part0023

Most research on pedophilia has drawn on studies of men who have been charged or convicted of sexual offenses involving children, usually contact sexual offenses. Some research has been conducted on noncontact offenses as well, including indecent exposure or voyeurism involving children, child pornography offenses, and online sexual solicitations of minors. The criminal justice population has received the most attention because of the clinical and public interest in sexual offenses involving children. The advantages to studying pedophilia in criminal justice samples are that (a) the clinical and social needs mean more funding and institutional support for research; (b) collateral information obtained by police and other authorities is available; and (c) research on etiology, assessment, and intervention has applied value.

The disadvantages to studying pedophilia in criminal justice samples include that (a) the self-report bias is likely strongest with this group, given they are facing serious consequences, including sentencing, treatment, and release decisions; and (b) only about half of the men who have committed contact sexual offenses involving children have pedophilia. As I discuss further in the next chapter , nonpedophilic offenders against children might seek sex with a nonpreferred person because of a desire for sexual gratification and an indifference to the possible harm to the child, disinhibition because of substance use or cognitive impairment, opportunism, or a lack of other sexual options because of their unattractiveness, lack of resources, or poor social skills.

Men who have committed sexual offenses have, by definition, engaged in criminal and antisocial behavior. They are likely to differ in meaningful ways from persons with pedophilia who have not committed such offenses or who have committed sexual offenses that have not yet been detected. For example, one would expect pedophilic offenders who have been detected by authorities to be more persistent (more offenses mean a greater likelihood of being reported), more antisocial (to overcome any inhibitions they have against offending), and possibly higher in sexual preoccupation than nonoffending persons with pedophilia or undetected offenders.

It is also the case that the nature of the sexual offense involving children may matter. Having sexual contacts with more young children is associated with a greater likelihood of pedophilia; as noted in the previous chapter , offending against boys or against unrelated victims is also informative. Noncontact sexual offenses such as indecent exposure directed at children tend to have many victims, suggesting pedophilia, but an unanswered question is the extent to which the strength of an activity paraphilia, such as exhibitionism or voyeurism, competes with age or gender preferences. Having many noncontact child victims may reflect a dominant motivation of exhibitionism or voyeurism rather than pedophilia, although it could involve both. On the other hand, most heterosexual men who engage in exhibitionism expose only to women, not to men or children, suggesting gender and age interests still orient activity paraphilias. Last, evidence indicates that child pornography offending is strongly associated with pedophilia, even more so than contact sexual offending (Seto, Cantor, & Blanchard, 2006). In their meta-analysis, Babchishin, Hanson, and VanZuylen (2015) found that dual offenders, who had committed both child pornography and contact sexual offenses, were the most likely to have pedophilia.

Our explanation in the Seto et al. (2006) study is that individuals seek pornography content that matches their sexual interests (see Ogas & Gaddam, 2011). For example, straight men might occasionally look for gay male pornography, but they usually do not repeatedly view it or amass large collections of it. Using parallel logic, men who have been criminally charged for child pornography offending are likely to be sexually interested in children. However, sexual offenses against children include both those who have pedophilia or hebephilia as well as nonparaphilic individuals who are acting opportunistically (e.g., offending against a 12-year-old girl who looks older than she is). Combining pedophilic and nonpedophilic offenders produces a lower average index of sexual response to children than combining predominantly pedophilic child pornography offenders with a minority who committed these crimes for other reasons.

A different set of selection effects may occur when studying pedophilia in clinical settings. Here, persons with pedophilia—whether they have offended or not—would be expected to be higher in psychopathology, distress, and/or impairment than persons with pedophilia seen outside of clinical settings. This is the expectation because those who are distressed or experiencing impairment are more likely to seek mental health services or to experience distress because of pressure to seek clinical help from family, friends, or partners. Persons with pedophilia who are not distressed by their sexual interest in prepubescent children and not facing negative consequences as a result—ostracism, criminal justice sanctions for committing sexual offenses—are much less likely to be seen in clinical settings or to admit to having pedophilia if seen for unrelated reasons.

It would be very valuable to have more research on self-identified persons with pedophilia who were not seen in clinical or criminal justice settings. This research is more feasible now thanks to online surveys that can protect participant anonymity and confidentiality using techniques such as not logging Internet Protocol (IP) addresses, framing questions to avoid any incriminating information (e.g., details about specific child victims that could trigger mandatory reporting requirements), and maintaining secure, encrypted servers. I discuss this newer survey research after mentioning several initial survey efforts, to provide comparisons over time and methodologies.

