Chapter 3: Sex Offender Typologies

Introduction

Sexual violence remains a serious social problem with devastating consequences. However, resource scarcity within the criminal justice system continues to impede the battle against sexual violence. The challenge of "making society safer" not only includes the need for resources, but also requires a comprehensive understanding of accurate offense patterns and risk. (For a discussion of adult "Sex Offender Risk Assessment," see Chapter 6 in the Adult section.) This knowledge may be used to devise offense typologies, or classification systems, that will inform decisions regarding investigation, sentencing, treatment and supervision. (For more on "Effectiveness of Treatment for Adult Sex Offenders," see Chapter 7 in the Adult section.)

Although other typologies exist, this chapter only includes the classification systems that have been empirically derived and validated. Two empirically validated typologies — Massachusetts Treatment Center: Child Molester Version 3 (MTC: CM3) and Rapist Version 3 (MTC: R3) (Knight & Prentky, 1990) — were not included because some researchers (e.g., Barbaree et al., 1994; Camilleri & Quinsey, 2008; Hudson & Ward, 1997) have questioned their clinical utility.¹

The crossover offending section encompasses more than 25 years of research using different methodologies and populations. Although not considered a classification system due to the dynamic nature of the offense pathways, the self-regulation model (SRM) was reviewed due to its clinical utility and relationship to risk. SRM has been validated using several offender populations and methodologies. Due to the limited scope of this chapter, this review focuses on adult sexual offenders, although some juvenile studies are included, where relevant. (For a discussion of "Etiology and Typologies of Juveniles Who Have Committed Sexual Offenses," see Chapter 2 in the Juvenile section.)

Summary of Research Findings

Traditional Typologies

The majority of theories regarding sexual deviance postulate that sexual offenders specialize in types of victims and/or offenses (Simon, 1997). Researchers have developed specific classification-unique offender characteristics (Knight & Prentky, 1990; Simon et al., 1992). Most of these typologies imply that victimization (i.e., who is a potential victim) is linked to the specific type of sexual offender (e.g., rapists sexually assault adults/peers, child sexual abusers sexually assault children).

Traditional typologies have been developed to provide a comprehensive understanding of deviant sexual behaviors required for treatment intervention and effective supervision. However, classifying sexual offenders has been shown to be problematic. Sexual offenders exhibit heterogeneous characteristics, yet they present with similar clinical problems or criminogenic needs (e.g., emotional regulation deficits, social difficulties, offense supportive beliefs, empathy deficits and deviant arousal); the degree to which these clinical issues are evident varies among individual offenders (Gannon, Terriere & Leader, 2012; Ward & Gannon, 2006). Indeed, this heterogeneity challenges effective risk management and treatment of offenders (Martinez-Catena, Redondo, Frerich & Beech, 2016). Overall, traditional typologies have demonstrated considerable problems, as indicated by inadequate definitions and inconsistent research findings. In addition, most of the typologies have failed to address treatment issues and to predict recidivism (Camilleri & Quincy, 2008; Knight & Prentky, 1990). (For information on "Adult Sex Offender Recidivism," see Chapter 5 in the Adult section.) This section reviews the most frequently used and empirically tested sex offender typologies for child sexual abusers, rapists, female offenders and internet sexual offenders.

Child Sexual Abusers

Finkelhor (1984) provides the most comprehensive definition of child sexual abuse: Child sexual abuse is the use of force/coercion of a sexual nature either when the victim is younger than age 13 and the age difference between the victim and the perpetrator is at least five years, or when the victim is between 13 and 16 and the age difference between the victim and perpetrator is at least 10 years. In this definition, coercion does not necessarily imply a direct threat. Child sexual abusers often develop a relationship with a child to manipulate him or her into compliance with the sexual act, which is perhaps the most damaging component of child sexual abuse (John Jay College, 2004). Indeed, a defining feature of child sexual abuse is the offender's perception that the sexual relationship is mutual and acceptable (Groth, 1983).

Differences Between Child Sexual Abusers and Rapists

Child sexual abusers have been difficult to classify as they vary in economic status, gender, marital status, ethnicity and sexual orientation. Child sexual abusers are often characterized as exhibiting poor social skills, having feelings of inadequacy or loneliness, having greater sexual problems or being passive in relationships (Cortoni & Marshall, 2001; Groth, 1979; Maniglio, 2012; Marshall, 1993; Whitaker et al., 2008). They differ from rapists with respect to thought processes and affect, and often describe their offending behaviors as uncontrollable, stable and internal; whereas rapists attribute their offenses to external, unstable and controllable causes (Garlick, Marshall & Thorton, 1996). Indeed, Whitaker et al.'s (2008) meta-analytic review of 89 studies indicates child sexual abusers have fewer externalizing behaviors compared to rapists. Child sexual abusers display deficits in information-processing skills and maintain cognitive distortions to deny the impact of their offenses (e.g., having sex with a child is normative; Hayashino, Wurtele & Klebe, 1995; Whitaker et al., 2008). In contrast, rapists display distorted perceptions of women and sex roles, and often blame the victim for their offense (O Ciardha, 2011; Polaschek, Ward & Hudson, 1997). With respect to affect, child sexual abusers assault to alleviate anxiety, loneliness and depression. Rapists typically assault as a result of anger, hostility and vindictiveness (Polaschek, Ward & Hudson, 1997). Many of these characteristics have been incorporated into the typologies of rapists and child sexual abusers (Camilleri & Quinsey, 2008; Groth, 1979; Knight & Prentky, 1990).

Pedophilic and Nonpedophilic Distinction

The most important distinction among child sexual abusers is whether they are pedophilic or nonpedophilic, because pedophilia has been shown to be a strong predictor of sexual recidivism (Hanson & Bussiere, 1998). Not all individuals who sexually assault children are pedophiles. Pedophilia consists of a sexual preference for children that may or may not lead to child sexual abuse (e.g., viewing child pornography), whereas child sexual abuse involves sexual contact with a child that may or may not be due to pedophilia (Camilleri & Quinsey, 2008). According to the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (American Psychiatric Association, 2013), a diagnosis of pedophilia requires an individual to have recurrent, intense and sexually arousing fantasies, urges or behaviors directed toward a prepubescent child (generally 13 years of age or younger) over a period of at least six months; to have acted on these urges or to be distressed by them; and to be at least 16 years old and at least five years older than the child victim. The World Health Organization, which publishes the International Statistical Classification of Diseases and Related Health Problems (WHO, 2010) defines Pedophilia as a sexual preference for children, boys or girls or both, usually of prepubertal or early pubertal by an adult.

Types of Child Sexual Abusers

One of the first typologies was formulated from the delineation of pedophilic and nonpedophilic child sexual abuse. Groth, Hobson and Gary (1982) classified child sexual abusers based on the degree to which the sexual behavior is entrenched and the basis for psychological needs (fixated-regressed typology). The fixated offender prefers interaction and identifies with children socially and sexually (Simon et al., 1992). These individuals often develop and maintain relationships with children to satisfy their sexual needs (Conte, 1991). In contrast, regressed child sexual abusers prefer social and sexual interaction with adults; their sexual involvement with children is situational and occurs as a result of life stresses (Simon et al., 1992). The majority of fixated child sexual abusers are individuals who sexually assault male children who are not related; regressed child sexual abusers often consist of incest offenders or offenders who sexually assault female adolescents (Priest & Smith, 1992). The fixated-regressed typology has been incorporated into the current models of sexual offending (e.g., self-regulation model; Ward & Hudson, 1998, 2000) discussed later in this chapter.

