Keywords

1 Introduction

Forty years ago, Martinson (1976, 1979) concluded that the search for programs that might reduce recidivism in offenders was useless because nothing works. Nowadays there is increasing evidence that this critique was wrong because there is a growing body of evidence that supports the conclusion that many programs and practices can reduce recidivism (Weisburd, Farrington, & Gill, 2017). However, there is also evidence of less effective practices and, importantly, of practices that may even cause harm. Therefore, this chapter attempts to determine the substantive conditions of effective treatment programs for violent and sexually violent offenders.

2 Programs for Offenders

The first cognitive-behavioral programs for offenders were developed during the 1970s, mainly in the Anglo-Saxon countries, followed by several Western European countries, sometimes almost a decade later. Most of these programs had or have a cognitive-behavioral approach that is based on social learning principles. Studies on the effect of these programs were conducted by comparing the rate of recidivism of a sample of treated offenders with that of a matched sample of untreated offenders. For example, the Reasoning and Rehabilitation program (R&R) designed by Ross, Fabiano, and Ewles (1988) was delivered in Canada and the United States in the early 1990s. This program is based on the proposal that offenders cannot realize their goals because they lack the necessary cognitive and behavioral skills. In 36 two-hour sessions, 6–12 participants extend or modify their cognitions and skills in such a way that the individual recidivism risk factors decrease. Robinson and Porporino (2001) reviewed several studies on the effects of R&R, including a study in more than 4000 Canadian offenders who completed the R&R program between 1989 and 1994 (Robinson, 1995). The recidivism rate in this study turned out to be 19.7% for R&R participants and 24.8% for the control group a year after discharge, which indicates a reduction of 20.5% in the treatment group.

A modified version of R&R was implemented in the United Kingdom from 1993 onwards, namely the Enhanced Thinking Skills (ETS) program with 20 sessions. Friendship, Blud, Erikson, Travers, and Thornton (2003a) compared a group of 667 offenders who followed R&R or ETS with a control group of 1801 offenders. Treatment turned out to produce a robust reduction in the probability of reconviction when other relevant variables were controlled for. For treated offenders, the percentage point reduction in reconviction was 14% in medium-low risk offenders and 11% in medium-high risk offenders. For the low- and high-risk offenders there was a trend in the expected direction, although the difference was not statistically significant.

Lipsey, Chapman, and Landenberger (2001) studied cognitive-behavioral programs focused on teaching new cognitive skills. The participants in these studies were defined as general criminal offenders, juveniles or adults, who are treated while on probation, incarcerated, or in aftercare/parole. Participants were at random or matched assigned to a cognitive-behavioral program or treatment-as-usual. A meta-analysis of fourteen studies from 1985 that met this condition showed that cognitive-behavioral programs were indeed effective interventions. Treated offenders on average recidivated at a rate of about two-thirds that of offenders in the treatment-as-usual control groups. Moreover, the most effective programs reduced recidivism rates to about one-third of the rate for untreated controls. However, most of the substantial recidivism reductions were produced by the programs that were designed by researchers, and no other comparisons were made for instance between the effects of programs in correctional institutions and those for offenders on probation or parole.

Landenberger and Lipsey (2005) performed a meta-analysis of 58 studies on the effects of cognitive-behavioral treatment programs on the recidivism of adult and juvenile offenders to find out which factors contributed to the positive effects of these programs.

Participants in these studies were criminal offenders, either juveniles or adults, treated while on probation, incarcerated/institutionalized, or during aftercare/parole, and were drawn from a general offender population. Offender samples that were selected for, or restricted to, persons committing specific types of offenses were not included. They found that the factors independently related to the effect sizes were (a) the risk level of the offenders, (b) the quality of the program implementation, and (c) the content of the program. The inclusion of anger management and interpersonal problem-solving in the program were associated with more substantial effects, while victim impact and behavior modification were associated with smaller effects. None of the major CBT brand-name programs, for instance, R&R (Ross et al., 1988) or Aggression Replacement Training (ART; Goldstein, Glick, & Gibbs, 1998), produced larger effects on recidivism than the other programs. According to these researchers, high-quality implementation, close monitoring of the performance of the programs, and adequate training for the providers seem to most strongly characterize effective CBT (Lipsey, Landenberger, & Wilson, 2007).