Some might argue that research on self-identified persons with pedophilia who are not involved in the criminal justice system and not presenting because of clinical concerns would be the best way to understand this chronophilia. It has several advantages, as noted. However, the biggest limitation is the reliance on self-report because of unconscious self-report biases and the conscious tendencies to present oneself in the best possible light, especially if recruitment takes places through pedophilia advocacy groups. However, some online research platforms allow the inclusion of viewing time or other measures that do not rely solely on self-report (e.g., Dombert et al., 2016).

People who self-identify as having pedophilia online may be polarized, either declaring their intent not to sexually offend and supporting a ban on adult–child sex or complaining about societal prohibitions against adult–child sex and suggesting it is a moral panic; the silent majority may not be represented in online discourse. Studies of clinical and criminal justice samples usually have other sources of information, whether that is criminal or other records, or reports from others, such as a romantic partner or family member regarding sexual offense history, sexual interests, and sexual behavior. Clinical or criminal justice samples might also have information from objective assessments.

Here, I provide a critical overview of studies of self-identified persons with pedophilia or hebephilia, clinical samples of men who are being assessed for or have been diagnosed with pedophilia or hebephilia, and criminal justice samples of men who have been charged or convicted of sexual offenses involving children. I am particularly interested in highlighting similarities and differences in study findings. Two questions are of particular interest: How do the various study groups differ from each other on factors that are important in understanding of the origins of sexual offending against children? How do they differ in factors that are candidates in the etiology of pedophilia?

In this section, I review the relatively small set of studies of self-identified persons with pedophilia or hebephilia in some detail. This is of particular interest because comparing the results of this work with research using clinical or criminal justice samples will illuminate what kinds of selection effects might occur, which would help contextualize and interpret the findings from different recruitment sources. Studying self-identified persons with pedophilia is also the best way of learning more about nonoffending persons with pedophilia and persons with pedophilia who have committed undetected offenses; this would also reveal possible selection effects in studying detected offenders. It also could shed much needed light on the factors that might protect individuals from acting on their sexual interest in children.

I begin with a survey of 77 members of the Paedophile Information Exchange, a defunct group of adult–child sex advocates and self-identified persons with pedophilia or hebephilia based in England in the late 1970s. With the support of the organization’s chairperson, G. D. Wilson and Cox (1983) were able to send out questionnaires to all members of the mailing list, estimated to comprise approximately 150 persons. All the respondents were men between the ages of 20 and 60, with a modal age between 35 and 40. Unlike clinical and criminal justice samples, which usually involve more persons with pedophilia attracted to girls, more of these self-identified persons with pedophilia reported to G. D. Wilson and Cox that they preferred boys (71%), with 17% attracted to both boys and girls and 12% preferring girls. The preferred age range for boys was 12 to 14, which would encompass both pedophilia and hebephilia, whereas the preferred age range for girls was between 8 and 10, which would predominantly involve pedophilia. The most parsimonious explanation for the age difference for younger girls for those respondents who were attracted to both boys and girls is that the age of puberty onset is younger for girls than for boys and, thus, self-identified persons with pedophilia would be attracted to younger girls in terms of age (recall Herman-Giddens et al., 1997; Herman-Giddens, Wang, & Koch, 2001).

A few of the respondents in the G. D. Wilson and Cox (1983) survey reported that their preference was for postpubescent adolescents. They might have been persons with pedophilia who were more attracted to older adolescents but who were still interested in a substantial way in prepubescent or pubescent children. Alternatively, because interest in postpubescent adolescents is also socially frowned upon, some members may have had this chronophilia (ephebophilia) and had joined the Exchange for lack of another community.

An interesting insight from these surveys that is unlikely to come from clinical or criminal justice samples is the fact that the attraction for some pedophilic individuals was not only sexual—When asked about their fantasies regarding children, 39 of the 77 respondents reported having sexual fantasies, but 22 reported having romantic or caring fantasies instead or as well. This was perhaps a sensitive question because 18 reported having no sexual fantasies and 7 gave no response, representing a third of the sample. (Nonresponse was an issue for other questions as well, raising concerns about survey design and self-censoring.) The psychological and physical characteristics that were attractive varied a great deal, such that no feature was endorsed by a majority of respondents. Astoundingly, to me, this kind of survey has not been repeated, so researchers still know very little about what persons with pedophilia find attractive about children and what they do know may be dated.

