Introduction

Sexually abusive behaviors have been evident throughout history [1, 2] with spectacular cases featuring in many historical accounts. Preeminent among these cases are descriptions of the offenses committed by Baron Gilles de Rais, who sexually molested and murdered countless numbers of children, and by the Marquis de Sade who secretly administered cantharides to prostitutes so he could enact vicious sexual acts on them. However, the first psychiatrically based description of what are now called the paraphilias was the account given by von Krafft-Ebing [3].

More detailed accounts of the history of work with sex offenders have been provided in earlier papers [4,5,6, 7••] so the present description of this history will be somewhat briefer and will focus only on adult male offenders. There are extensive research and treatment histories addressing juvenile sex offenders [8, 9] and sex offenders with cognitive impairments [10, 11], and there are also the beginnings of research and treatment endeavors with female sex offenders [12]. Limited space, however, dictates that this historical review will have to be restricted to adult male sex offenders and will be almost entirely limited to developments by behavioral and cognitive behavioral approaches as they have evolved in the English speaking world. Of course Europeans have a long history of attending to sex offenders and their efforts continue to flourish [13] but attention to this work will also have to be set aside here.

This focus on cognitive behavioral approaches is not meant to dismiss other ways of viewing these offenders. Indeed, early non-behavioral reports [14, 15] identified many of the problems and deficiencies of these perpetrators that inspired some of aspects of the initial foci of the behavioral movement. In particular, the extensive detailed description of the features of sex offenders that distinguished them from non-offenders contained in the report of Gebhard, Gagnon, Pomeroy, and Christensen [16] provided the foundation upon which much of the later theorizing, research, and treatment was generated by cognitive behaviorists. Also, several early non-behavioral treatment programs reported initial successful outcomes [17, 18]. Additionally, the excellent work derived from biological and medical perspectives [19,20,21] has significantly advanced our understanding of these problematic behaviors. Unfortunately, space restrictions and the limitations of the present author’s medical sophistication dictate that consideration of this literature be set aside.

Although theories concerning the origin of deviant sexual behaviors only fully flourished after research and treatment had made progress, this account will begin with an examination of the development of explanations of the origins of these behaviors, followed by a description of the emergence of research endeavors, with the generation of treatment programs occupying the final segment. In some sense this is a bit of a reverse of the order of the most important developments in work with sex offenders, but it is a fact that the early behavioral theories drove the initial applications of treatment, and research provided the capacity to evaluate the effectiveness of early interventions.

The Generation of Theory

The earliest attempts to explain the origins of deviant sexuality were outlined in the late nineteenth century by Binet [22], who is better remembered for initiating the assessment of intelligence, and by Norman [23] and von Krafft-Ebing [3]. While these early accounts evoke comparisons with more modern explanations, it was not until the 1960s that theory began to directly influence practice.

McGuire, Carlisle, and Young [24] proposed that deviant sexual interests drove deviant sexual behaviors and that these unusual interests were acquired by classical conditioning experiences. They offered as an example, the hypothetical experience of a young man stopping by the wayside to urinate, thinking he was safe from observation. The unexpected arrival of an attractive woman was said to have aroused sexual feelings in the man who thereafter evoked images of this scene during masturbation, thereby entrenching a penchant for exposing his erect penis to passing females. Of course this explanation was entirely speculative but it fitted with the emerging zeitgeist of the time in British therapeutic practice and research with sex offenders. As a result, McGuire et al.’s theory became widely popular and set the foundation for the subsequent restricted focus on conditioning-based treatments and assessments for men with unusual sexual interests.

This initial theory, and its later modification by Laws and Marshall [25], was eventually called into question when research revealed that conditioning processes do not appear to underlie the acquisition of deviant interests [26, 27]. In fact well before these challenges had been raised, it was assumed that conditioning-based treatment could not serve as an effective “stand-alone” intervention [28, 29]. This is not to say that conditioning processes are now considered to be entirely absent from the development of deviant interests, it is just that they are thought to play a more minor role than was previously accepted [30].