Pearson, Lipton, Cleland, and Yee (2002) conducted a meta-analysis of 69 research studies, published between 1968 and 1996, on the effectiveness of behavioral and cognitive-behavioral treatment in reducing recidivism for offenders. They defined behavior modification/behavior therapy as arranging contingencies of positive reinforcement to develop and maintain appropriate patterns of behavior, such as contingency contracting and token economies. Cognitive-behavioral treatments were described as treatments that include attention to cognitive and emotional processes that function between the stimuli received and the overt behaviors enacted, for example, problem-solving skills training and anger management training. They found that reduced recidivism was mainly reported in cognitive-behavioral interventions, rather than programs with standard behavior modification approaches. However, no information was provided whether or not the cognitive-behavioral programs were individually or group oriented.

2.1 Group Programs for Offenders

A meta-analysis of 20 studies of group-oriented cognitive behavioral programs for offenders was performed by Wilson, Allen, and MacKenzie (2000; cited in Lipsey et al., 2001). These programs were found to reduce recidivism by 20–30% compared to untreated control groups. However, some studies concerned general samples of offenders, others specialized types of offenders, for example, sex offenders, drug offenders, driving under the influence cases, or batterers. Therefore, much of the variability in effects found across studies may have been due to differences between types of offenders. In a later study, Wilson, Bouffard, and MacKenzie (2005) investigated the effectiveness of cognitive-behavioral group programs for offenders in 20 studies. They found that all higher quality studies reported positive effects favoring the cognitive-behavioral treatment programs. Specifically, actual reductions in recidivism were observed for Moral Reconation Therapy (Little & Robinson, 1986), R&R (Ross et al., 1988), and various cognitive-restructuring programs. More specifically, there was evidence for the effectiveness of programs that focus on cognitive skills and cognitive restructuring as well as programs that emphasize moral teachings and reasoning.

3 Programs for Violent Offenders

Turning specifically to programs for violent offenders, the effectiveness of cognitive behavioral group programs has been examined in some ways, including meta-analyses, systematic or narrative reviews, or studies examining the evaluation of individual programs. For instance, Dowden and Andrews (2000) conducted a meta-analysis of 35 studies to examine the effectiveness of correctional treatment in reducing violent recidivism. As expected, cognitive-behavioral and social learning strategies were associated with substantially larger treatment effects than those produced by non-behavioral approaches. Also, programs that predominantly targeted criminogenic needs produced greater treatment effects than programs that predominantly targeted non-criminogenic needs.

Polaschek and Collie (2004) reviewed nine studies on treatment programs with a cognitive-behavioral approach, and had an appropriate methodological rigor: two cognitive programs, three anger management programs, and four programs that used multiple intervention targets and strategies. Although most programs had small to large positive effects on violent and nonviolent recidivism, little could be concluded about the most promising theoretical framework for future program development.

A systematic review of 11 high-quality research reports was conducted by Joliffe and Farrington (2007) to assess the evidence on the effectiveness of interventions with adult male violent offenders and to identify potential mediators and moderators. Analysis of the reports suggested that interventions with violent offenders were effective both at reducing general and violent recidivism. Violent offenders with extensive criminal histories were more likely to recidivate than general offenders and tended to be more challenging to engage in treatment. Also, there was some evidence to suggest that those interventions of greater overall duration were more effective and that the greater duration per session was associated with greater effect for both general and violent recidivism. However, what could not be assessed in this review is what the optimal dosage of intervention might be. The evidence also suggested that interventions which addressed cognitive skills, anger control, used role play and relapse prevention and had offenders to complete homework were more effective than those interventions that did not.

3.1 Group Programs for Violent Offenders

Cortoni, Nunes, and Latendresse (2006) evaluated the Canadian Violence Prevention Program (VPP) for incarcerated violent offenders, who have committed at least two violent offenses and who are at high risk to commit future violent crimes. The goal of the VPP is to reduce the likelihood of violent recidivism among the male offender population. The intervention phase of the VPP consists of the following modules: Making Change (6 sessions), Violence Awareness (12 sessions), Anger Control (12 sessions), Solving Problems (10 sessions), Social Attitudes (10 sessions), Positive Relationships (8 sessions), Conflict Resolution (8 sessions), Positive Lifestyles (8 sessions), Self-Control (8 sessions), and Violence Prevention (12 sessions). Although VPP is not explicitly called a cognitive behavioral program, the modules indicate a cognitive-behavioral approach because they are, according to the researchers, based on social learning and information-processing theories. Five-hundred treated offenders were compared to a matched, non-treated comparison group of 466 offenders. Results showed that offenders who had completed the program had significantly lower rates of recidivism than the non-treated offenders.