Regarding their feelings about sex with adults, 14 respondents to G. D. Wilson and Cox (1983) viewed it negatively, 33 were indifferent, and 14 viewed it positively (presumably persons with nonexclusive pedophilia). In addition, nine somehow misunderstood the question so that their responses did not make sense, and seven did not respond or their answers could not be categorized.

Around the same time as the G. D. Wilson and Cox (1983) survey, Bernard (1979/1985) surveyed 50 members of a Dutch pedophilia advocacy group. Again, the group was young, with the majority of respondents reporting they were under the age of 40. Most (90%) were unmarried and had no children of their own, which would be expected of persons with exclusive pedophilia or hebephilia. Again, unlike the clinical or criminal justice samples described later, most (96%) of the respondents preferred boys, with the peak age being 12 or 13; the other two respondents liked boys and girls equally. A majority (56%) of the respondents admitted they were currently having sexual contact with a child, and a similar slight majority (54%) had previously been convicted of sexual offenses involving children. Bernard did not comment on the overlap—for example, on how many of those with prior convictions were currently having sexual contacts with children. A minority (14%) reported having sexual contacts with more than 50 children in their lifetime.

Many of the Bernard (1979/1985) participants were at peace with their sexual interest in children, with most saying they would not change it even if it was possible to change it. Similarly, many of the respondents in the G. D. Wilson and Cox (1983) survey were not distressed by their pedophilia: Three were bitter or angry with society, putting the blame on societal views, but 27 reported having happy, proud, or other positive feelings about their sexual interest in children. A quarter of the respondents in Bernard reported they became aware of their pedophilia before the age of 15, and two thirds said they had sexual contact with a child by the age of 20. This also can be compared with the previously mentioned B4U-ACT survey of 193 respondents (98% male) who reported an average age of 14 for awareness of their sexual attraction to children, as well as to Tozdan and Briken (2015).

Li (1991) interviewed 27 self-identified persons with pedophilia about their views. A third of the sample reported they thought their pedophilia was innate, and a similar number thought the stigma associated with pedophilia was culturally relative, that is, it would be accepted in other cultures or times. Over half the sample mentioned specific child characteristics they found attractive, and similar to the respondents in G. D. Wilson and Cox (1983), these included psychological characteristics, such as innocence, affection, and warmth. The respondents reported that relationships or sex with children were more satisfying than the same with adults, indicating a true preference for children.

Riegel (2004) conducted an online survey of 290 anonymous self-identified persons with pedophilia, again most indicating they were sexually attracted to boys. Reflecting the impact of the Internet on access to child pornography (see Seto, 2013), across the past 15 years, relative to surveys conducted in the 1970s and 1980s when child pornography involved physical objects (e.g., photographs, magazines, films) that were expensive and difficult to obtain, most of Riegel’s respondents had viewed child pornography online. A majority of these respondents thought viewing child pornography reduced, rather than increased, their urges to have sexual contacts with children, but the veracity of this opinion was not evaluated. Child pornography could provide an outlet for sexual fantasy and masturbation, so one would see a negative correlation, or it could increase desire for sexual contact with a real child resulting in a positive correlation. As discussed in the next chapter on the origins of sexual offending, it is likely that both effects exist, where the direction and magnitude depend on individual differences, just as general pornography use could increase or decrease sexual aggression directed at women for different men (see Hald & Malamuth, 2015; Seto, Maric, & Barbaree, 2001).

The results of these surveys—G. D. Wilson and Cox (1983), Bernard (1979/1985), and Riegel (2004)—suggest that being boy-attracted makes someone with pedophilia or hebephilia more likely to become involved in these kinds of advocacy groups. A similar trend was found for more boy-attracted individuals in a help-seeking sample described by Beier et al. (2009), with equal numbers (45% each) preferring only boys or only girls, and the remainder preferring both boys and girls. This striking observation raises the possibility that boy-attracted persons with pedophilia are either less likely to act on their sexual interests or are less likely to be reported or discovered if they do act and, thus, are underrepresented in clinical or criminal samples. Consistent with this idea is evidence that boys are less likely to disclose sexual abuse than girls, especially older boys (see Paine & Hansen, 2002). At the same time, among identified offenders, those who have boy victims are more likely to have pedophilia and are more likely to sexually reoffend.