The first more complex theory was not so much an explanation of the origin of deviant sexual behavior, but rather an account of what factors needed to be operative for a man to commit a sexual offense. Finkelhor [31] proposed that four “preconditions” had to be present for a man to sexually molest a child. He must (1) view the idea of sex with a child as emotionally satisfying; (2) be sexually aroused by the child; (3) be unable to meet his sexual needs appropriately; and (4) somehow overcome his inhibitions about sexually abusing a child. This theory had a significant influence among treatment providers but this early impact seems to have faded over time.

In an attempt to integrate biological dispositions, childhood experiences, exposure to the influences of the media, the impact of disruptions during adolescence and early adulthood, and the presence of opportunities to offend, Marshall and Barbaree [32] outlined what they called an “integrated theory.” While this theory appears to have had an influence among theorists, it has been criticized for being overly ambitious by attempting to explain all types of sex offenders, and for failing to specify its treatment implications [33]. Acknowledging these valid criticisms, Marshall and Marshall [34] narrowed their focus to the influence of attachment disruptions among child molesters across the full spectrum of developmental stages (i.e., childhood, adolescence, early adulthood, middle age, and old age) since in at least some child molesters, these are the ages of their first offense. This latter version of their theory was not simply restricted to child molesters but was even more limited to just those offenders who had molested children known to them who Marshall, Smallbone, and Marshall [35] had called “affiliative child molesters.” Affiliative child molesters, Marshall et al. said, included family members, teachers, clergy, care workers, sports coaches, scout leaders, and other men who had continuing relationships with, and ready access to, children. These types of child molesters account for approximately 90% of all men who sexually abuse children [36].

In terms of the breadth of its impact, the most influential theory in recent years has been Ward’s [37] Good Lives Model (GLM). Consistent with Maslow’s [38] earlier notion, Ward suggested that all humans strive for fulfillment even if they are not always aware of this. Ward identified 11 areas of functioning in which he claimed humans seek to attain competence. This theory directly suggests that therapists dealing with all types of offenders should adopt a focus on developing the strengths of their clients in these 11 areas of functioning, rather than attempting to simply decrease the offenders’ deficits. As a consequence, Ward’s proposal led to the development of a variety of similar strength-based approaches to treating sex offenders [39]. In addition to the GLM’s implications for treatment, Ward and Gannon [40] spelled out its relevance for etiology and for the management of sex offenders.

In addition to the conditioning theory mentioned earlier, there have been other theories that have attempted to account for specific treatment relevant aspects of sex offenders. These proposals have covered the origin of distorted cognitions [41, 42], underlying schemas about women and children [43], problems in empathy for others [44], low self-esteem [45], and notions about the offenders’ inability to regulate their emotions [46]. In addition, Smallbone and Cale [47] have developed a theory about the situational cues that trigger sexual abuse, and Harris [48] and Polaschek [49••] have outlined accounts of the processes that lead to desistance or relapses after discharge from institutional settings.

Each of the comprehensive models discussed here, and the single factor theories, has led to important developments in research and treatment.

The Emergence of Research

Very little research focused on sex offenders until the 1950s and 1960s, when more experimentally based views began to emerge [50]. In the treatment of sexually anomalous behaviors, these behaviorally based approaches initially focused on the elimination of deviant sexual interests that were thought to underpin paraphilic behaviors. As a result, one aspect of early research was devoted to the development of methods to measure sexual interests. Freund [51] described a device that enclosed the penis and was, thereby, able to detect changes in penile volume as the subject became aroused. These changes were taken to indicate degrees of sexual interest in whatever stimuli were being presented to the subject. Later, rather more simple devices were developed that assessed changes in the circumference of the penis in response to sexually provocative stimuli [52, 53]. Both volumetric and circumferential devices remain the accepted standard for detecting sexual interests [54] and for providing a basis for diagnosing the paraphilas [55, 56].

