part0024

The first question that needs to be asked in trying to explain sexual offending is whether a special theory is required—after all, sexual offending is a serious crime and might be sufficiently explained by more general theories of criminal behavior. As mentioned in the previous chapter , evidence supports this generalist idea: Many sex offenders have also committed nonsexual offenses; the onset of sexual offending typically occurs after the onset of nonsexual offending; and identified sex offenders are more likely to commit new nonsexual offenses than sexual offenses when followed over time (Caldwell, 2002; Hanson & Bussière, 1998). Moreover, as I discuss in further detail in Chapter 7 , many important risk factors for sexual recidivism are general criminal risk factors, such as offender age, prior criminal history, and antisocial personality traits (Hanson & Bussière, 1998; Hanson & Morton-Bourgon, 2005). These same factors also predict recidivism among other offenders, including general criminals (Gendreau, Little, & Goggin 1996), offenders with mental disorders (Bonta et al., 2014), and youth with a history of juvenile delinquency (Lipsey & Derzon, 1998). Last, as discussed in more depth in Chapter 8 , interventions designed to address general criminogenic needs, such as emotion regulation, problem solving, and substance use, can reduce recidivism among sex offenders (Borduin, Schaeffer, & Heiblum 2009; Letourneau et al., 2009).

This converging evidence does suggest that many sexual offenses against children can be explained by general theories of crime (Gottfredson & Hirschi, 1990; Loeber & Farrington, 1997; Moffitt, 1993). Gottfredson and Hirschi’s (1990) low self-control theory of crime suggests that sexual offending is expected when a motivation to seek sexual gratification is combined with low self-control and opportunity. However, the general theory assumes conventional motivations (e.g., sexual assault committed because the perpetrator desires sex, assault because the perpetrator wants to hurt the target, stealing because the perpetrator wants the property) and does not specifically account for atypical motivations such as pedophilia or hebephilia. Indeed, a welcome influx of researchers has studied sexual offending from different criminological perspectives, including developmental life course (criminal career) approaches, routine activities approaches, and situational crime prevention (Lussier & Blokland, 2014).

Much less evidence has been found on factors that explain the onset versus persistence of sexual offending, as is evident when comparing the studies mentioned in this chapter versus those discussed in Chapter 7 . Given this gap, a valuable study was conducted by Duwe (2012), who developed an actuarial tool to predict first-time sexual offending (defined as conviction for sexual offense within 4 years of opportunity after release from prison) in a large sample of 9,064 general offenders in Minnesota who had no prior history of sexual offending and who were released from prison between 2003 and 2006. The significant predictors were all indicators of general criminality: Any juvenile convictions, recent minor (in severity) assaults, robberies committed under age 21, first-degree burglaries, false information to police, suicidal history, and participation in prison-based chemical dependence treatment. These indicators suggest that a pattern of antisocial and criminal behavior can predict the onset of sexual offending among already identified general offenders. It would be fascinating to also have measures of atypical sexual interests to see whether they perform as expected, but such data were not available to Duwe because the general offenders had not committed any sexual offenses and thus they would not have been expected to undergo assessment of atypical sexual interests. Other factors that were considered and that were not selected in Duwe’s analysis included supervision failures, other mental health indicators (e.g., having a mood disorder, having any mental disorder diagnosis), other criminal history variables (e.g., total number of convictions, drug offense convictions), and demographic characteristics (age, education). However, scores on a general measure of criminogenic risk and needs did not predict the onset of sexual offending.

A recently published study by Babchishin et al. (2017) is germane to this discussion. We found that a set of theoretically or empirically identified parental and birth factors predicted onset of sexual offending in a large Swedish male cohort, including young parent age, parent violent offending history, and parent psychiatric history, as well as small size for gestational age and small head circumference; however, many of these factors also predicted the onset of nonsexually violent offending, suggesting none of these early risk factors were unique to sexual offending.

Nonetheless, evidence supports the idea that some sexual offenses, and some sexual offenders, are unique and that both common and specific factors explain sexual offending against children. Seto and Lalumière (2010) looked at what distinguishes adolescent sex offenders from other adolescent offenders. No comparable meta-analysis has been conducted for adult sex offenders, a much larger research literature, but systematic and narrative reviews consistently identify atypical sexual interests as the most distinctive factor (Lalumière, Harris, Quinsey, & Rice, 2005; Laws & O’Donohue, 2008; Seto, 2008, 2013). In their systematic review, Whitaker et al. (2008) examined studies comparing offenders against children with offenders against adults, non–sex offenders, or nonoffenders. Compared with non–sex offenders, sex offenders against children were more likely to have atypical sexual interests, high sex drive, sexualized coping, and offense-supportive cognitions. They were also more likely to have interpersonal problems, such as social skill deficits, loneliness, anxiety, and difficulty with intimate relationships. Sex offenders against children were also less likely to have externalizing problems. The largest differences were found for sexuality variables.