Bailey, Bernhard, and Hsu (2016) conducted an online survey of 1,189 men, recruited from pedophilia or hebephilia websites, who self-identified as being attracted to children. Distinguishing by gender, 29% percent of the sample reported that their highest attraction was to prepubescent girls, suggesting pedophilia, with 17% reporting an exclusive attraction. Twenty-eight percent reported that their highest attraction was to pubescent girls, suggest of hebephilia, with 13% reporting an exclusive attraction. The corresponding values among the respondents who were interested in boys was 18% for prepubescent boys (7% exclusive) and 36% for pubescent boys (21% exclusive). The remaining participants reported their maximum attraction was to adolescents or adults, but they still reported substantial sexual interest in children, suggesting they were nonexclusive.

Analyzing data from the same sample, Bailey et al. (2016) found that most of the participants did not report any detected sexual offenses. Only questions about detected offenses were asked, to avoid the risk of self-incrimination in the survey; it is almost certain that some of these participants had committed undetected sexual offenses. Detected sexual offending was associated with opportunity (older, repeatedly working in positions with high access to children), relative strength of pedophilia (relative attraction to children versus adults, attraction to boys), repeatedly falling in love with children, and struggling not to offend. Consistent with the idea that sexual abuse history may play a role in sexual offending against children (see the next chapter ), childhood sexual abuse history was correlated with detected sexual offending. In contrast, permissive attitudes about adult–child sex and frequency of sexual fantasies about children were unrelated to sexual offending.

Observations of posts on pedophilia forums are also relevant, but I discuss only Holt, Blevins, and Burkert (2010); interested readers should also see Durkin and Bryant (1999), Malesky and Ennis (2004), and O’Halloran and Quayle (2010). Conclusions that could be drawn from this set of studies is that some pedophilia forum participants believe that children can be interested in sex and can initiate and/or that children can benefit from the sexual attention of loving, caring adults. Holt et al. (2010) qualitatively analyzed 705 threads from five online pedophilia forums. Several themes were identified, the most prominent being marginalization, a sense of being ostracized and beleaguered because of stigma, public fear/hatred of pedophilia, and the conflation of pedophilia with sexual offending against children. Reflecting the stigma associated with pedophilia, Holt et al. noted that language used by participants often referred to “boy love” or “girl love” rather than pedophilia. In general, Holt et al. found mutual support and reinforcement of group values. Users also talked about sexual contacts with children, but not in graphic detail (e.g., talking about kissing and touching). The norm seemed to be disapproving of any sexual contact that involved perceived psychological or physical harm, with some discussion of attractive features (e.g., boys in sports uniforms) and puberty recognized as liminal:

One of the reasons that studies of self-identified persons with pedophilia are valuable is that they can include nonoffending persons with pedophilia, who are less likely to appear in clinical samples and by definition do not end up in criminal justice samples. To illustrate this point, many of the help-seeking persons with pedophilia or hebephilia assessed in the Dunkelfeld Prevention Project had committed sexual offenses at some point in their lifetimes, either by having sexual contact with children or by interacting with child pornography (Neutze, Seto, Schaefer, Mundt, & Beier, 2011). The opportunity for prevention is that some of the respondents had no known contact sexual offenses and might be supported from crossing that line, and even among those who had sexually offended against children or engaged with child pornography, some had been inactive in the past 6 months and might be helped to remain inactive.

Another online survey is germane. As mentioned in Chapter 1 , the organization B4U-ACT surveyed 193 respondents (98% male) in 2011, with a modal age of 30 to 39 years and a plurality (48%) from the United States. These respondents reported their average age of first attraction to children was 12 (median was 13), and 85% reported experiencing this attraction before adulthood. The average age of awareness of having pedophilia was 14, median age of 16, with two thirds becoming aware before adulthood. This difference in ages of onset reflects the lag in understanding one’s sexual attraction to children and applying a label to that attraction. It likely also reflects a growing awareness of difference: Many 12-year-olds are attracted to same-age peers, but their interest shifts as they get older, whereas pedophilic individuals do not show this shift in age interests. This is a topic I return to in discussing the etiology of pedophilia in Chapter 5 .

Houtepen, Sijtsema, and Bogaerts (2016) carried out a qualitative interview study of 15 self-identified male persons with pedophilia in Europe recruited through online pedophilia forums. Most of the sample was unknown to authorities, yet the majority (11 of 15) had been involved with child pornography at some point. One participant had been convicted of contact sexual offending, but five admitted to having sexual contact with a child at some point; much of this offending took place in adolescence, and a number of participants emphasized they no longer offended and that their sexual attraction to children was less important as they grew older.