These measuring instruments are commonly described as either “penile plethysmography” or “phallometry.” These devices have also served as a way to assess changes in deviant interests resulting from the effects of treatment which was also one of their original purposes. However, questions began to be raised about these assessment methods. O’Donohue and Letourneau [57] questioned the psychometric properties of phallometry and its lack of standardization across different sites. Marshall and Fernandez [58] noted the unsatisfactory reliability over repeated assessments of these measures and Bailey [59] raised concerns about measurement error. As a result, more recent research has examined alternative “indirect” strategies to assess sexual interests. These measures involve approaches designed to disguise the intent of the tests so that offenders will be less likely (or able) to fake normal interests. Evidence for the value of these recent alternatives is limited with the emotional Stroop test having the strongest empirical support [60].

The impetus for research into the skills necessary to form effective intimate relationships was triggered by Marshall’s [61] proposal that intimacy deficits are the bases that drive sexual offending. This proposal generated extensive research with much of the findings having been recently summarized by Beech and Mitchell [62]. In an interesting variation on the usual approach to social and intimacy deficits, Keenan and Ward [63] suggested that such deficiencies among sex offenders might be partly explained by their failure to have developed the skills involved in what is known as “theory of mind.” This notion initially emerged from research on child development but has subsequently been applied to the analysis of a broad range of problems. It essentially involves the capacity, or lack thereof, to infer the intentions, thoughts, and emotions of others and to adjust one’s actions accordingly [64]. Recent research has shown that indeed sex offenders do have significant deficits in theory of mind skills [65].

Oddly, very little attention has been paid to the quite obvious deficits among sex offenders in the sexual skills and sexual knowledge necessary to effective functioning in affectionate adult intimate relations. However, the relevance of research derived from the general literature on healthy human sexual functioning has recently been detailed by Marshall, Hall, and Wo [66]. Hopefully Marshall et al.’s outline will prompt researchers to address these issues with sex offenders.

Various other specific topics that are typically addressed in treatment programs for sex offenders have received some research attention. These include empathy for others, particularly for the victims [67]; low self-esteem [68]; sexual preoccupation [69]; and deficits in coping skills [70]. Recently, researchers have also begun to examine features of child pornographers and internet offenders [71].

An influential body of research emerged in the 1990s and early 2000s. This research was aimed at identifying factors that predict sexual offending. This program of study was prompted by the findings of Andrews [72, 73] indicating that effective treatment programs must contain specific elements. Treatment benefits were evident only when (1) most resources were allocated to the highest risk offenders; (2) empirically established, potentially modifiable risk factors were targeted; and (3) therapists adopted an approach that had been shown to facilitate the formation of an effective therapeutic alliance. These three principles of effective programming were described, respectively, as risk, needs and responsivity. Recently, Hanson, Bourgon, Helmus, and Hodgson [74] demonstrated that these principles also apply to the treatment of sex offenders. Marshall and his colleagues [75] showed that therapists working with sex offenders produced the desired changes only when they displayed empathy and warmth, were rewarding of progress, and offered suggestions when clients were struggling. These features of effective therapists are a match for Andrews’ findings concerning his responsivity principle.

Andrews’ early reports launched the need in the sex offender field to identify risk factors pertinent to these clients. Two sets of factors were needed: static risk factors and dynamic (also called “criminogenic”) risks. Static risks refer to unchangeable features derived from the client’s history (e.g., extent of offense history, diversity of offenses, types and number of victims, gender of victims, the offenders’ age) while dynamic risks concern problematic features that are nevertheless open to the possibility of change (e.g., deviant sexual interests, deficits in social and intimacy skills, poor emotional, sexual and behavioral regulation, offense-facilitating beliefs). Several researchers were able to identify static risk factors relevant to sex offenders [76, 77] while other studies generated the empirical bases for the criminogenic needs of sex offenders [78].