Victim Characteristic Distinction

Of the traditional models, the victim gender-relationship typology is the only model that has demonstrated clinical utility because it accounts for much of the variability in child sexual abuse, addresses treatment issues and is related to recidivism (Camilleri & Quinsey, 2008). The gender of the victim remains an important distinction among child sexual abusers because this factor has been shown to be a strong predictor of sexual reoffense (Hanson & Bussiere, 1998), although exactly what can be predicted is unclear. One study showed that male child sexual abusers who assault males are twice as likely to recidivate in comparison to offenders who abuse females (Quinsey, 1986). Yet, contradictory findings have also been reported in the literature. Several studies found that child sexual abusers who sexually assault females report over twice as many victims as same-sex child offenders (Abel et al., 1981). More recent studies have shown that mixed-gender child sexual abusers reported the highest number of victims (Cann, Friendship & Gozna, 2007; Stephens, Seto, Goodwill & Cantor, 2016), offenses (Simons & Tyler, 2010) and the highest rates of risk for reoffense (Abel et al., 1988; Kleban, Chesin, Jeglic & Mercado, 2013). However, after controlling for number of victims, mixed-gender offenders were not more likely to sexually recidivate compared to child sexual abusers who offend against males and females exclusively (Stephens et al., 2016). Overall, small sample sizes and reliance on official records have limited the extensive investigation of this group.

Extrafamilial child sexual abusers are more likely to be diagnosed with pedophilia and are often unable to maintain adult relationships.

Within this typology, child sexual abusers are also categorized based on their relationship to the victim (i.e., intrafamilial or extrafamilial). According to Rice and Harris (2002), intrafamilial child sexual abusers (i.e., incest offenders) are less psychopathic, less likely to report male victims, cause less injury, are less likely to exhibit pedophilia and have lower sexual and violent recidivism rates. Intrafamilial child sexual abusers are less likely to have antisocial tendencies (e.g., criminal history, substance abuse) and atypical sexual interests (Seto, Babchishin, Pullman & McPhail, 2015). Seto et al. (2015) also found intrafamilial offenders display fewer offense-supportive beliefs and interpersonal deficits than extrafamilial child sexual abusers. Extrafamilial child sexual abusers are more likely to be diagnosed with pedophilia (Seto et al., 2015) and are often unable to maintain adult relationships (Prentky et al., 1989). Although intrafamilial child sexual abusers substitute a child for an adult sexual partner, they often maintain their adult sexual relationships (Miner & Dwyer, 1997). Studies have reported that intrafamilial child sexual abusers have fewer victims as compared to extrafamilial sexual offenders (Miner & Dwyer, 1997) and lower rates of sexual recidivism (Stephens et al., 2016). These studies relied primarily on official records (i.e., criminal convictions), which do not take into account the possibility that many incest offenders may have undisclosed victims to whom they are not related. Nonetheless, the gender/relationship typology is the most frequently used and researched typology of child sexual abusers.

Rapists

In comparison to child sexual abusers, rapists are more likely to be younger, to be socially competent and to have engaged in an intimate relationship (Gannon & Ward, 2008). Rapists differ from child sexual abusers in that they tend to be of lower socioeconomic status and are more likely to abuse substances and exhibit a personality disorder (e.g., antisocial disorder) or psychosis (Langstrom, Sjostedt & Grann, 2004). In addition, rapists often display the following criminogenic needs: intimacy deficits, negative peer influences, deficits in sexual and general self-regulation and offense-supportive attitudes (e.g., justification of the sexual offense and feelings of entitlement in relation to the expression of a strong sexual desire) (Craissati, 2005).

Rapists and Violent Offenders

Rapists have been found to have a greater number of previous violent convictions, and they tend to use greater levels of aggression and force than child sexual abusers (Bard et al., 1987). Likewise, rapists are more likely to reoffend violently rather than sexually. A meta-analysis conducted by Hanson and Bussiere (1998) found that of 1,839 rapists, 19 percent (n = 349) sexually recidivated and 22 percent (n = 405) violently recidivated over an average follow-up of five years.² The researchers assessed recidivism from several studies that reported the commission of another sex crime (e.g., rape) or violent crime (e.g., assault) through reconviction records (84 percent), arrest records (54 percent), self-reports (25 percent) and parole violation records (16 percent).³ They caution that these findings are based on diverse methods and follow-up periods.

Rapists have been shown to resemble violent offenders or criminals in general. Similar to violent offenders, Simon (2000) found that rapists displayed significant diversity in their offense records in comparison to child sexual abusers and had committed equivalent proportions of drug-related offenses, thefts and burglaries. Harris, Mazerolle and Knight (2009) suggest that rape can be explained by the general theory of crime. Rapists are versatile criminals who engage in many different types of crime over time; sexual offending reflects only one manifestation of an underlying antisocial condition (Gottfredson & Hirschi, 1990).

Types of Rapists

The majority of traditional rapist typologies have focused on the relationship to the victim, degree of aggression, motivation, sexual versus nonsexual nature of the assault and degree of control (impulsive vs. planned). Like child sexual abusers, rapists are often classified by their relationship to the victim (i.e., stranger vs. acquaintance). Seventy-three percent of rapists know their victims (Bureau of Justice Statistics, 2012). Acquaintance rapists are characterized as coercive, less violent and less opportunistic than stranger rapists (Bruinsma, 1995). In contrast, stranger rapists are more hostile and use more expressive violence (i.e., inflicting pain or injury as the goal itself) toward women (Polaschek, Ward & Hudson, 1997).

Rapists have been shown to resemble violent offenders or criminals in general. Acquaintance rapists are less violent and opportunistic than stranger rapists, who are more hostile and use expressive violence.

Rapists have also been classified based upon motivational characteristics. Groth (1979) created a typology based upon the degree of aggression, the underlying motivation of the offender and the existence of other antisocial behaviors, which resulted in four types of rapists. The power-reassurance or sexual-aim rapist is characterized by feelings of inadequacy and poor social skills and does not inflict injury upon his victims (National Center for Women and Policing, 2001). The violence used by the power-reassurance rapist is only sufficient to achieve the compliance of the victim or to complete the sexual act. Such an individual may perceive that the victim has shown a sexual interest in him, or that by the use of force the victim will grow to like him (Craissati, 2005). The power-assertive or antisocial rapist is impulsive, uses aggressive methods of control and abuses substances. His sexual assaults are often unplanned and he is unlikely to use a weapon (Groth, 1979). The third type of rapist is the anger-retaliation or aggressive-aim rapist, who is motivated by power and aggression. This individual sexually assaults for retaliatory reasons and often degrades or humiliates the victim.

The fourth type is the sadistic rapist, who reenacts sexual fantasies involving torture or pain. Sexual sadism is defined as the repeated practice of cruel sexual behavior that is combined with fantasy and characterized by a desire to control the victim (MacCullock et al., 1983). This type is characterized by extensive planning and may often result in sexual murder (Groth, 1979). Although it has been reported in only 5 percent of rapists (see Craissati, 2005, for a review), sexual sadism has consistently been shown as a strong predictor of both sexual and violent recidivism (Hanson & Morton-Bourgon, 2005).

Characterized by Groth's (1979) anger-retaliation rapist, Ramirez, Jeglic and Calkins (2015) examined the relationship between pervasive anger and the use of physical and verbal aggression (including use of a weapon) during a sexual offense. Additionally, the study compared child sexual abusers and rapists with respect to levels of expressive anger and use of violence during the commission of the crime. Records of 571 offenders were reviewed and coded to assess anger (using a pervasive anger measure) and violence used during the sexual offense. Findings indicated rapists were rated as exhibiting more expressive anger than child sexual abusers. Regardless of victim type, sexual offenders who used violence (physical and verbal, but not a weapon) during the sexual offense were evaluated as angrier than those who did not use violence. Taken together, findings provide support for Groth's conceptualization of the third type of rapist.

Although inherently useful for research purposes, these traditional rapist typologies demonstrate little clinical utility because they exclude the irrational cognitions (i.e., offense-supportive beliefs) displayed by most men who commit rape (Hudson & Ward, 1997).