The majority (11 of 15) of Houtepen et al.’s (2016) respondents reported that they gradually became aware of their sexual attraction to other children, similar to their peers at early puberty, but then their age interest remained stable as they grew older, while their peers shifted developmentally. The other four said they became aware of their attraction to children suddenly, by seeing a child image or by having sexual contact with children. Seven respondents said they were also sexually aroused by adults and could express this through fantasy and masturbation.

Ten of the 15 respondents said their attraction to children was not purely sexual, they also had romantic feelings, either by falling in love with a child or having romantic fantasies about a child. Consistent with G. D. Wilson and Cox’s (1983) sample, eight reported both physical (e.g., “their beauty,” “bodily shapes”) and psychological characteristics (e.g., “openness,” “naivete”) were attractive to them; only four mentioned appearance only (similar to many teleiophilic men being interested in features other than physical appearance of women). Ten mentioned how having close relationships with children was very satisfying, for example, having a friendship or being in a mentor role, again consistent with the idea that this is not a purely sexual interest.

Indeed, reflecting the moral landscape in which discussions of pedophilia and hebephilia take place, the general public view is highly negative and stigmatizing, with some glimmers of understanding or compassion. But what about those individuals with pedophilia (or hebephilia) who are not interested in refraining from acting on their attraction to children, think the laws and norms are wrong, and who in fact advocate for the decriminalization of adult–child sex? Researchers do not know about the prevalence of unabashed pedophilic and hebephilic individuals and how they differ from those who do want to refrain.

Neutze, Grundmann, Scherner, and Beier (2012) compared detected and undetected offenders in a subsample of 345 self-referred persons with pedophilia or hebephilia (60% had pedophilia) who were assessed between 2005 and 2010 and admitted to one or more sexual offenses. Fifty-seven percent were not known to the authorities; 29% of the sample was facing legal charges; and the rest had a criminal history. Seventy-nine percent had committed child pornography offenses and 63% had committed contact sexual offenses; 42% had committed both kinds of sexual offenses. Although detected and undetected offenders shared more similarities than differences, the two groups showed some significant differences on theoretically and clinically meaningful factors (see Chapters 4 and 7 , this volume): Undetected offenders were younger, more educated, and more likely to be currently employed than detected offenders. No group differences were found in offense-supportive cognitions, such as the idea that children can freely consent to sex, empathy, or interpersonal deficits such as loneliness or hostility to women. Undetected offenders scored higher in sexual self-regulation problems, including being more likely to report other paraphilic sexual interests.

In another study, Schaefer et al. (2010) compared 97 individuals whom they described as potential offenders, because they were sexually interested in children, and 63 undetected Dunkelfeld offenders. Data came from a computer administered screening, not the full clinical assessment reported by Neutze et al. (2012). All participants admitted having sexual fantasies about children. The undetected offenders were more likely to be concerned about their risk of offending in the future and were more likely to have sought help. Undetected offenders were also more likely to be fathers (which does not necessarily mean their undetected victims were related children). The two groups did not differ, however, in distress about sexual fantasies regarding children, sexual history (age of onset of fantasies, awareness of fantasies by age 20, sexual experience with adults, relationship status), or the age or gender of children in their fantasies. In Schaefer et al. (2010), undetected offenders differed from nonoffenders in a number of potentially important ways. Nonoffending participants reported more interest in boys and were higher in their stated intention not to offend.

Clinical samples are another study group to understand pedophilia and its relationship with sexual offending against children, overlapping both with self-identified pedophilic and hebephilic individuals (who self-refer because of distress or other concerns) and individuals who are involved with the criminal justice system (and who self-refer or are referred for evaluation or treatment). Much more is known about clinically referred individuals who have criminal charges or convictions than those who have committed undetected offenses or who have no known offenses. I begin with an unpublished study by Fedoroff, Smolewska, Selhi, Ng, and Bradford (2001), who found that 26 of 316 (8%) consecutively assessed individuals diagnosed with pedophilia were self-referred and had no known child victims. These 26 nonoffending persons with pedophilia were significantly less likely than pedophilic offenders to have any kind of criminal history, although a third still had a criminal record of some kind. They were also more likely to report having a sexual abuse history, more likely to be virgins, and more likely to use pornography; whether it was mainstream or child pornography was not specified. No group differences were found in biographic characteristics (e.g., age, education, substance use) or in aspects of sexual history (e.g., age at first intercourse, age of first sexual partner).