One area of research that has important practical implications, concerns the evaluation of the effectiveness of treatment for sex offenders. A comprehensive meta-analytic study by Hanson et al. [79] included outcome data on 43 treatment programs. This report showed that among untreated sex offenders 16.8% reoffended over an extended post-release follow-up while only 12.3% of the treated men recidivated. More extensive meta-analyses were reported by Lösel and Schmucker [80, 81, 82••] who found results similar to those reported by Hanson et al. [79].

As will be seen in the next section on treatment, there has been a recent shift away from a deficit- focused treatment approach to one that emphasizes the sex offenders’ strengths. This shift was inspired by Ward’s [37] Good Lives Model. However, the only appraisal of a strength-based treatment program with sex offenders to date appears in the report by Olver, Marshall, Marshall, and Nicholaichuk [83]. This study examined the long-term outcome of Marshall, Marshall, Serran, and O’Brien’s [84] strength-based sex offender treatment program which also complies with Andrews’ principles of effective offender treatment. Of the 579 sex offenders treated in this program during the period 1991 to 2001, only 4.2% reoffended during the 8-year follow-up period. This recidivism rate was significantly lower than was evident among a matched group of untreated men (20.2%) and lower than that shown in a matched group treated by Correctional Service of Canada’s (CSC) standard program (10.7%). These differences in effects were even more pronounced when only high risk offenders were examined. However, it is important to note that the standard CSC program, which was also based Andrews’ RNR principles, produced significant reductions in rates of recidivism when compared to the untreated men.

Thus, not only do Olver et al.’s findings encourage optimism about the potential benefits of a strength-based approach, they also offer support for the more general idea that treatment with sex offenders can be effective if programs conform to Andrews’ principles. Finally, when sex offenders are effectively treated, the results not only serve to protect the public, it has also been shown that they save considerable costs associated with the investigation, prosecution, and incarceration of repeat offenders [85].

The Emergence and Development of Treatment

Initial treatments addressing paraphilic behaviors were generated in the late 1800s. Both Charcot and Magnan [86] and von Schrenck-Notzing [87] outlined methods for resolving unusual sexual interests, and the first approach to successfully overcome an attraction to children was described in a single case report by Moll [88]. However, little seems to have been done after these early attempts until a series of single case reports emerged in the 1950s [89], mostly conducted by staff at the Institute of Psychiatry of the University of London. Psychiatrists at this Institute reported the effectiveness of their conditioning-based treatment of several types of paraphilias [90,91,92]. These Pavolvian approaches remained the primary focus of treatment as programs began to be developed in North America [93, 94], although it was not long before the limits to these treatments became apparent. Marshall [95], for example, pointed to the need to develop procedures to both enhance normative sexual interests and to develop the social skills necessary to act on this changed orientation. Marquis [96] provided a strategy to enhance prosocial arousal and Barlow [97] outlined the necessary elements of social skills training.

Without any clear empirical basis, approaches in the 1970s began to broaden the issues addressed in treatment [28, 29]. Social ineptitude was assumed to be an essential treatment target, as were distorted ideas about women’s desires and about children’s interests in sex. It was also assumed that sex offenders lacked empathy toward others. These changes in treatment remained entrenched throughout the 1980s and early 1990s, with other elements being progressively added [98] again in the absence of supporting evidence.

In the early 1980s, Marques [99] proposed framing sex offender treatment around Marlatt’s [100] relapse prevention (RP) approach that he had developed in his work with drug and alcohol problems. RP became instantly popular, although again in the absence of relevant evidence, and soon spread across North America as the basis for the treatment of sex offenders [101]. Later criticisms of RP as applied to sex offender treatment [102], and the disappointing results of a carefully designed evaluation [103], led to a call to markedly curtail the deployment of RP [104].