Finally, Polaschek, Yesberg, Bell, Casey, and Dickson (2016) compared 121 violent offenders who had completed treatment in a New Zealand high-risk special treatment unit with 154 violent offenders who were eligible to attend the treatment program but had not done so. The treatment comprised a structured, closed-group cognitive-behavioral intervention combined with a “hierarchical democratic therapeutic community approach.” The group intervention included sessions on areas such as offense-supportive thinking, mood management, problem-solving skills, and relapse prevention planning. In addition to the treatment program, offenders took part in several activities within the unit community and developed a reintegration plan for release from prison. The program was delivered to closed groups of 10 offenders for approximately 250 h over 25 weeks and remained in the treatment unit for 10–12 months. Completers of the treatment program were found significantly less likely to breach parole, to be reconvicted for any offense, to be reconvicted for violence, and to be imprisoned for a new offense within 12 months after release.

3.2 Aggression Replacement Training (ART)

ART is a North American multimodal intervention developed by Goldstein et al. (1998) to improve prosocial behavior in children and adolescents who perform aggressive and violent behaviors. However, as this intervention was the starting point for the development of the treatment programs as described in Chap. 6, it will also be considered here. Goldstein et al. (1998) regarded aggressive behavior as a result of inadequate emotional control, a limited repertoire of social skills, and a lack of moral standards and values. Consequently, ART has three main components, namely anger management, social skills, and moral reasoning. In the original version, ART takes 10 weeks with three sessions per week, one for each component. The groups usually include 6–8 participants. Homework assignments are given during all modules of ART in order to achieve generalization of the learned skills to new situations. In 1999, the state of Washington (USA) decided to investigate ART together with Functional Family Therapy (FFT; Alexander & Parsons, 1982) and Multisystemic Therapy (MST; Henggeler, 1999). In a study by the Washington State Institute for Public Policy (WISP) an experimental group of 704 juvenile offenders with a medium to high recidivism risk was compared with a control group of 525 juvenile offenders. ART resulted in 24% lower recidivism than a control group throughout 18 months in the experimental group (Barnoski, 2004).

More recent research has yielded mixed results. Gunderson and Svartdal (2010) found that the behavior of 77 Norwegian children (girls and boys) who followed ART improved according to trainers, parents, and teachers but in the control sample of 63 children positive changes were also observed. Similar findings were yielded in a study by Koposov, Gunderson, and Svartdal (2014) who investigated ART in a sample of 145 Russian children (girls and boys). They found a significant improvement of social skills among children who had followed ART but in the control sample of 90 children an improvement in social skills and a reduction of problem behaviors were also found. A study by Langeveld, Gunderson, and Svartdal (2012) in 112 Norwegian children and adolescents (girls and boys) showed that after ART social competence was significantly improved and behavioral problems significantly decreased. During awaiting period before the start of the training program, no changes in social competence and behavioral problems were found. Finally, Hornsveld, Kraaimaat, Muris, Zwets, and Kanters (2015) explored the effects of ART in a group of 62 Dutch violent young men aged 16–21 years, who were required to follow a treatment program in a forensic psychiatric outpatient clinic. During the waiting period, the patients did not change on most measures, although they displayed a significant increase in anger. The training produced significant decreases in physical aggression and social anxiety and showed trends toward a decline in self-reported hostility, general aggression, and anger. After the training, the patients scored comparably with a reference group on measures of hostility and aggressive behavior. However, Brännström, Kaunitz, Andershed, South, and Smedslund (2016) concluded from 16 studies that there is insufficient support for the effectiveness of ART. They also noted that almost half of the studies were conducted by researchers who had an interest in the research.

ART has been applied not only to juvenile but also to adult offenders. Hatcher et al. (2010) studied ART in a group of British violent adult offenders serving community sentences in the Probation Service. An experimental group of 53 violent offenders who were required to follow ART was compared with a matched control group of 53 violent offenders who were not obliged to follow the training. Twenty offenders in the experimental group were reconvicted, compared with 27 offenders in the comparison control group. Because of these findings, Hatcher et al. (2010) concluded that “the ART programme may be effective with adult males in community settings” (p. 529). Hornsveld, Nijman, and Kraaimaat (2008) found some support for the effectiveness of an ART-based program in a sample of 89 Dutch violent forensic psychiatric inpatients, while an exploratory study by Zwets et al. (2016) in Dutch violent forensic psychiatric inpatients demonstrated that ART in combination with psychomotor therapy (PMT) resulted in clinically significant improvements of observed social behavior, observed aggressive behavior, and self-reported anger, but there were no differences in treatment effects between the experimental sample of 22 inpatients group and the control sample of 15 inpatients on primary outcome measures. However, on secondary outcome measures, the experimental sample displayed somewhat more improvement than the control sample.