Because all the data were derived from self-report, it is possible that participants in Fedoroff et al. (2001) lied about their criminal or sexual history; for example, exaggerating their sexual history with adults or minimizing their sexual behavior with children or with child pornography use. The limitations of having only self-reports are particularly salient for the offending group, which was usually assessed for legal reasons, whereas the self-identified pedophilic individuals were self-referred and therefore presumably less likely to deny or minimize problems; if anything, the self-identified may have exaggerated their problems because of their distress. This reliance on self-report can be mitigated to some extent in some clinical settings, which have some access to collateral information, such as reports from partners or family members, referral documents, and criminal record checks.

The biggest source of data about clinically seen pedophilic individuals who are not necessarily in conflict with the law is the Dunkelfeld Prevention Project, which I return to again when discussing innovative intervention approaches in Chapter 9 (this volume). I have already discussed several studies from this ground-breaking work. The Dunkelfeld Project began as a pilot project in Berlin in 2004, using mass media advertisements to recruit help-seeking pedophilic and hebephilic individuals from the community for assessment and possibly treatment using cognitive–behavioral principles. The ambition of the Dunkelfeld Project—which has now spread to multiple clinics through Germany—was to prevent sexual offenses against children by reaching individuals concerned about their sexual interest in children.

Beier et al. (2009) reported that 808 individuals contacted the Project in the first 38 months from its launch. Of these individuals, 358 (44%) traveled to the outpatient clinic in Berlin to complete the assessment; the average distance traveled, by those who reported this, was approximately 200 km (124 miles). Of the 358 individuals who were assessed, 60% were classified as pedophilic, 28% were classified as hebephilic, and the remainder were classified as teleiophilic. Focusing on the pedophilic group, 61% were deemed to be exclusive in their interest in children; equal numbers (45% each) preferred boys or girls, with 10% preferring boys and girls. A small minority (15%) had evidence of another paraphilia (e.g., fetishism, exhibitionism, voyeurism). Regarding help seeking, 48% had sought mental health services in the past 6 months. A majority of the combined pedophilic and hebephilic group (86%, not reported separately) had admitted their sexual interest in children to someone they trusted, such as a family member or friend. Preliminary evaluation results from this project were reported by Beier et al. (2015) and are discussed in Chapter 9 (this volume).

Most research on pedophilia has focused on criminal justice samples of adult male offenders, driven by the numbers and the public and professional concerns about sexual offenses involving children. Research on pedophilia is an offshoot of the policy and practice concerns, rather than being a funded and supported endeavor of its own, to understand this paraphilia and human variation in sexual interests more broadly. It is only relatively recently—mostly in the past decade—that more attention is being paid to pedophilia separate from the criminal concerns.

I have suggested that approximately half of identified sex offenders against children are likely to have pedophilia, with variation depending on multiple parameters, including how samples are recruited, how pedophilia is operationalized, and how pedophilia is assessed or ascertained. For sample characteristics, the prevalence of pedophilia depends on factors such as legal status, where one would expect more offenders being willing to admit their sexual interest in prepubescent children if they are being assessed as part of a postconviction evaluation for treatment than if they are being assessed as part of presentencing; recruitment source, where individuals seen in a specialty sexual offender clinic are more likely to be suspected of pedophilia than individuals seen in a general assessment service, because it will be offenders who are suspected of having pedophilia or who show more indicators, such as multiple sexual offenses, who are more likely to be referred to a specialty clinic; and sexual offending history, where as I have already discussed, offenders with boy victims, multiple child victims, younger child victims, and unrelated child victims are more likely to be pedophilic. In sum, my estimate of an approximately 50% prevalence of pedophilia among identified sex offenders is a generalization across samples; it does not mean that individual settings should expect half of the offenders they see to have pedophilia.

The prevalence of pedophilia also depends on whether both exclusive and nonexclusive forms are included, whether it refers only to prepubescent children or includes interest in pubescent children, and whether the focus is on pedophilia rather than pedophilic disorder. The prevalence also depends on whether pedophilia is assessed through self-report, sexual behavior, or objective testing. As noted, self-report is the easiest information to obtain but also the easiest to distort, intentionally or not.

Sexual behavior provides a conservative measure of pedophilia because pedophilia is not the sole motivation for sexual offending, and the role of pedophilia differs across offender groups—as I discuss shortly, pedophilia is more common in child pornography offenders than contact offenders, probably because they are less constrained by opportunity than offenders who offend directly against children and because seeking child pornography online despite so much other pornography available may be a more specific or stronger indicator. Pedophilia is least likely among sexual solicitation offenders, most of whom approach young adolescents between the ages of 13 and 15 (i.e., persons below the legal age of consent but unlikely to indicate pedophilia). Phallometric testing is highly specific as it is usually conducted, but a price is paid in sensitivity. Phallometric testing can also be affected by faking and by refusal to participate, where one might think that those motivated to fake their sexual response patterns or to refuse testing are more likely to be pedophilic than those who agree to participate and do not (obviously) fake.

