resulting in the need for motivational approaches to treatment (Prescott, 2009;Yates, 2009b).
Briefly, the GLM proposes that, like other human beings, sexual offenders are goal-directed and seek to acquire fundamental primary human goods, defined as actions, experiences, and activities that are intrinsically beneficial to individual well-being and that are sought for their own sake (Ward & Gannon, 2006;Ward & Stewart, 2003). Examples of primary human goods, also termed common life goals (Yates & Prescott, 2011, 2012, include relationships and friendships, happiness and sexual pleasure, being independent, and attaining peace of mind or emotional equilibrium. The GLM posits that sexual offending results from maladaptive strategies (termed secondary or instrumental goods) that individuals use to attain these life goals. For example, an offender may desire intimacy, but as a result of factors such as emotional identification with children, turns to children to meet this need. Similarly, an individual may utilize substances to regulate mood and attain peace of mind, or utilize aggression and violence to achieve the goal of independence and autonomy. In the GLM framework, the problem does not lie in the life goals of each individual but in the ways they attempt to achieve these goals, which lead to life problems and to sexual and other offending.
In treatment using the GLM approach, there is an explicit focus on assisting individuals to attain important and valued life goals in pro-social, non-harmful ways Yates et al., 2010;Yates & Prescott, 2011). This model, unlike the RNR and RP, also explicitly utilizes approach rather than avoidance goals in treatment. That is, rather than focus solely on those activities and behaviors in which clients cannot engage, treatment includes actively working toward and attaining important life goals. Thus, for example, treatment actively assists clients to attain independence and autonomy without abusing others, to achieve intimacy without engaging in sexual activity with children, to experience sexual pleasure in non-harmful and healthy ways, and so forth. This is an important element, given that approach goals are more easily attainable and sustainable over the long-term than are avoidance goals (Mann et al., 2004). Using the GLM approach, it is hypothesized that will not only offenders be assisted to attain greater well-being but also that dynamic risk factors will be mitigated, thereby reducing risk to reoffend (Ward & Stewart, 2003), responsivity will be better addressed in treatment, and offenders will be more motivated to change and to participate in treatment. Importantly, the GLM cannot be implemented in the absence of risk management approaches and explicitly targeting criminogenic needs in treatment -to do so runs the risk of ignoring important risk factors and potentially increasing recidivism Yates et al., 2010). In order to ensure the inclusion of risk factors and risk management, the GLM has been integrated with the self-regulation model in a comprehensive approach to assessment and treatment Yates, Kingston, & Ward, 2009;Yates & Ward, 2008;Yates, in press;Yates et al., 2010;Yates & Prescott, 2011) that is consistent with the principles of effective correctional intervention, and that employs practices and techniques from demonstrated effective interventions.
Research into the GLM as an approach to sexual offender treatment is in its infancy, although does provide some preliminary support. For example, good lives constructs have been found to be differentially associated with offense characteristics , as well as static risk to re-offend, dynamic risk factors, and sexual offense pathway , thus suggesting the potential utility of the GLM with sexual offenders with respect to risk, need, and self-regulation. In one of the first empirical investigations of the model, Simons, McCullar, and Tyler (2006) found that, compared to an RP approach, offenders participating in GLM-based treatment were more likely to complete treatment, remained in treatment longer, and were rated by therapists as more motivated to participate in treatment. In addition, pre-/post-treatment comparisons indicated that offenders participating in either program improved similarly on social skills, victim empathy, and problem-solving ability. However, those who participated in the GLM approach demonstrated significantly greater improvements compared to clients who received the RP approach, and demonstrated significantly better coping skills post-treatment. Importantly, offenders participating in the GLM program were found to drop out from treatment at much lower rates ). Conversely, Harkins, Flak, Beech, and Woodhams (2012) found no differences in attrition rates between and their RP and GLM programs, although qualitative data suggested that both therapists and clients perceived the program as more positive and future-oriented than RP.

 Conclusions and Future Directions
This article provides a necessarily brief overview of current and emerging approaches to the treatment of sexual offenders. Research clearly supports the application of an approach in which treatment intensity varies on the basis of the risk for recidivism posed by offenders, with treatment being more effective when it is applied to higher risk cases, when it targets known dynamic risk factors for recidivism, and when it is responsive to the characteristics of individual offenders. Research further demonstrates that targeting specific known risk factors using cognitive-behavioral methods, is most effective in reducing recidivism among sexual offenders. An explicit skills-based approach is recommended in order to enable treatment participants to change cognition, affect, and behavior such that these become entrenched their behavioral repertoire. Emerging research suggests that the traditional relapse prevention approach, not demonstrated to be effective to date, is overly simplistic in its conceptualization of pathways to offending, and the self-regulation approach, with its broader conceptualization of offense process, dynamics, and motivations, is gaining substantial support for its application to the sexual offender treatment. Research also indicates that there are essential characteristics of the intervention itself and of therapists, that are associated with improved outcomes. In a related vein, the good lives model, has been found to have some validity and, importantly, to be associated with increased motivation and reduced treatment attrition, although additional research into this model is required. It is suggested that treatment of sexual offenders can be most effective when it is based in empirically-demonstrated models and methods and in a manner integrates various approaches (Yates, in press), and that such an approach shows the greatest promise of reducing recidivism, improving offenders' lives, and contributing to community safety and growth.
 References