In the next section, I summarize the major models of sexual offending against children: Finkelhor’s (1984) four-factor model; W. L. Marshall and Barbaree’s (1990) integrated theory; Hall and Hirschman’s (1992) quadripartite model; Ward and Siegert’s (2002) pathways model; Ward and Beech’s (2006) integrated theory of sexual offending; and Stinson, Sales, and Becker’s (2008) multimodal self-regulation theory. I am particularly keen to (a) highlight similarities and differences that can serve as a launching point for developing a comprehensive model to explain sexual offending against children and (b) identify research studies that can test them. I also discuss a developmental life course perspective from criminology that recognizes sexual offending typically occurs after the onset of nonsexual offending and that many sex offenders are generalists rather than specialists. Any comprehensive explanation of sexual offending needs to include factors that have been identified in general criminology research, as well as the etiology and effects of atypical sexual interests such as paraphilias or hypersexuality. Theories or models developed specifically to explain sexual offending against adults are not reviewed here; readers are instead encouraged to read Knight and Sims-Knight (2003), Lalumière et al. (2005), and Malamuth (2003), as well as the book-length reviews of these theories in Ward, Polaschek, and Beech (2006) and in Stinson et al. (2008).

David Finkelhor is a tremendously important scholar in the scientific study of sexual offending against children, and he was one of the first to describe a multifactorial explanation of this form of offending (Finkelhor, 1984). His preconditions model has three factors reflecting motivations to engage in this behavior—sexual arousal to children (pedophilia), having more affinity for children than for adults (emotional congruence with children), and feeling unable to meet one’s emotional and sexual needs in adult relationships (blockage)—and a fourth factor representing the overcoming of inhibitions to commit sexual offenses (disinhibition), which can include trait factors such as personality (impulsivity, recklessness) and state factors such as intoxication. Finkelhor (1984) further noted that four conditions had to be met for sexual offenses to be committed against children: (a) motivation to commit the offense; (b) overcoming internal inhibitions, such as fear of detection or arrest; (c) overcoming external inhibitions, such as the presence of other people; and (d) overcoming any resistance by the child.

Several years later, W. L. Marshall and Barbaree (1990) described a developmentally informed model that integrated ideas about biological vulnerabilities, such as perturbations in male-typical tendencies toward aggression and engaging in sexual behavior; adverse childhood experiences, such as abuse or neglect; and the development of self-regulation problems and other deficits that increased the likelihood of sexual offending against children. W. L. Marshall and Barbaree emphasized a role for social skills deficits, where some individuals are unable to develop appropriate relationships with peers and thus turn instead to children for their sexual and emotional needs. In their view, any positive sexual experiences with children could fuel fantasies about sex with children that could be reinforced through masturbation and further sexual behavior involving children. W. L. Marshall and Fernandez (2000) later acknowledged a role for pedophilia, mostly for offenders with multiple, unrelated child victims. The view seemed to be that most offenses against related children, and many offenses against girls, reflected nonpedophilic motivations.

Around the same time as W. L. Marshall and Barbaree (1990), Hall and Hirschman (1992) modified their four-factor model of sexual offending against adults to specifically explain sexual offending against children. This model has some overlap with Finkelhor’s (1984) model, with the four factors in Hall and Hirschman being sexual arousal to children, cognitions that can justify sexual offending against children, affective dysregulation, and personality problems. The first factor also relates to pedophilia; the second overlaps with emotional congruence with children; and the last two factors are related to disinhibition. Hall and Hirschman suggested that these factors could operate singly or in combination but that particular factors would be more important for different groups of offenders. Thus, Hall and Hirschman were also describing a typology of sexual offenders against children, including a pedophilic offender group that was expected to have more child victims and unlikely to have nonsexual offenses, and an antisocial group high on affective dysregulation that would be more opportunistic, more violent, and more likely to commit nonsexual offenses. Individuals high on offense-supportive cognitions were thought to be more likely to plan their offenses and to select related child victims. The personality problem factor was thought to be a general vulnerability that increased the likelihood of sexual arousal to children, offense-supportive cognitions, or affective dysregulation.

Ward and Siegert (2002) tried to integrate what they considered to be the best elements of each of the earlier models in explaining sexual offending against children. They again identified four clusters of problems that increase the likelihood of sexual offending: distorted sexual scripts, which can include sexual fantasies and sexual arousal involving children; cognitive distortions; intimacy and social deficits; and emotional dysregulation. Notice the overlap with the factors identified by Finkelhor (1984) and by Hall and Hirschman (1992). Ward and Siegert suggested different pathways led to sexual offending, each distinguished by a particular cluster of problems; a fifth pathway represented sexual offending by antisocial individuals who have committed a variety of crimes.

Ward and Beech (2006) expanded on the Ward and Siegert (2002) pathways model with an integrated theory of sexual offending, a broad model that encompasses evolutionary selection pressures, neurobiology, and learning. They also incorporated psychological evidence regarding the systems involved in emotion, social functioning, and sexual arousal. Ward and Beech (2016) updated their discussion of this integrated theory, noting that evolutionary selection pressures included male preferences for casual sex and propensities to use violence, and explaining how their theory could cross different levels of explanations.