A few studies have found much higher proportions of pedophilia through phallometric testing, possibly reflecting differences in the sources of those samples, methods used, and comparison groups. For example, Chaplin, Rice, and Harris (1995) found a high proportion of pedophilia in a sample of sex offenders against children assessed at a maximum-security psychiatric facility, using an optimized phallometric procedure. Seto et al. (2006) found that more than two thirds of child pornography offenders assessed at a specialty clinic using a well-validated phallometric test showed greater sexual arousal to children than to adults when assessed phallometrically.

In the following sections, I briefly summarize key findings from research on identified sex offenders, recognizing that these studies are often mentioned again throughout this book. The most notable findings are the similarities and differences between online and contact offenders, evidence that identified offenders are more antisocial and have more extensive criminal histories, which is in line with the differences reported for comparisons of detected and undetected offenders and in line with explanations of sexual offending against children ( next chapter ).

Unlike the prevalence of attraction to boys observed among self-identified pedophilic individuals, the majority of adjudicated sex offenders have girl victims, and the majority of self-identified victims are girls (e.g., Seto & Lalumière, 2001; Seto, Stephens, Lalumière, & Cantor, 2017; Stoltenborgh, van IJzendoorn, Euser, & Bakermans-Kranenburg, 2011). It is unclear why this large discrepancy exists between the reports of self-identified individuals compared with the reports of self-identified victims or data on officially known victims. Being attracted to boys over girls is a significant correlate of having pedophilia (see Chapter 2 , this volume) and is a strong predictor of sexual recidivism among identified sex offenders (see Chapter 7 ). One possibility, already raised, is that those who are attracted to boys are more likely to be online and/or to participate in surveys. Another is that those who are attracted to boys are less likely to act on their interest, but once they do act, they are more likely to persist. These explanations are inconsistent with results from Bailey and colleagues (Bailey, Bernhard, & Hsu, 2016), however, because the majority of self-identified individuals in their samples preferred girls and being attracted to boys was associated with detected sexual offending. A third possibility is that those who offend against boys are less likely to get caught.

Most identified offenders are known to their victims, and many contact sexual offenses involve related child victims; the proportion of related victims increases as victim age decreases, reflecting the greater access of family members to very young children (Snyder, 2000). This may reflect both a reporting and detection bias, in which victims are more likely to disclose regarding known perpetrators, especially related perpetrators, and it is easier to prosecute a case involving a known perpetrator than when victims do not know or are barely acquainted with the perpetrators. As I discuss in more detail in Chapter 6 , the differences between offenders with related versus unrelated child victims are meaningful (Seto, Babchishin, Pullman, & McPhail, 2015). Briefly here, those with related child victims were less likely to be pedohebephilic and less antisocial in terms of criminal history. In a subsequent meta-analysis, Pullman, Sawatsky, Babchishin, McPhail, and Seto (2017) also found differences between incest offenders who are genetically versus sociolegally related to their child victims.

Of particular importance for this book are the major differences between pedophilic and nonpedophilic sex offenders, and between hebephilic and nonhebephilic offenders. As discussed in Chapter 2 , pedophilic offenders are more likely to have boy victims, multiple child victims, prepubescent victims, and unrelated victims. Conversely, nonpedophilic offenders are more likely to have only single girl victims who are older and related to them. Comparisons of pedophilic and nonpedophilic sex offenders against children suggest that pedophilic offenders are less antisocial, for example, scoring lower on measures of psychopathy (Strassberg, Eastvold, Wilson Kenney, & Suchy, 2012), and less likely or no different in showing signs of executive dysfunction (Schiffer & Vonlaufen, 2011; Suchy, Whittaker, Strassberg, & Eastvold, 2009). These results are consistent with the idea that nonpedophilic sex offenders tend to be higher on facilitation factors because they do not have pedophilia as a motivation for sexually offending against children (see the next chapter ). In follow-up studies, pedophilic sex offenders against children are more likely to sexually reoffend (Hanson & Morton-Bourgon, 2005; see also Chapter 7 , this volume).