Female Sexual Offenders

Differences between male and female sexual offenders are identified in the literature. In contrast to male sexual offenders, female offenders are more likely to sexually assault males and strangers (Allen, 1991; Vandiver, 2006). Female sexual offenders report different offense-supportive cognitions than males. Specifically, their beliefs are gender-specific; they perceive female abuse as less harmful, men control women and their partner's needs are paramount (Gannon, Hoare, Rose & Parrett, 2010). Studies have also shown that female sexual offenders are less likely than male sexual offenders to sexually reoffend (Freeman & Sandler, 2008). For example, Cortoni and Hanson (2005) found a female sexual recidivism rate of 1 percent over a five-year average follow-up period with a sample of 380 females.

Yet the most evident distinction between male and female offenders is that female offenders are more likely to sexually assault with another person or group (i.e., co-offenders). In a sample of 227 female sexual offenders, Vandiver (2006) found that 46 percent offended with another person and the majority of these co-perpetrators were male (71 percent), 62 percent offended with one individual and 38 percent offended within a group. Studies have differentiated female co-offending according to whether the female participated in an active or passive role (Grayston & De Luca, 1999; Nathan & Ward, 2002). Females who take an active role in the abuse engage in direct sexual contact with the victim. Females who participate passively do not engage in direct sexual contact; instead, these women may observe the abuse but not intervene, procure victims for others to sexually assault or expose children to pornography or sexual interaction (Grayston & De Luca, 1999).  

Typologies of female offenders include the co-offender and the teacher lover/heterosexual nurturer.

Recently, more extensive typologies of female sexual offending have been developed to summarize these female offense characteristics (Matthews, Mathews & Speltz, 1991; Nathan & Ward, 2002; Vandiver & Kercher, 2004). Most of the typologies differentiate female offenders based on the presence of a co-offender, the age of the victim and the motivation for the offense.

Compared to females who abuse alone, females who co-offend are more likely to abuse females and familial victims, to commit multiple sexual offenses (Wijkman, Bijeveld & Hendriks, 2011) and are more likely to be arrested for a non-sexual offense (Vandiver, 2006). Females who abuse alone are more likely to abuse unrelated males and to be diagnosed with a mood disorder (Muskens, Bogaerts, van Casteren & Labrijn, 2011; Vandiver, 2006). Gillespie and colleagues (2014) found a greater prevalence of sexual dissatisfaction, substance abuse, depression, denial and involvement with known offenders among co-offending females. Prior to the offense, female offenders who sexually abuse alone exhibited a greater need for power or dominance, need for intimacy, negative mood state, extensive offense planning and abusive fantasies.

Females who co-offend with a male (i.e., accompanied abusers) have been described as emotionally dependent, socially isolated and displaying low self-esteem (Matthews, Mathews & Speltz, 1991; Muskens et al., 2011; Nathan & Ward, 2002).These individuals are further differentiated based on the use of coercion by the accomplice. Female offenders coerced into sexual offending are motivated by fear and dependence upon the co-offender (Matthews, Mathews & Speltz, 1991; Muskens et al., 2011). Although they initially perpetrate under duress, some later initiate the abuse on their own (Saradjian & Hanks, 1996). These females have been shown to report a history of childhood sexual and physical abuse. Female offenders who accompany a male co-offender and take an active role in the abuse have been shown to be motivated by jealousy and anger and often offend in retaliation (Nathan & Ward, 2002).

Female offenders who sexually abuse alone (i.e., self-initiated abusers) are differentiated based upon age of the victim and motivation for the offense (Nathan & Ward, 2002). One typology, the teacher lover/heterosexual nurturer, describes female offenders who sexually abuse adolescent boys within the context of an acquaintance or position-of-trust relationship (Matthews, Mathews & Speltz, 1991; Vandiver & Kercher, 2004). These females exhibit dependency needs and often abuse substances. They are less likely to report severe child maltreatment; instead, their sexual abuse behaviors often result from a dysfunctional adult relationship and attachment deficits. Female offenders within this category attempt to meet intimacy and/or sexual needs through sexual offending.

Self-initiated female offenders who sexually assault prepubescent children, also referred to as predisposed offenders, have been shown to display significant psychopathologies (Matthews, Mathews & Spletz, 1991). They are more likely than other female offenders to display symptoms of post-traumatic stress disorder (a serious psychological condition that occurs as a result of experiencing a traumatic event) (Foa, Keane & Friedman, 2000) and depression. These female offenders report extensive physical and sexual abuse by caregivers. Researchers contend that they are often motivated by power (i.e., to reenact their childhood trauma, this time as the aggressor) and sexual arousal.

Recently, additional typologies have been added to describe female offenders who sexually assault adult or postpubescent females (Vandiver & Kercher, 2004). Female offenders who engage in the exploitation or forced prostitution of other females have been reported to be motivated by financial gain and have higher number of arrests for nonsexual crimes. Cortoni, Sandler and Freeman (2014) found females convicted of promoting prostitution of a minor tend to be younger at age of first conviction, have a greater history of incarceration and exhibit general criminality (e.g., noncompliance with supervision, antisocial personality) than traditional female sexual offenders. Female offenders who themselves sexually assault other female adults often offend within an intimate relationship as a form of domestic violence (i.e., aggressive homosexual offenders). They are motivated to assault out of anger, retaliation and jealousy.

To reduce the incidence and prevalence of sexual violence in the future, there remains a need for etiological research to provide an empirical basis for treatment interventions.

Although these female typologies are useful to describe offense characteristics, they (like the male typologies) do not provide a theoretical framework for the etiology of sexual offending (Logan, 2008). (For a discussion of the "Etiology of Adult Sexual Offending," see Chapter 2 in the Adult section.) To reduce the incidence and prevalence of sexual violence in the future, there remains a need for etiological research to provide an empirical basis for treatment interventions for female offenders.

Internet Offenders

The widespread availability of pornography on the internet has facilitated the development and maintenance of sexual deviance (Delmonico & Griffin, 2008; Quayle, 2008). The internet has been used as a vehicle for child sexual abuse in at least three ways: viewing pornographic images of children, sharing pornographic images of children and luring or procuring child victims online (Robertiello & Terry, 2007). Individuals download pornographic pictures of children to aid arousal and masturbation, as a collecting activity, as a way of facilitating social relationships and as a substitute for child sexual contact (Quayle & Taylor, 2003).

Internet offender typologies: impulsivity/curiosity, fueling sexual interests, accessing victims/disseminating images, seeking financial gain.

In comparison to child sexual abusers, internet child pornography offenders reported more psychological difficulties in adulthood and fewer sexual convictions (Webb, Craissati & Keen, 2007). In this study of 90 internet offenders and 120 child sexual abusers (Webb, Craissati & Keen, 2007), internet offenders were more likely to succeed in the community (4 percent characterized as failures) and less likely to engage in sexually risky behaviors (14 percent) as compared to child abusers (29 percent and 26 percent, respectively). Formal failure was defined by reconviction, violation and return to prison. With respect to demographics, the majority of offenders are male, younger than other sexual offenders and likely to be of white European descent (Webb, Craissati & Keen, 2007; Quayle, 2008; Seto, Hanson & Babchishin, 2011; Wolak, Finkelhor & Mitchell, 2012). In a recent meta-analysis, Seto, Hanson and Babchishin (2011) reported that in a sample of 2,630 online offenders, 4.6 percent recidivated sexually after an average follow-up period of four years. Likewise, of 983 online offenders, 4.2 percent recidivated with a violent offense. With respect to risk factors, Seto and Elke (2008) reviewed Canadian police files of 282 child pornography offenders to examine sexual contact and predictors of recidivism; 10.3 percent of the sample sexually recidivated and 6.6 percent violently recidivated. Researchers reported substance abuse and criminal history predicted future contact sexual offenses; self-reported sexual interest in children, criminal history and substance use problems predicted future violent offending among child pornography offenders.