The core of Stinson et al.’s (2008) model is that biological vulnerabilities can combine with early adversities to produce maladaptive self-regulation strategies. This in turn raises the risk of antisocial behavior in an attempt to regulate self, including substance use and problematic sexual behavior such as excessive pornography use. Sex can be a powerful reinforcer and can entrench sexual behavior through sexual fantasy, masturbation, and orgasm. Drawing on ideas proposed by Kafka (1991, 2003) and Kafka and Hennen (2003), Stinson et al. suggested that, for sexual offenders, paraphilia-like behavior can develop rather than (or as well as) compulsive or excessive conventional behavior, such as use of pornography as a result of accidental pairings of sexual arousal with stimuli (e.g., during childhood sexual experimentation) or inaccurate misattribution of spontaneous erections or arousal as a result of strong emotions, such as anger. Stinson et al. (2008) recognized that many children might have early sexual experiences with peers and not develop paraphilia-like sexual interests or behavior. Other factors must be in play as well, which might include personal vulnerabilities; early adversity; and situational factors, such as a lack of appropriate sexual outlets or lack of consequences to early problematic behavior.

Even from these brief descriptions, it is clear these models have many similarities. All them recognize a role for sexual arousal to children, although they differ in how important this motivation is and how it might develop. They all recognize that cognitive and social deficits can play a role in sexual offending against children, and all of them recognize that both internal (disinhibition) and external (opportunity) obstacles to offending must be overcome for offending to occur. Although not necessarily stated, all these models were developed to explain sexual offending against children by men; their generalizability to explaining sexual offending by women or by adolescents is unspecified. However, these models also have some notable differences. They differ, for instance, in the weight given to agreed-upon factors; some suggest a single factor is sufficient (G. C. N. Hall & Hirschman, 1992), whereas others suggest factors can combine to increase or decrease risk (Finkelhor, 1984; W. L. Marshall & Barbaree, 1990; Ward & Siegert, 2002). Last, several models—W. L. Marshall and Barbaree (1990), Ward and Siegert (2002), and Ward and Beech (2006)—introduced developmental and biological elements.

Ward et al.’s (2006) and Stinson et al.’s (2008) critical review of these models found that all have limitations in terms of their unifying power, internal consistency, predictive validity, heuristic value, falsifiability, or parsimony. A particular concern—for me—is how to evaluate and compare these theories, because many concepts are not clearly operationalized. For example, Ward and Beech (2006) attempted to be consilient and comprehensive in their integrated theory of sexual offending, but the breadth of the theory is also a limitation because it is vague about how specific constructs, processes, and mechanisms operate and interact to influence sexual offending.

A general criticism of all these models is that they do not pay sufficient attention to important theoretical and empirical advances in developmental and general criminology. This reflects a tendency to view sexual offending as unique and, therefore, sexual offenders as a unique population. Yet many sex offenders have also committed nonsexual offenses, some go on to commit nonsexual offenses, and some nonsexual offenders go on to commit sexual offenses. As I discuss later in Chapter 7 , many general crime risk factors identified in the forensic and criminological literatures are also potent predictors of sexual recidivism among identified sex offenders (e.g., offender age, general criminal history).

A limitation of the field rather than the theories themselves is that none of them have been tested in their entirety, using large samples of sex offenders; reliable and valid measures; and structural modeling, clustering, or other techniques to examine whether concepts are indeed related in the way the models presuppose. Instead, aspects of models are tested, for example through group comparisons showing differences on one or more of the candidate explanatory factors. It would be particularly helpful in advancing the field to pit theories against each other.

The motivation–facilitation model (see Figure 4.1 ) is strongly influenced by Finkelhor’s (1984) preconditions model and by Gottfredson and Hirschi’s (1990) self-control theory of crime, which suggests low self-control facilitates acting on normative but antisocial impulses (e.g., taking things one wants even though they belong to others, hitting someone when angry, using illegal drugs). It is also related to other models, particularly Stinson et al.’s (2008) emphasis on the role of ineffective or maladaptive self-regulation, which can be seen in the impact of facilitation factors. Similar to other models, the motivation–facilitation model draws on multiple lines of research, as summarized in Seto (2008, 2013) and in Pullman, Stephens, and Seto (2016). It differs from Finkelhor’s (1984) model by describing antecedents for pedophilia, including hebephilia, and by considering other sexual motivations for sexual offending, such as hypersexuality. The motivation–facilitation model differs from low self-control theory because it specifically considers atypical motivations for sexual offending (paraphilias, hypersexuality, intense mating effort) and their origins, whereas low self-control theory is agnostic about the nature of motivations for sexual offending. The motivation–facilitation model was later expanded to explain online child pornography and sexual solicitation offending (Seto, 2013). In principle, however, the model could apply to other forms of sexual offending, including sexual assaults of women by men; noncontact offending involving exhibitionism or voyeurism; and even less common crimes, such as bestiality or other illegal pornography use.