3.3 Conclusions and Recommendations

The above studies show that programs that follow a cognitive-behavioral approach are the most successful in reducing recidivism among offenders in general and in violent offenders in particular. There are also strong indications that programs that focus on criminogenic risk factors yield better results than programs that do not. These studies also suggest that interventions that aim to teach prosocial cognitive skills and to learn to control anger, using role play, relapse prevention, and homework assignments, are more effective than interventions that do not. A limitation, however, is that the meta-analyses do not provide information about the treatment integrity of the programs studied. Manuals turn out not to be public and hardly any mention is made of the knowledge and experience of those who led the programs.

4 Programs for Sexually Violent Offenders

Group treatment with sex offenders is utilized far more frequently than individual treatment. A 2013 survey of North American sex offender programs showed that 90% of the programs were group based, although there are hardly studies which explored whether one modality is more effective than another (Serran, Marshall, Marshall, & O’Brien, 2013).

4.1 Recidivism of Treated and Untreated Sex Offenders

Studies on the effect of the first cognitive-behavioral programs for sexual offenders were conducted by comparing the rate of recidivism of a sample of treated sex offenders with that of a matched sample of untreated sex offenders. The programs focused not only on the modification of deviant sexual orientations but also on improvement of relational skills and a more mature sexual orientation (Barlow, 1973). Due to the introduction of the relapse prevention model in the early 1980s (Pithers, Marques, Gibat, & Marlatt, 1983), modules were added to promote empathy, social skills, relational skills, anger management, and self-esteem (Marshall & Hollin, 2014; Marshall & Laws, 2003). Pithers, Kashima, Cumming, Beal, and Buell (1988) defined relapse as a process of successive steps which result in sexually violent behavior, namely negative mood, fantasies about deviant sexual behavior, distorted cognitions, making plans for committing sexual offenses, deviant masturbation fantasies, and finally committing the offense. Therefore, the relapse prevention-based programs aimed to provide sex offenders insight into the situations that trigger an offense and into the various associated thoughts, feelings, and behaviors involved. Also, sex offenders learned coping skills for dealing with future problems or risk situations, and they had to draw up a plan that would minimize the risk of unexpected risk situations (Marshall & Anderson, 2000).

The most influential study on the effect of a program for sexual offenders that included a relapse prevention component was the Sex Offender Treatment and Evaluation Project in California (SOTEP; Marques, Day, Nelson, & West, 1994). In their controlled study, Marques, Wiederanders, Day, Nelson, and Van Ommeren (2005) compared an experimental group of 259 sex offenders with a matched group of 225 sex offenders who did not follow the program but participated in the investigation, and a control group of 220 sex offenders who did not follow the program and did not want to participate in the study. The treatment program at the Atascadero State Hospital lasted 2 years and included three sessions of 90 min a week. After discharge, participants in the clinical program followed a 1-year aftercare program (SOAP). Noteworthy, the researchers found no difference between the three conditions for sexual or violent recidivism throughout 8 years, but the sex offenders who “met the program’s treatment goals had lower re-offense rates than those who did not” (p. 79). However, in the study of Marques et al. (2005) the base rate of recidivism for the untreated offenders was low (13–14%) compared to 10.8% in the experimental group.