Several typologies have been created to categorize internet offenders. In their review of internet offenders, Beech and colleagues (2008) summarized these typologies into four groups. The first group consists of individuals who access pornographic images impulsively and/or out of curiosity. This group includes those who never exhibited sexual problems until they began using the internet (Delmonico & Griffin, 2008). The second group is composed of individuals who access or trade pornography to fuel their sexual interest in children (Beech et al., 2008). For these individuals, the internet facilitates an extension of an already-existing pattern of sexual deviance (Delmonico & Griffin, 2008). The third group consists of sexual offenders who use the internet as part of a pattern of offline contact offending, including those who use it to acquire victims and/or disseminate images that they produce (Beech et al., 2008; Delmonico & Griffin, 2008). The fourth group consists of individuals who download pornographic images for nonsexual reasons (e.g., financial gain). To date, studies have not examined the personality characteristics, criminogenic needs or risk factors of these offenders. In addition, it is not known if these offenders are pedophiles and whether they view pornographic images more than the general population (Quayle, 2004).

A recent qualitative typology was proposed by Tener, Wolak and Finkelhhor (2015), identifying four types of internet offenders who use online communications to commit sex crimes against minors. The typology was derived from 75 case narratives described by law enforcement of offenders who met victims online or knew them and used the internet for communication. Data analysis identified four dimensions of offense characteristics: patterns of online communication, online identities, nature of the relationship and levels of expertise (number of victims, child pornography involvement, sophistication of strategies and awareness of criminality) (Tener et al., 2015). The dimensions were summarized into four internet offender types on a continuum of level of crime expertise (high to low).

The highest level of crime expertise, the experts, consists of sophisticated offenders who systematically procure victims. The experts utilize extensive planning, manipulation and techniques to procure victims. Likewise, they have the ability to use more than one strategy to achieve their goal of sex, for which they spend an extensive time refining. Expert offenders typically meet their victim(s) online and strategically manipulate them into a sexual relationship using a false identity. They sexually assault multiple victims (often hundreds) without emotional attachment; the expert type may offend alone or collaborate with others to acquire victims and/or to share porn. Indeed, pornography use (including production) is prevalent among this type. Expert offenders have explicit awareness of the criminality and use extreme methods to prevent detection. Thirty-two percent of the cases were classified as the experts.

Contrary to the experts, cynical offenders know their victims or, if they meet online, they fabricate or present true identities. Regardless, the online relationship usually progresses into physical meetings. Cynical offenders have less knowledge of offending, fewer skills, fewer victims (one typically) and spend less time manipulating than the expert offenders. In addition, their victim selection is based solely upon personal preference. Like expert offenders, sex is the goal and cynical offenders are not emotionally involved with the victim, are aware they are committing a sex crime and use pornography but the involvement is less extensive (Tener et al., 2015). The cynical type comprised 35 percent of the cases.

Unlike expert and cynical offenders, attention-focused offenders meet victims online with the intention of developing a genuine relationship. As a result, they develop feelings for and become emotionally involved with a minor. Contrary to the other types, the attention-focus offender does not use manipulation. The offender and the victim are mutually interested in a sexual relationship, which may or may not include an awareness of the victim's age (actual) and the criminality of the relationship. If the age is unknown and later revealed, attention-focused offenders often continue with the relationship, even if it remains online only. Pornography is seldom involved in these cases. Twenty-one percent of cases were identified as attention-focused.

The sex-focused type of offender has the lowest level of crime expertise. The online communication in these cases originates on sex-oriented internet sites to arrange a physical, sexual encounter (Tener et al., 2015). Initially, the sex-focused type is not seeking sex with minors; instead, they are looking for immediate sexual interaction with adults. As such, no grooming or manipulation is involved; there is no planning; they present true identities; and the relationship is intentionally brief. Frequently, the sex-focused type inadvertently meets a minor online posing as an adult; they continue the sexual encounter after the victim's age is revealed. From the beginning of the encounter, both parties are interested in immediate sexual gratification without emotional attachment. Tener and colleagues (2015) conceptualize this relationship as exchanges or deals. Twelve percent of cases were categorized as sex-focused.

All of the 75 cases were considered sex crimes and resulted in arrests. Differences among the types of offenders who procure victims online are useful for understanding the heterogeneity of internet sexual offending for need-based intervention (Tener et al., 2015).

For more on "Internet-Facilitated Sexual Offending," see chapter 4 in the Adult section.

Limitations of Traditional Typologies: Crossover Offending

Traditional typologies rely on an official record and/or self-report data. Over 25 years of research (including victim and offender studies) have shown that only 1–3 percent of offenders' self-admitted sexual offenses are identified in official records (Abel et al., 1988; English et al., 2003; Heil, Ahlmeyer & Simons, 2003; Tjaden & Thoennes, 2006). These studies reported a "crossover effect" of sex offenders admitting to multiple victims and offenses atypical of criminal classification. Specifically, studies (e.g., Abel et al., 1988; English et al., 2000; Heil, Ahlmeyer & Simons, 2003; O'Connell, 1998) have shown that rapists often sexually assault children and incest offenders often sexually assault children both within and outside their family. These findings are consistent among populations (e.g., community, prison, parole, probation) and methodologies (e.g., guaranteed confidentiality, polygraph testing). This section reviews the evidence of crossover offending, which challenges the validity of traditional sex offender typologies (those that are based on a known victim type).

Crossover offending presents significant challenges to traditional sex offender typologies.

Despite differences in location and supervision status of offenders, crossover offending has been reported in studies using guaranteed confidentiality, anonymous survey or treatment with polygraphy⁴ (Abel et al., 1988; Emerick & Dutton, 1993; English et al., 2003; Heil, Ahlmeyer & Simons, 2003; O'Connell, 1998; Simons, Heil & English, 2004; Weinrott & Saylor, 1991; Wilcox et al., 2005). The findings indicate that offenders, on average, admit significantly more victims and offenses than are documented in official records. Using official record databases containing 1,345 incarcerated sexual offenders, Cann et al. (2007) found only 8 percent of offenders had assaulted both adults and children, 9 percent victimized males and females and 14 percent of incest offenders sexually abused children not related to them. Cann et al. (2007) attributed the low prevalence of crossover to the use of official records. Using polygraph testing combined with treatment, Heil, Ahlmeyer and Simons (2003) examined offense patterns of 223 incarcerated and 266 paroled sex offenders. This study found that the average number of victims reported in official records (two for incarcerated offenders and one for paroled offenders) increased to 18 and three, respectively, after polygraph testing. The average number of offenses reported in official records increased from 12 for incarcerated offenders and three for paroled offenders to 137 and 14 respectively, after polygraph testing.

These studies have also demonstrated that male sexual offenders engage in crossover sexual offending at higher rates than reported in other studies (e.g., Guay et al., 2001; Marshall, Barbaree & Eccles, 1991; Smallbone & Wortley, 2004). Age crossover (i.e., victimizing both children and adults) ranged from 29 to 73 percent (Simons Heil & English, 2004; Wilcox et al., 2005). Of further interest is the high percentage of official record-identified rapists who admit child sexual victimization. Studies have reported prevalence rates from 32 to as high as 64 percent; the majority of studies found rates in the range of 50 to 60 percent (Abel & Osborn, 1992; English et al., 2000; Heil, Ahlmeyer & Simons, 2003; O'Connell, 1998; Wilcox et al., 2005). With respect to gender crossover (i.e., victimizing both males and females), findings have been relatively consistent and range from 20 to 43 percent (Abel & Osborn, 1992; Elliott, Browne & Kilcoyne, 1995; English et al., 2000; Heil, Ahlmeyer & Simons, 2003). The majority of offenders who assault males have also assaulted females (63–92 percent), but not the reverse (23–37 percent). With respect to relationship crossover, studies have shown that 64–66 percent of incest offenders report sexually assaulting children who they were not related to (Abel and Osborn, 1992; English et al., 2000; Heil, Ahlmeyer & Simons, 2003).

Recent studies have found similar rates of crossover (referred to as polymorphism) using official records coupled with treatment and assessment files. Using a combination of official record and treatment files, Stephens et al. (2016) found 41 percent of 751 adult male sexual offenders did not exhibit stability in victim type (i.e., polymorphism). The most common type of polymorphism consisted of age (35 percent); 13 percent of offenders were gender polymorphic and 11 percent were relationship polymorphic. In addition, polymorphic offenders had a greater number of sexual assault victims. Although Stephens et al., found age and relationship crossover was associated with higher rates of sexual recidivism, the relationship was no longer significant after controlling for the number of victims.