As was shown in the “The Emergence of Research” section, the 1990s saw the beginnings of the development of risk assessment instruments that eventually identified both static and dynamic risks, thereby providing the empirical bases for distributing resources according to static risk levels and narrowing the focus of treatment to established criminogenic factors (i.e., dynamic risks). Treatment for sex offenders could now be based on firm empirical foundations. Unfortunately, not all treatment programs have followed these empirical principles as is evident by an examination of the most recent Safer Society Survey of the majority of North American programs [105]. This survey revealed that only 3.1% of community-based programs, and just 10.1% of institutional programs, identified Andrews’ RNR model as their primary guide. These observations are indeed disappointing.

There seems little doubt that the most influential model in sex offender treatment in recent years has been Ward’s [37] Good Lives Model (GLM). This model emerged from the “positive psychology” movement and like that approach, the GLM emphasized focusing on developing sex offenders’ strengths with the idea being that the deficits that drove them to offend would, as a result of developing competing skills, simply disappear. The GLM identified 11 specific areas of human striving that it declared should be the targets of treatment. This model clearly appealed to many treatment providers as indicated by the fact that a significant number of North American sex offender programs began to describe their approach as modeled on the GLM. However, when the actual practice implementations of these programs were examined in detail, only one was found to approximately conform to GLM practices [106]. On the other hand, the GLM did inspire a proliferation of other strength-based programs that were either an adaptation of Ward’s model or a limited integration of some aspects of the model.

As was shown in the “The Emergence of Research” section, at least one of these strength-based approaches [84] has been shown to be effective and has produced greater reductions in reoffending than was evident for a standard deficit-focused program. However, it is important to note that even the standard programs are actually effective as was evident in the “The Emergence of Research” section. Thus, from the available evidence, it appears that when sex offender treatment is based on the principles of risk, need, and responsivity, and when the focus on criminogenic needs is aimed at building strengths rather than directed at eliminating deficits, then treatment can produce significant reductions in subsequent re-offense rates [83]. This observation of the effectiveness of a strength-based treatment bodes well for the future of psychological treatments for adult sex offenders.

Conclusions

This paper has provided an overview of the developments in psychological theory, research and treatment of adult sex offenders as these developments have unfolded particularly since the early 1960s. Developments in these three areas have, for the convenience of this review, been addressed separately although in the course of the history covered here, each area has fed thinking and adjustments in each other domain.

Theories developed over time to become more complex but also more restricted in their focus. Beginning with Pavlovian-derived conditioning notions, theories of the origin of sexual offending began to recognize a wide range of perspectives and to include many other factors than simple conditioning. General theories addressed the origins of sexual offending behaviors while specific theories have focused on each aspect of the offending behaviors.

Research has both followed and chased the development of theories and has expanded to examine a broad range of issues thought to drive sexual offending behaviors. Perhaps the two most important foci of research have been the empirical development of risk assessment instruments and reports on the evaluation of treatment. The outcome of research has provided the bases for allocating treatment resources where the greatest benefits might be derived, for selecting the appropriate targets to be addressed in treatment, and has established the appropriate way for therapists to deliver treatment. The results of a series of meta-analytic evaluations of reports of treatment outcome have encouraged confidence in the capacity of well-designed programs to produce significant reductions in rates of reoffending among sex offenders.

Treatment programs have evolved from simply targeting deviant sexual interests to progressively more complex approaches. Ward’s Good Lives Model has influenced a move away from an exclusive focus on targeting sex offenders’ deficits, to give more pronounced attention to building the strengths of these offenders in order to offset deficits. The implementation of strength-based treatment appears to offer greater promise for improving effectiveness than the traditional exclusive attention to the clients’ deficits.

Hopefully, this historical review will encourage further developments in theory, research, and treatment with sex offenders. There is still much to be done to improve our understanding of these difficult clients and to make treatment more effective so that fewer of these men will continue to harm innocent women and children after their release from custody.