More or less parallel to developments in the United States, the Sex Offender Treatment Program (SOTP) was implemented in Britain in 1992 (Grubin & Thornton, 1994). This program included a Core and a Better Lives Booster program for all offenders, as well as an Extended program that provides a more prolonged and more intensive program for high-risk and high-need sex offenders with specific modules on assessment grounds. The Core program consisted of challenging distorted cognitions, promoting empathy, and drawing up an offense scenario to develop relapse prevention strategies. For example, the modules of the Extended program focused on a lack of social skills or insufficient anger management. Finally, the Booster program consisted of integrated skills training, which reconsidered the relapse prevention strategies from the Core program and the Extended Program skills. Individual group meetings could also be added to the group meetings of the three programs. The Core and Booster programs lasted 140–160 h (Grubin & Thornton, 1994). However, an evaluation of the SOTP by Friendship, Mann, and Beech (2003b), comparing a group of 647 treated sex offenders to a group of 1910 matched untreated sex offenders, yielded disappointing results. After 2 years, the percentage of sexual recidivism in the treated group was not significantly lower than in the untreated group, but significant differences were found between the two groups in the recidivism rate when sexual offenses were combined with violent offenses. Friendship and colleagues (2003b) attributed the results primarily to the low base rate of sex offenses, but also to insufficient treatment integrity (Hollin, 1995). Other debilitating factors were that the program changed in between (Mann & Thornton, 2000), the training level of the therapists/trainers (psychologists, recruitment staff, and prison guards) varied greatly, and supervision appeared to be insufficient. A recent large-scale evaluation (Mews, Di Bella, & Purver, 2017) concluded that the Core SOTP was not producing reductions in reconviction (and perhaps in some cases, increased reconvictions) and so the SOTP suite of program was withdrawn by the English and Welsh Prison Service.

However, an evaluation of international studies in the period 1970–1998 by Grossman, Martis, and Fichtner (1999) indicated that both hormonal interventions and cognitive-behavioral therapy programs lead to a 30% reduction in recidivism. Outpatient treatments were more effective than clinical treatments, possibly because the outpatients had a lower recidivism risk than the participants in the clinical programs. An early meta-analysis conducted by Hanson et al. (2002) reviewed 43 studies conducted before 2000. Averaged across all studies, the sexual offense recidivism rate was lower for the treatment groups (12.3%) than the comparison groups (16.8%). Currently, mainly cognitive-behavioral treatments are associated with reductions in both sexual recidivism (from 17.4% to 9.9%) and general recidivism (from 51% to 32%). Older forms of treatment (operating before 1980) appeared to have little effect.

A meta-analysis of 69 studies including 22,181 sex offenders by Lösel and Schmucker (2005) also showed that programs with a relapse prevention component produce positive results, while approaches such as castration, hormone treatment, and some psychotherapeutic approaches show no impact at all. Similar findings were found by De Ruiter, Veen, and Greeven (2008) in a meta-analysis of five studies with a total of 578 rapists, which demonstrated that cognitive behavioral therapy aimed at relapse prevention resulted in a decrease in sexual and general recidivism. The authors concluded that the positive treatment effects found were mainly observed after intensive and relatively long-term intramural treatment, which in some cases was followed by an outpatient care program.

As an alternative to the principles of relapse prevention, Ward and Hudson (2000) developed a Self-Regulation Model with a focus on the various ways in which individuals realize key goals. The model has nine stages for sex offenders: event, desire for forbidden sex, determination of goals, determination of approach, being in a high-risk situation, relapse, sexual offense, evaluation after the offense, and attitude toward future offenses. The authors distinguished four pathways that result in a sexual offense, namely (a) avoidant/passive: the offender loses self-control, (b) avoidant/active: the offender firstly shows avoidance behavior but loses self-control, (c) approach/automatic: the offender commits the sexual offense without thinking, and (d) approach/explicit: the offender commits the sexual offense consciously and feels sexually satisfied (Ward & Hudson, 2000). Following the pathways of this model, Bickley and Beech (2002) could classify 87 child sexual abusers to one of these four paths. Approach sex offenders reported more cognitive distortions, among others, on the consequences of the abuse for the victims than avoiding sex offenders with boys or boys and girls as victims. Avoidant sex offenders often had a long-term relationship with an adult partner during the offense and had mostly abused girls. Compared to active sex abusers, passive sex abusers more often blamed the circumstances and were more often convicted of sexual offenses. Yates and Kingston (2006) also supported the Self-Regulation Model and investigated 80 intra-familial child sexual abusers, extra-familial child sexual abusers, and rapists. On intra-familial child sexual abusers, avoidant/passive path was found to be applicable, and to rapists both the approach/automatic and the approach/explicit path. The validity and usability of the Self-Regulation Model in a treatment context were later evaluated by Kingston, Yates, and Olver (2014) in a group of 275 male sex offenders following one of Canada’s Correctional Service (Yates et al., 2000) treatment programs. Results indicated that participation in treatment resulted in moderate to large sized improvements from pretreatment to posttreatment on a dynamic risk assessment measure and several self-reported treatment targets. These changes were, in some cases, differentially associated with self-regulation pathways, suggesting that offense pathway is a clinically relevant variable when evaluating treatment change. Concerning outcome, individuals following the approach pathways, particularly the approach-automatic pathway, demonstrated higher failure rates than individuals following avoidant pathways. However, many of these differences were less pronounced when taking risk of recidivism into account. Because no control group was used in this study, the empirical support for the Self-Regulation Model is for the time being limited.