Sims and Proeve (2010) studied the prevalence and patterns of offending against more than one victim type and its relationship to risk of sexual recidivism. To avoid underestimation of sexual offending using official records, a victim was counted if there was a conviction, charge or arrest and the sample included only sexual offenders against children and adolescents with more than one victim. The prevalence rate for 128 sexual offenders who crossed over into at least one domain (i.e., age, gender, relationship) was 63 percent, a rate comparable to studies using guaranteed confidentiality, anonymous survey or treatment with polygraphy. In this study, 48 percent of sexual offenders reported adult and child victims, 22 percent sexually offended against both genders and 26 percent crossed over into the relationship domain. Sexual offenders who victimized both adult and child victims and crossed over in the relationship domain had more victims than those who did not crossover. Crossover was not associated with sexual recidivism rate or frequency of offending. With respect to victim type stability over time, repeat offending was not stable in the age domain; offenders with victims 5 years old or younger show the least stability (i.e., more likely to assault victims from different age categories). Offenders showed the most stability in the gender domain particularly whose index victim was female.

Studies have also shown the rates of age crossover are higher when adolescents are included as a distinct age category (vs. adults and children) (Kleban et al., 2013). In their examination of crossover offending, Kleban et al. (2013) compared three age categories: adolescent victims' ages ranged from 13 to 17 years old, child victims consisted of those under age 13 and adults were 18 years and older. Using archival records of 789 incarcerated offenders, this study examined the prevalence of crossover offending from three victim categories across three domains (age, gender and relationship). Kleban et al. (2013) also examined risk variables to determine which characteristics differentiate crossover offenders from offenders who have a stable offense history. Results indicated 35 percent of offenders had multiple victims at index offense and 26 percent of offenders were convicted of a previous sex crime. Of those with multiple victims at index crime, 13 percent sexually assaulted both genders; 14 percent had a conviction that included a combination of child, adult and adolescent victims; 13 percent had assaulted victims from multiple relationship categories. Of the repeated offenders, 21 percent offended against victims of both genders; 40 percent victimized a combination of adults, adolescents and children; and 48 percent assaulted victims from multiple relationships. Kleban et al. collapsed child and adolescent categories and found only 9 percent of offenders had victims from both age categories. When adolescents were excluded completely, only 5 percent of offenders crossed over from child to adult victim, which suggests examining crossover offending using broad categories of adult and child results in an underestimation of age crossover. Likewise, offenders were more likely to offend against adolescents and children rather than adults and children, suggesting victims should be classified according to developmental characteristics not age. These findings illustrate the importance of a comprehensive history for treatment and management decisions.

Among female sexual offenders, Heil, Simons and Burton (2010) reported similar findings with respect to offense patterns. Using polygraph testing, Simons and colleagues (2008) examined the offense patterns of incarcerated female sex offenders and female sex offenders who had been released in the community. The sample consisted of 74 incarcerated adult female sexual offenders and 22 female sexual offenders in the community who were under supervision at the Colorado Department of Corrections. All participants received cognitive-behavioral treatment. Offense patterns disclosed during treatment with polygraph testing revealed similar findings to those of male offenders. Female sexual offenders reported more extensive offense patterns (i.e., number of victims and offenses, crossover offending) than otherwise indicated by their criminal history.

Simons and colleagues (2008) demonstrated that the average number of victims — reported in official records as one for both incarcerated offenders and offenders in the community — increased to four and three, respectively, after polygraph testing. The average number of offenses increased from 33 for incarcerated offenders and five for offenders in the community to 44 and 13, respectively. In comparison to female sexual offenders in the community, incarcerated female sexual offenders reported significantly more offenses, but these groups were comparable in the number of victims. After polygraph testing, 21 percent of incarcerated females and 11 percent of female offenders in the community reported age crossover (i.e., offending against children and adults). Both incarcerated offenders (30 percent) and those in the community (21 percent) disclosed relationship crossover (i.e., offending against individuals from more than one relationship). This study indicates that female sexual offense patterns may be less extensive than those of male sexual offenders. Nonetheless, this research indicates that female offenders report poor sexual boundaries regarding illegal behaviors and they also disclose legal, but sexually problematic, behaviors. In addition, female offenders were more likely to co-offend than male offenders. Based on polygraph testing, these co-offenses were seldom coercive and the majority of women sexually assaulted alone either before or after the co-offense.

Polygraph testing has also recently been used to distinguish internet offenders who commit "hands-on" child sexual assault from those who do not attempt physical sexual contact. Some internet sex offenders do not attempt physical contact or engage in hands-on sexual offending (e.g., Surjadi et al., 2010; Quayle & Taylor, 2003; Webb, Craissati & Keen, 2007). This classification is important because those individuals who view or download child abuse images but do not have inappropriate contact with children may not pose a direct threat. A recent meta-analysis examined the prevalence of child sexual abuse among internet offenders. Seto, Hanson and Babchishin (2011) reviewed 24 studies and found that 12.5 percent of internet offenders engaged in contact sexual offending as indicated by official records; however, this rate increased to approximately 50 percent using self-report. In this meta-analysis, only one study used polygraph testing to verify the self-report. Bourke and Hernandez (2009) demonstrated significant increases in the number of previously undisclosed victims, offenses and paraphilic interests when self-report is corroborated through polygraph examination. Using polygraph testing, these researchers examined the prevalence of hands-on sexual offending among 155 internet child pornography offenders. Prior to testing, 74 percent (n = 115) of the internet offenders had no known sexual contact with children. After polygraph examination, 85 percent of 155 (n = 132) offenders disclosed hands-on sexual abuse. These findings suggest that crossover to hands-on offending may be more prevalent among internet offenders and further support the use of the polygraph to classify offenders. However, additional research is needed in this area due to the limitations of this study. The sample consisted of volunteers and the majority reported hands-on offenses prior to internet pornography use. Future research should differentiate between those who view pornography and later commit sexual abuse from those who use pornography as a supplement to or a substitute for sexual contact. (For more on "Internet-Facilitated Sexual Offending," see Chapter 4 in the Adult section.)

Taken together, crossover findings suggest that traditional typologies based on victim type may not be useful to allocate resources, evaluate risk or devise individualized treatment interventions. Although crossover findings have been reported in numerous studies using different methodologies, some suggest that the prevalence of age crossover or multiple paraphilias is overstated, particularly in studies that use polygraph testing. Kokish, Levenson and Blasingame (2005) report that 5 percent of individuals stated that they provided false admissions in response to a deceptive result on a polygraph exam. In addition, Marshall (2007) contends that very few sexual offenders commit more than one type of offense.

Accurate self-reporting of victim and offense information remains critical for risk assessment. For example, risk assessment instruments are derived from research demonstrating intrafamilial child sexual abusers present a lower risk of reoffense than extrafamilial offenders (Hanson & Bussiere, 1998); as such, incest offenders are more likely to be assigned as lower risk, consequently receiving less treatment services and supervision. Indeed, crossover research suggests legislative decisions based upon index crime may underestimate an offender's risk level (Kleban et al., 2013). However, the relationship between crossover offending and recidivism are not conclusive. Cann et al. (2007) found crossover offenders scored higher on risk assessment than offenders reporting one victim type; yet Kleban et al. (2013) failed to find a relationship between crossover and sexual recidivism risk. Studies are more likely to find a relationship if age crossover is examined. Harris, Knight, Smallbone and Dennison (2011) found crossover offenders were more likely to sexually recidivate compared to offenders who assault children only; offenders whose victims are adults and incest offenders. In this study, crossover offenders violently recidivated comparably to offenders who assault adults exclusively. After a 15-year follow-up, Parent, Guay and Knight (2011) found offenders who assaulted children and adults (43 percent) reoffended more frequently and sooner than child sexual abusers (3 percent) and rapists (19 percent). Findings from a study conducted by Stephens et al., (2016) indicate number of victims explains the relationship between age crossover and sexual recidivism. 

According to Gannon, Beech and Ward (2008), when offense crossover is disclosed, assigned risk level increases because child sexual abuse of males (i.e., gender crossover), impulsivity and regulation deficits (as suggested by age crossover), and stranger victims (i.e., relationship crossover) are shown to be significantly associated with sexual recidivism (Hanson & Morton-Bourgon, 2004; Stephens et al., 2016). Likewise, Levenson, Becker and Morin (2008) emphasize the importance of understanding crossover offending patterns to enhance safety planning for the offenders and the community. To address the issue of heterogeneity and crossover offending with respect to offender typologies, researchers (e.g., Robertiello & Terry, 2007) have suggested that the best way to regard typologies is as a continuum rather than discrete categorizations, and they emphasize the importance of classifying offenders based on offense characteristics that have been shown to be related to recidivism.

Recent Advances in the Development of Sexual Offense Patterns

The interaction of biological and social learning factors influences the development of sexual offending.

Recent models of the sexual offense process have been devised to include etiological theories of sexual offending and treatment-relevant factors. Assessment, classification and treatment should be formulated from rehabilitation theories, which are integrative practice frameworks that contain elements of etiology, ethics and research (Ward, Yates & Willis, 2011). They are based on clusters of behaviors and psychological processes to account for the heterogeneity of offending. The heterogeneity of traits, experiences and criminal history of sexual offenders requires more advanced methods of categorizing offense behaviors (Martinez-Catena et al., 2016). The most promising are the developmental pathways of sexual offending model, the self-regulation model and the specialist vs. generalist model. These models take into account problematic behaviors, distorted thought processes and offense histories. Developmental factors have been shown to be predictive of high-risk sexual behaviors, treatment failure, and dynamic risk (Craissati & Beech, 2006), and the self-regulation model has been shown to be associated with static and dynamic risk for reoffense (Yates & Kingston, 2006; Simons et al., 2009). The generalist theory of crime (Gottfredson & Hirshi, 1990) has also been examined in sexual offender research (e.g., Lussier, Proulx & LeBlanc, 2005). Similar to crossover findings, studies have shown that few sexual offenders "specialize" in sexual offending (Harris, Mazerolle & Knight, 2009; Lussier, Proulx & LeBlanc, 2005). Specialization has been associated with child sexual abusers who sexually prefer children, while rape has been associated with criminal versatility (Harris, Mazerolle & Knight, 2009). This section reviews models that may ultimately replace traditional typologies to inform treatment and management of sexual offenders. (For more on "Sex Offender Management Strategies," see Chapter 8 in the Adult section.)

Developmental Histories of Sexual Offenders

Due to advanced statistical methods that evaluate the unique and combined contributions of risk factors, more comprehensive descriptions of the psychological processes, developmental histories and offense patterns have been devised to explain sexual deviance. Although they are not described as typologies, they have been shown to be related to different trajectories of offending and they are able to identify criminogenic needs, which have been shown to be predictive of sexual recidivism (Craissati & Beech, 2006; Martinez-Catena et al., 2016).

Etiological research has suggested that it is the interaction of biological and social learning factors that influences the development of sexual offending behaviors (Ward & Beech, 2008). Researchers explain that genetic factors may predispose an individual to pursue a specific human need (e.g., sex or intimacy), but it is the environmental experiences that provide the methods through which these needs are met — either appropriately through the development of relationships or inappropriately through the use of violence (Ward & Beech, 2008). Negative developmental experiences figure prominently in many models of sexual offending behavior.

In a recent study of 614 incarcerated sexual offenders, childhood emotional abuse and neglect predicted sexual victimization and later sexual offending behavior (Jennings, Zgoba, Maschi & Reingle, 2014). In comparison to the general population, incarcerated sexual offenders were more than three times as likely to report sexual abuse during childhood, two times as likely to experience physical abuse, 13 times as likely to experience emotional abuse and more than four times as likely to experience neglect (Levenson, Willis & Prescott, 2016). These individuals were found to be at increased risk for substance abuse and for criminal diversity (Levenson & Socia, 2015).

Indeed, a recent meta-analysis has confirmed the association between the experience of sexual abuse and subsequent sexual offending against children (Jespersen, Lalumiere & Seto, 2009). Yet, not all sexual offenders report being sexually victimized during childhood. Research findings indicate that there may not be only one type of abuse that serves as a developmental risk factor for later sexual offending. Instead, multiple types of abusive experiences, or a pathological family environment, may precede offending behaviors (Dube et al., 2001). Researchers have also suggested that different types of maltreatment may be associated with different types of sexual offending behaviors (e.g., Lee et al., 2002; Simons, Wurtele & Heil, 2002). This section reviews the current research findings that compare the developmental risk factors of various offender characteristics.

Child Sexual Abusers

Researchers have found that child sexual abusers exhibited heightened sexuality in childhood. Meta-analysis results indicate that juveniles who commit sexual offenses were more likely than non-sex offenders to have been exposed to sexual violence, sexual abuse, emotional abuse and neglect (Jespersen, Lalumiere & Seto, 2009). Within the adult sex offender population, Simons, Wurtele and Durham (2004) found that child sexual abusers, as compared to rapists, reported more experiences of child sexual abuse, early exposure to pornography, sexual activities with animals and an earlier onset of masturbation.

Rapists

Rapists, when compared to child sexual abusers, reported more frequent experiences of physical abuse, parental violence and emotional abuse.

In contrast, the childhood histories of rapists appear more indicative of violence. Simons, Wurtele and Durham (2004) found that rapists, when compared to child sexual abusers, reported more frequent experiences of physical abuse, parental violence, emotional abuse and cruelty to animals. Researchers contend that physical abuse, parental violence and emotional abuse result in externalizing behaviors only when they are considered in combination (Lee et al., 2002; McGee, Wolfe & Wilson, 1997). As an illustration, Beauregard, Lussier and Proulx (2004) found that physical and verbal abuse during childhood led to antisocial behavior and callous personality traits, both of which led to aggressive sexual fantasies. Likewise, Salter and colleagues (2003) indicate that the combination of physical violence, domestic violence, emotional abuse and neglect predicted subsequent sexual offending. In a prospective study of childhood abuse histories, Widom and Massey (2014) reported individuals who experienced physical abuse and neglect (but not sexual abuse) were at significantly increased risk for arrest for sexual offenses in comparison to those who did not experience abuse. Researchers (e.g., Craissati, McClurg & Browne, 2002a) explain that an individual who has been raised in an emotionally impoverished environment is often unable to identify his emotions in an accurate manner and, as a result, is likely to become confused when confronted with emotionally charged situations. These individuals often react to confusing situations with overt aggression.

Crossover Offenders

In studies that examined the developmental risk factors of crossover offenders or indiscriminate offenders (e.g., Heil & Simons, 2008; Simons, Tyler & Heil, 2005), findings indicate that indiscriminate offenders report childhood histories of both violence and heightened sexuality. Indiscriminate offenders, also known as mixed offenders, report sexually abusing both adults and children equivalently. With respect to heightened sexuality, Simons, Tyler and Heil (2005) found that indiscriminate offenders were less likely than child sexual abusers to be sexually abused, but they were more likely to report early sexual experiences with peers (before age 10), to have witnessed sexual abuse as a child, and to have had more frequent exposure to pornography before age 10. Similar to child sexual abusers (i.e., 62 percent), 58 percent of indiscriminate offenders reported an early onset (before age 11) and high frequency of masturbation. A great majority of indiscriminate offenders (81 percent) disclosed engaging in bestiality during childhood in comparison to fewer child sexual abusers (59 percent) and rapists (30 percent). With respect to childhood violence, both indiscriminate offenders and rapists described childhood experiences consistent with physical and emotional abuse. However, indiscriminate offenders were exposed to domestic violence significantly more frequently than rapists. Results indicated that parental violence and bestiality were strong predictors of crossover offending.