A population study by Hanson and Harris (2000) of the dynamic risk factors in 208 recidivated and 201 matched non-recidivated Canadian sex offenders revealed that the first group was characterized by limited social support, positive attitudes toward sexual violence, antisocial lifestyles, limited self-control strategies, and problems with supervision. Also, the recidivists appeared to have experienced more subjective tension and anger before the new offense. The different types of sex offenders appeared to provide similar risk factors, although fewer factors were significant among the girl’s sexual abusers than among the boy’s sexual abusers, or the rapists. These findings may be on the conservative side because the recidivists had followed significantly fewer treatment programs than the non-recidivists.

In a critical discussion of the relapse prevention model, Hanson (2000) suggested that treatment programs should focus more on risk factors that initiate criminal behavior (for example, deviant sexual interests) and risk factors that prevent relapse (for example, self-regulation strategies) than on risk factors that can result in a sexual offense. Hanson’s (2000) critique is in line with the Risk-Need-Responsivity principles advocated by Andrews and Bonta (2010) who showed that recidivism risk of violent offenders is related to antisocial history, antisocial personality, antisocial cognitions, antisocial network, family and/or relationship problems, limited education and/or no work, insufficient leisure activities, and substance abuse. Hanson, Bourgon, Helmus, and Hodgson (2009) found that these factors also apply to sex offenders. A meta-analysis of 23 high-quality studies revealed that the recidivism risk of treated sex offenders (10.9%) was significantly lower than for untreated sex offenders (19.2%). Programs based on the Risk-Need-Responsivity principles gave the greatest decrease in sexual and general recidivism. Smid, Kamphuis, Wever, and Van Beek (2014) conducted a survey in which a sample of 106 Dutch treated sex offenders was compared to a group of 188 untreated sexual offenders, using the Static-99R. The treatment lasted an average of 8 years and 4 months and consisted of individual or group psychotherapy, a relapse prevention program, practical training and/or education, and any medication. The researchers found a treatment effect in the high-risk offenders, but not in the average- or low-risk offenders, comparing the pre-measurement with the post-measurement.

Some studies into the effectiveness of treatment programs concerned only child sexual abusers. Långström et al. (2013) found evidence of a reduction in recidivism risk as a result of pharmacological, psychological, or psychoeducational interventions based on eight methodologically sound studies in child sexual abusers. Grønnerød, Grønnerød, and Grøndahl (2015) supported this finding in their meta-analysis of 14 studies with a robust research design on psychotherapeutic interventions, which showed that the effect sizes of such interventions in this population are very small. Out of ten studies conducted by Walton and Chou (2015), which investigated the impact of psychological interventions on child abusers, there were only three who could demonstrate a lower recidivism risk. These three well-controlled studies involved cognitive-behavioral treatment programs.

In 2004, Ward and Marshall suggested that treatment of negative offense-related behavior in sex offenders (for example, cognitive distortions) should be complemented by promoting positive behavior (for example, intimate relationships skills), thus giving them the chance to reduce future offenses. Ward and colleagues introduced the Good Lives Model, which focuses on the potential of sex offenders to improve the quality of their lives. According to this model, ten main aspects of existence have to be addressed in the preparation of a treatment plan: healthy life, knowledge, play and work, independence, inner peace, relationships, community spirit, spirituality, happiness, and creativity (Wills, Yates, Gannon, & Ward, 2013). However, Marshall and Marshall (2014) noted that, to date, hardly any controlled research has been conducted on the effectiveness of Good Lives Model-based treatment programs. Therefore, they advocated a balanced approach, addressing both negative and positive behaviors (Marshall et al., 2005). Their treatment program (Marshall, Marshall, Serran, & Fernandez, 2006) focuses on both negative factors and sexual abuse in childhood, inability to attachment and intimacy, deviant sexual interests or sexual compulsion as well as strengthening positive (protective) factors such as learning coping skills, emotion regulation, and building a more or less satisfying existence. Marshall, Marshall, Serran, and O’Brien (2013) concluded that effective programs meet four criteria: (1) they focus on dynamic criminological factors, (2) are offered in a user-friendly manner, (3) apply procedures that allow criminogenic needs to be changed, and (4) place treatment in a positive framework according to the principles of the Good Lives Model. In a study by Olver, Marshall, Marshall, and Nicholaichuk (2018), the early version of Rockwood’s prison-based sex offender program that has elements of a strength-based approach was compared with sex offender treatment programs of the Correctional Service of Canada and with no-treatment sex offenders. Both treatment groups displayed significantly lower rates of both sexual and violent reoffending when compared with the no-treatment sex offenders, but the Rockwood program generated the lowest recidivism rates.