Female Sexual Offenders

Similar to indiscriminate offenders (of both genders), the majority of female sexual offenders report both violent and sexualized childhoods (Heil, Simons & Burton, 2010). Of a subsample of 42 female sexual offenders, Simons and colleagues (2008) reported that the majority (81 percent) had been sexually abused by multiple perpetrators at a young age with high frequency. Female offenders masturbated later than male offenders (i.e., during adolescence instead of childhood) and with less frequency, but like male offenders who abuse children, they are more likely to masturbate to their abuse experiences and report masturbation to deviant fantasies during adolescence. Likewise, many female offenders were exposed to pornography before age 10, but early exposure is significantly more prevalent among male sexual offenders. Similar to male offenders, females report engaging in bestiality during adolescence, but the prevalence rates for females are significantly lower than for child sexual abusers and indiscriminate offenders of both genders. Similar to indiscriminate offenders, Simons and colleagues (2008) also found that the majority of female sexual offenders reported physical abuse, emotional abuse and witnessing of domestic violence. Although the frequency of physical abuse among female sexual offenders was less than for males, females were more likely to be abused by both male and female perpetrators. Yet, female sexual offenders were more likely than male offenders to report witnessing violence perpetrated by a female; male rapists and indiscriminate offenders more often witnessed violence by a male perpetrator.

Similar findings were reported among 47 female sexual offenders assessed on a childhood experiences measure by Levenson, Willis and Prescott (2015). Compared to the general population, female sexual offenders had more than three times the odds of being sexually abused; four times the odds of experiencing verbal abuse; and more than three times the odds of neglect and having an incarcerated family member. Female sexual offenders experienced multiple adverse childhood experiences consisting of sexual abuse, neglect, verbal abuse and substance abuse in the home, which was associated with having younger victims. These findings emphasize the importance of trauma-informed care with female sexual offenders.

Attachment

In addition to childhood abuse, the majority of sexual offenders (93 percent) exhibited insecure attachment (Marsa et al., 2004). According to researchers, childhood adversities may result in the failure to establish secure attachment bonds to parents (Cicchetti & Lynch, 1995). Marshall (1993) contends that the failure of sex offenders to develop secure attachment bonds in childhood results in their failure to develop sufficient social skills and self-esteem necessary to achieve intimacy with adults. Recent models of sexual deviance suggest that poor parental bonding enhances the effects of child maltreatment and may subsequently initiate the processes that lead to sexual offending by creating vulnerability in the child (Marshall & Marshall, 2000), a lack of empathy for others (Craissati, McClurg & Browne, 2002b), or intimacy deficits (Ward et al., 1995).

Poor parental bonding enhances the effects of child maltreatment and may contribute to sexual offending by creating vulnerability, a lack of empathy and intimacy deficits.

Early attachment research recognized four patterns of attachment: secure attachments that develop when caregivers are consistently responsive to their child's needs; insecure-ambivalent (anxious) attachments that develop when caregivers respond inconsistently to the needs of their child; insecure-avoidant attachments that develop when caregivers are consistently unresponsive to their child's needs; and insecure-disorganized attachment, a category established to describe children who fail to demonstrate a coherent pattern of response to parental separation (Ainsworth & Bowlby, 1991). Recently, attachment style has been associated with different types of offending. Rapists have been shown to exhibit avoidant parental attachments, whereas child sexual abusers display anxious or ambivalent attachment (Simons & Tyler, 2010; Simons, Wurtele & Durham, 2008; Ward et al., 1995). Studies have found that indiscriminate and female offenders were more likely to exhibit disorganized attachment (Simons, Tyler & Heil, 2005; Simons, Wurtele & Durham, 2008).

Maniglio (2012) summarized the influence of attachment on sexual offending behavior. A sexual offender's insecure attachment style during childhood affects the development of social skills and self-confidence, which in turn, prevents initiation or maintenance of intimate relationships during adulthood. Instead, sexual offenders may use deviant sexual fantasies as a means to achieve intimacy, power and control, absent of reality. Maniglio (2012) explains deviant fantasy as a means to achieve intimacy or autonomy creates a disposition to sexually offend. Likewise, in their recent meta-analysis comparing intrafamilial to extrafamilial offenders, Seto et al. (2015) found that family abuse and poor attachment played an important role in the etiology of intrafamilial sexual offending. This finding confirms family dysfunction as an important etiological factor in sexual offending.

Etiological Theory

Taken together, these findings support Marshall and Barbaree's (1990) integrated theory of sexual offending, which postulates that individuals who experienced child maltreatment are likely to exhibit distorted internal working models of relationships, which result in poor social skills and emotional self-regulation. The lack of social skills, especially during adolescence, is likely to result in rejection by others, which in turn will decrease self-esteem, increase anger and produce cognitive distortions about peers and relationships. Negative emotions combined with cognitive distortions may increase the intensity of sexual desire and deviant sexual fantasies (e.g., those about children, whom they perceive as less threatening). Masturbation to these fantasies may serve as a coping mechanism from stress, as a means to exert control, and ultimately, as a behavioral rehearsal to sexual offending. These developmental factors interact with disinhibiting factors (e.g., intoxication, stress, negative affect) and the presence of a potential victim to impair an individual's ability to control their behaviors, which in turn may result in a sexual offense. The emotional and psychological reinforcement of the behavior may be approach oriented (i.e., to achieve a goal directly) or avoidant oriented (i.e., to avoid an unpleasant result). The actual sexual offense combined with cognitive distortions serves to maintain sexual offending behaviors.

The assessment of developmental risk factors is important to determine the criminogenic needs of the individual offender; the assessment also contributes to static predicting (Craissati & Beech, 2006). Consistent with Marshall and Barbaree's (1990) integrated theory of sexual offending, bestiality and masturbation to abuse experiences contribute to the development of deviant sexual interest and frequent masturbation suggests problems with emotional self-regulation. Frequent masturbation coupled with frequent pornography use increases the likelihood of sexual compulsivity. Sexualized coping often serves as a means of meeting needs of intimacy and control (Grady, Levenson & Bolder, 2016). Likewise, insecure attachments suggest intimacy deficits, empathy deficits, antisocial lifestyle and social difficulties. Violence in the home has been shown to be predictive of antisocial lifestyle, hostile attitudes toward women, emotional callousness and hostile masculinity (Malamuth et al., 1991), all of which suggest pro-offending attitudes toward rape. In addition to difficulties with self-regulation, a heightened sexual childhood may lead to the development of child sexual abuse-supportive beliefs (e.g., sexual entitlement, sex with a child is beneficial). As summarized by Craissati and Beech (2006), developmental experiences (sexual and violent experiences and insecure attachment) predict dynamic risk that, when combined with static markers (e.g., male victims, single status), increase the likelihood of reoffense.

Self-Regulation Model

Ward and Hudson (1998, 2000) developed a nine-stage model of the sex offense process, which takes into account the heterogeneity of sexual offending. The self-regulation model (SRM) summarizes the offense process by examining situational precipitants (e.g., desire for deviant sex), cognitive distortions (whether entrenched or function to justify the offense), degree of control over behavior (i.e., impulsiveness or extensive planning), evaluation of sexual assault after the offense and attitude with respect to future offending (positive or negative). SRM contends that individuals are goal-directed as sexual abusers and offend to achieve a desired state — either to satisfy or to avoid offending.

This model proposes that four pathways lead to sexual offending. Two pathways characterize offenders who attempt to avoid offending (avoidance oriented) but do not have adequate strategies (i.e., they have either underregulation or misregulation of self-control) to avoid the undesired behavior (the sexual offense). The two remaining pathways characterize individuals who seek to achieve goals associated with sexual offending (approach oriented) and experience positive feelings as a result. These approach-oriented individuals vary with respect to self-regulation; some of them exhibit deficient self-regulation (i.e., impulsivity), whereas others display intact, effective self-regulation. Thus, the assessment of SRM offense pathways depends on whether the offender attempted to avoid (indirect) or to engage (direct) in the sexual offense, the ability to self-regulate (underregulation, misregulation, effective regulation) and the degree of awareness associated with the sexual offense (implicit or explicit).

The avoidant-passive pathway consists of an offender who attempts to prevent offending (indirect route) but does not have the ability or awareness to prevent the offense (underregulation, implicit awareness). Similarly, the avoidant-active pathway is characterized by the desire to avoid offending (indirect), but the offender uses counterproductive strategies to control deviant thoughts and fantasies (misregulation, explicit awareness). For example, an individual who follows the avoidant-active pathway masturbates to deviant fantasies as an alternative to acting on these fantasies, but this behavior inadvertently increases his/her likelihood to offend. In contrast, the approach-automatic pathway is characterized by the impulsive desire to sexually offend and assault (direct route). Indeed, approach-automatic pathway offenders fail to control their behavior as they respond to situational cues on the basis of well-entrenched cognitive-behavioral scripts that support sexual offending. Individuals on the approach-explicit pathway desire to sexually offend (direct), but they carefully plan their offenses (effective regulation, explicit). Individuals on the approach pathways experience positive emotional states from offending; cognitive dissonance is absent. These offenders do not experience an internal conflict after the offense because they achieved their goal to sexually offend.  

Research on SRM supports the validity of the model and its use in classification and treatment. Specifically, SRM pathways have been shown to differentiate offense characteristics and static and dynamic risk. With respect to offense pathways, incest offenders have been shown to follow the avoidant-passive pathway (Bickley & Beech, 2002, 2003). Rapists are more likely to follow the approach-automatic pathway because their goal is to offend, but they offend impulsively to situational cues (Yates, Kingston & Hall, 2003). Child sexual abusers who offend against male victims are more likely to follow the approach-explicit pathway (Simons & Tyler, 2010). Their goal is to offend and they carefully plan their offenses by establishing relationships with their victims. The indiscriminate (or crossover) offenders who sexually assault both children and adults of both genders and from multiple relationships are more likely to follow the approach-automatic pathway (Simons, McCullar & Tyler, 2008; Simons & Tyler, 2010).

Specialist vs. Generalist Model

The specialist vs. generalist model is another theory that explains the sexual offense process, taking into account the risk and needs of offenders. Although the implicit assumptions about sexual offenders are that they engage in distinct types of crimes and differ significantly from nonsexual offenders, some sexual offenders have been shown to be more versatile in their criminal behaviors and to share attributes with nonsexual offenders. (Lussier, Proulx & LeBlanc, 2005). According to this model, sexual offenders may be characterized as specialists who commit sexual crimes persistently or as generalists who do not restrict themselves to one type of crime; they commit different crimes over time (Lussier, 2005).

One of the assumptions of the traditional explanatory models of sex offending (i.e., the specialist) is that offenders who sexually abuse children engage in sexual offending exclusively. This model has been shown to have a distinct etiology — specifically, a history of childhood sexual abuse (Burton, 2003; Marshall & Marshall, 2000). As previously discussed, developmental studies have demonstrated the association between childhood sexual experiences and sexual abuse of children (Jespersen, Lalumiere & Seto, 2009). Child sexual abusers who are specialists are more likely than generalists to exhibit sexual deviance and sexual preoccupation and to have an emotional congruence with children (Groth, 1979; Harris, Mazerolle & Knight, 2009; Laws & Marshall, 1990).

Similar to rapists, generalist (versatile) offenders resemble violent nonsexual offenders (Craissati, 2005; Langstrom, Sjostedt & Grann, 2004; Simon, 2000). The generalist theory contends that offenders participate in a broad array of activities that are manifestations of low self-control and impulsivity, such as excessive alcohol use, unprotected sex and reckless driving (Gottfredson & Hirschi, 1990). Hanson (2002) concluded that, in addition to sexual deviance, variables such as low self-control, criminal lifestyle, impulsivity and opportunity are important factors associated with sexual offending. Sexual offenders (the majority of rapists and a subset of child sexual abusers) have demonstrated substance abuse and relationship problems, antisocial behavior in adolescence, employment instability and evidence of psychopathy (Harris, Mazerolle & Knight, 2009; Lussier, Proulx & LeBlanc, 2005).

Lussier, Proulx and LeBlanc (2005) examined whether sexual offending among 388 convicted sexual offenders could be explained by a generalist theory of crime using structural equation modeling. They reported differences among child sexual abusers and rapists and concluded that, similar to traditional typologies, the offense patterns of rapists were versatile and that rapists displayed extensive antisocial tendencies. In contrast, child sexual abusers were more likely than rapists to specialize in sexual offending.

Harris, Mazerolle and Knight (2009) examined 374 male sexual offenders to compare these models of sexual offending. The researchers found that the majority of sexual offenders followed the generalist model. Rapists and child sexual abusers exhibited extensive criminal histories, substance abuse issues, antisocial tendencies and psychosis. In addition, few rapists specialized in sexual crimes. Those who did specialize in sexual crimes were more likely to exhibit characteristics similar to child sexual abusers, such as sexual deviance and sexual preoccupation. As Lussier, Proulx and LeBlanc (2005) found, the specialist model was evident in child sexual abusers. Child sexual abusers assessed as specialists were more likely than nonspecialists to know the victim, exhibit sexual preoccupation and display emotional congruence with children.

Francis, Harris, Wallace, Knight and Soothill (2013) examined the life course of 780 sexual offenders in civil commitment treatment between 1959 and 1984. Specifically, this study investigated distinct trajectories of offending, comparing generalist crime to specialist crime. Results identified four trajectories to sexual offending, which varied according to offense pattern. Differences were found with respect to criminal onset, length of criminal career, age of peak offending and time of entry into treatment. Overall, late onset was associated with child sexual abuse and early onset (younger) was associated with rape. Likewise, findings indicated sexual offending began later than nonsexual offending and three out of four groups exhibited a decrease in frequency with age. Two groups (low-rate and high-rate limited) offended at an earlier age than the other groups (low-rate persistent and high-rate persistent). Low-rate persistent offenders (56 percent of the sample) began offending during late teens and offended less than once per year with the highest point in their 30s. This group was equally as likely to commit rape as child sexual abuse. High-rate limited offenders (24 percent) exhibited an earlier age of onset and offended most frequently (average twice per year) during their late 20s. This group consisted mostly of rapists. This trajectory was consistent with the generalist pattern and the decline in offending occurred during their 50s. The third group or high-rate accelerators (12 percent) began offending during their 20s and their offending increased until mid-40s; this group consisted primarily of child sexual abusers. The fourth group was classified as late onset accelerators (8 percent). They began sexual offending during their late 20s and the offending behaviors increased to its peak during their mid-50s. The majority of these offenders sexually assaulted relatives (i.e., incest offenders). Taken together, findings indicate there are distinct trajectories of offending based upon onset, frequency and persistence.

These findings are consistent with many traditional typologies of rapists and child sexual abusers; however, the results suggest that the generalist vs. specialist model is a better way to assess sexual offenders, regardless of victim type. Future research in this area is needed to further identify factors that characterize specialist offenders from generalist offenders.

Summary

Advances in developmental risk factors and offense pathways can assist with risk and need evaluation, but additional research is needed to develop models of sexual deviance.

The prevention of sexual violence requires a balance of community safety with effective resource allocation. Current research emphasizes the importance of a comprehensive approach to sex offender typologies, through the assessment of criminogenic needs (dynamic risk) and offense patterns, not based upon the type of victim exclusively (Martinez-Catena et al., 2016). Recent advances in our knowledge of developmental risk factors and offense pathways can assist with risk and need evaluation, but additional research is needed to develop more extensive models to explain sexual deviance. Nonetheless, through a comprehensive understanding of treatment needs and subsequent effective intervention, an offender can attend to the process, learn skills and alternative strategies to sexual violence and, ultimately, strive to live a healthy lifestyle without offending.