4.2 Recidivism and Other Treatment Outcomes

The promising results of cognitive-behavioral treatment programs for sexually violent offenders at the beginning of the twenty-first century were followed by studies in which the differences between a pretest at the beginning and a posttest at the end of a program were related to recidivism risk, and no longer was a comparison made between treated and untreated groups.

Williams, Wakeling, and Webster (2007) found that 211 British sex offenders who had followed the Adapted Sex Offender Treatment Program (ASOTP) for lower intellectually functioning sex offenders scored afterwards significantly better on the dynamic items of risk assessment instruments. The program included sexual education, accountability for the offense, victimization, and the learning of coping skills. For evaluating purposes, self-report questionnaires that measured cognitive distortions, attitudes about victims, victim empathy, coping skills, self-esteem, and loneliness were used. Recidivism risk at the beginning of the treatment did not influence changes in problem behaviors, nor did the type of crime. In a follow-up study of 3773 British sex offenders who had completed treatment, Wakeling, Beech, and Freemantle (2013) calculated significant changes of the four domains in the Structured Assessment of Risk and Need (Thornton, 2002): (1) sexual interests; (2) pro-offending attitudes; (3) socio-affective problems; and (4) self-regulation problems. Analyses indicated that those whose scores were in the average range before and after treatment were reconvicted at a significantly lower rate than those whose scores were not in the average range after treatment on selected psychometric scales. Additionally, participants who were deemed to have changed overall on three of the four risk domains were reconvicted at a lower rate than those who were deemed not to have changed on these domains. An overall treatment change status was also computed, but this did not add significantly to the predictive validity of a modified version of an actuarial risk assessment tool (RM2000, Thornton et al., 2003).

Olver, Kingston, Nicholaichuk, and Wong (2014) also found among 392 detained sex offenders that change in self-report questionnaires for cognitive distortions, aggression/hostility, empathy, loneliness, intimacy, and taking responsibility correlated with recidivism risk when controlling for pre-measurements. Noteworthy in this study, however, was that the investigated offenders at the pre-measurement did not score “in a particularly pathological manner on the battery of psychometric measures” (p. 549).

Due to the sensitivity of self-report questionnaires for socially desirable responding, in a number of studies treatment results have been measured using risk assessment instruments. In addition to self-report questionnaires for cognitive distortions, intimacy, and loneliness, in their study of 313 Canadian sex offenders Nunes, Babchishin, and Cortoni (2011) used three risk assessment instruments, including the Stable 2000 (Hanson, Harris, Scott, & Helmus, 2007) for the measurement of psychological risk factors. The treatment consisted of a short version (40–48 h) or an extended version (200–224 h) of the National Sex Offender Program (NaSOP), including self-regulation, cognitive distortions, deviant sexual arousal and fantasy, social skills, anger- and emotion regulation, empathy, and victim-empathy. The short program was given to both community and detained sex offenders, the long program was intended exclusively for detained sex offenders. Nunes et al. (2011) found that investigated sex offenders improved on both self-report questionnaires and risk assessment instruments, but those who followed the short program showed a greater improvement than those who completed the long program. However, the participants in the short program had a lower recidivism risk, measured with the Static-99 (Hanson & Thornton, 1999), than the participants in the long program. Olver, Wong, Nicholaichuk, and Gordon (2007) showed in their study of 321 Canadian sex offenders that positive changes in the psychological items of the Violence Risk Scale-Sexual Offender Version (VRS-SO, Wong, Olver, Nicholaichuk, & Gordon, 2003) were related to a decline in recidivism risk. All psychological risk factors together also contributed to improved predictions of sexual recidivism. In a combined study by Olver, Beggs Christofferson, Grace, and Wong (2014), these findings were confirmed in a larger group of Canadian and New Zealand sex offenders. However, Beggs and Grace (2011) believed that there is stronger evidence of the relationship between treatment outcome and recidivism risk when using different behavioral measurement instruments. Therefore, in their study of 218 sex offenders, they used three different instruments for determining treatment outcome, namely self-report questionnaires, the psychological items of the VRS-SO, and a modified version of the Standard Goal Attainment Scaling (SGAS; Hogue, 1994). The combined score on the three measurements for treatment results correlated significantly positive with recidivism risk. Consequently, the authors concluded that positive treatment results contributed to a decrease in recidivism risk.

4.3 Conclusions and Recommendations

Studies comparing a treated and an untreated sample of sex offenders as well as studies where the differences between a pre- and post-measurement were related to recidivism after discharge‚ resulted in a significant but modest reduction of recidivism. However, these studies have several methodological shortcomings. Firstly, most effect studies do not distinguish any subgroups (e.g., intrafamilial child sexual abusers versus extrafamilial child sexual abusers) with their specific risk factors. Thus, it remains unclear on which of these factors a treatment program should be focused and for which subgroup this should be meant. Another problem is that, just as with programs for violent offenders, the manuals for the programs studied are usually not public, although the interventions used are sometimes briefly described (Olver & Wong, 2013). As a result, it is difficult to determine what exactly the content of the programs was, and which interventions were effective and which were not. Which behavioral change must be achieved in order to decide on discharge or subsequent treatment in a community program remains unnoticed. Besides, it is unclear how the problems of individual participants were analyzed and then addressed in conjunction with the group program. No outcome study, except Friendship et al. (2003), reported on treatment integrity (Hollin, 1995), namely that treatment was conducted by well-trained and supervised therapists who followed the treatment manual (Mann, 2009).

Consequently, Brooks-Gordon, Bilby, and Wells (2006) concluded by an analysis of nine studies that “the uncertainty on the effectiveness of psychological interventions remains” (p. 460). A Cochrane review of ten studies in which included 944 sex offenders also concluded that psychological treatments, both behavioral and psychodynamic, do not automatically result in a decline in recidivism risk (Dennis et al., 2012). In their opinion, the state of affairs is characterized as follows: “Currently available evidence does not support the belief that once the individual has been treated, their risk of offending is reduced” (p. 2). Several studies have found that high-risk sex offenders benefit more from treatment programs than those with average or low recidivism (Beggs & Grace, 2011; Olver et al., 2007; Olver, Nicholaichuk, Kingston, & Wong, 2014; Smid et al., 2014; Wakeling et al., 2013). Beggs and Grace (2011) noted, however, that “higher pretreatment scores leave more room to show genuine improvement” (p. 185).

5 General Conclusions

Treatment programs for offenders are effective when they focus on the dynamic factors that are related to criminal behavior (Andrews & Bonta, 2010). These factors are antisocial personality pattern, antisocial cognition, antisocial associates, family/marital circumstances, school/work, leisure/recreation, and substance abuse. The factor “antisocial personality pattern” refers in most offenders to an antisocial personality disorder (Raine, 1993). An antisocial personality disorder may hardly be cured, but it is an important factor that refers to the ability and the motivation to change behavior. Further, the factors “antisocial associates,” “family/marital circumstances, school/work,” and “leisure/recreation” are broad domains which require several functional analyses for the determination of the various problem behaviors that are playing an important role in such a domain (Chap. 4). Substance abuse is a problem behavior for which we refer to Leukefeld, Gullotta, and Gregrich (2011) or Weekes, Moser, Wheatley, and Matheson (2013).

For the time being, the advice of Mann, Hanson, and Thornton (2010) should be followed regarding the content of cognitive-behavioral programs for sexually violent offenders. In the first place, these programs have to focus on empirically supported risk factors such as deviant sexual interest, offense-supportive attitudes, emotional congruence with children, lack of intimate relationships with adults, impulsiveness, poor problem solving, resistance to control, rancor, and antisocial associates. In the second place, programs should be devoted to promising risk factors such as hostility toward women, antisocial personality pattern, lack of concern for others (callousness), and inadequate coping skills. Finally, programs should not pay attention to factors such as depression, general social skills deficits, poor victim empathy, and lack of motivation.

Several publications stress other non-substantive factors that are related to the effectiveness of treatment programs for offenders, such as the education and experience of the providers and in closed institutions of the staff on the ward, facilities for research on the effectiveness of the program, and finally the support of the management during the implementation and the performance of the program. These conditions are part of treatment integrity (Cooke & Philip, 2001), which are discussed in Chap. 8.