Other plausible differences between pedophilic and nonpedophilic offenders are potentially confounded. For example, one might think that duration of offending is another pedophilia indicator, where those who offend for longer periods of time are more likely to be pedophilic (see cases involving teachers or coaches or priests who offend over decades). But duration of offending is confounded with incest offending, where incest offenders may perpetrate multiple offending acts over long periods of time, given their ongoing access to the child victims, whereas extrafamilial offenders might have only a single contact with a particular victim.

As I discuss further in the next chapter , an important theoretical and practical question is whether sexual offending against children is part of a broader pattern of criminality or whether unique factors and mechanisms that underlie sexual offending against children need to be considered. Evidence supports both specialist and generalist explanations. Some perpetrators are specialists, committing only sexual offenses against children, perhaps only children of the same gender and similar age/maturity. Some sex offenders against children have committed only sexual crimes and otherwise live prosocial lives, for example, maintaining steady employment and stable relationships with family and others in their community. On the other hand, many sex offenders against children also have committed nonsexual offenses such as assaults, theft, and drug-related offenses (e.g., Lin & Simon, 2016).

Some offenders have victimized children of both genders, children of different ages, both children and adults, or have also committed noncontact sexual offenses involving exhibitionism or voyeurism (Abel, Becker, Cunningham-Rathner, Mittelman, & Rouleau, 1988; Bradford, Boulet, & Pawlak, 1992; Heil, Ahlmeyer, & Simons, 2003). For example, over a quarter (28%) of Abel et al.’s (1988) sample of offenders against children had both boy and girl victims, and 8% had also offended against adults. More recent studies show higher levels of overlap, especially when undetected sexual offenses are included (Cann, Friendship, & Gozna, 2007; Heil et al., 2003). This evidence of so-called crossover offending belies the idea that someone who only has known offenses against girls ages 10 to 12 poses no risk to younger or older girls, or no risk to boys; although it is more likely that any new victim would be a prepubescent to pubescent girl, other children might still be at risk. At a minimum, this points to the existence of both generalist and specialist sex offenders, which has implications for theories and interventions. D. A. Harris, Mazerolle, and Knight (2009) found that offenders against children tend to be specialists, in contrast to offenders against adults, who tend to be generalists. The relevance of sexual and nonsexual offending to risk for sexual recidivism is discussed in Chapter 7 , this volume. Characteristics of sexual offenses against children, and how these characteristics might be relevant to prevention efforts, are discussed again in Chapter 8 .

I recently synthesized the literature on online sexual offending in Seto (2013), and readers are encouraged to look at that book for a broader and deeper discussion of the theoretical and empirical questions. The understanding of online sexual offending has grown quickly over the past decade, reflecting the fact that more (sexual) offending involves online technologies in the developed world, just as more of people’s lives are now online. Nonetheless, I focused on child pornography offending and sexual solicitation of minors because almost no research has been conducted on Internet-facilitated sexual offending against adults, other illegal pornography offending, online offending involving exhibitionistic or voyeuristic behavior, or sex trafficking. Dozens of studies have examined questions regarding online offender characteristics, contact offending history, and sexual recidivism.

An informative, recent summary of research on online offender characteristics was provided by Babchishin et al. (2015). Most notable here is that an even higher proportion of child pornography offenders are male (99% or higher) than contact sex offenders against children (80%–90%) or offenders more generally (Babchishin, Hanson, & Hermann, 2011)—all of which indicates that having pedophilia, being online, and seeking child pornography is even more male-typical than other sexual offending against children. Studies of identified child pornography offenders find many show pedophilia and they are more likely to show pedophilia or hebephilia than contact sex offenders, with dual offenders having both child pornography and contact sexual offenses being the most likely to have pedophilia (Babchishin et al., 2015).

Studying the involvement of children and youth in sex work could reveal something about men who are sexually interested in young people and act on these interests by offering money, drugs, or other goods in exchange for sex. I noted the absence of good data on this social problem in the 2008 edition of this book; it is unfortunate that not much has changed since then, except for a change in language and conceptualization from referring to child or juvenile prostitution to referring to the sexual trafficking of minors.

Commercial sex work in North America involving children below the age of 12 is rare, compared with sex work involving minors between the ages of 12 and 17, as indicated in North American surveys, which shows that the average age of entry into sex work is in adolescence or early adulthood; however, it must be noted that these data are 20 or more years old (McClanahan, McClelland, Abram, & Teplin, 1999; Potterat, Rothenberg, Muth, Darrow, & Phillips-Plummer, 1998; Silbert & Pines, 1983). It is possible that the age of entry is lower in countries in which child labor is more common and in which poverty increases the likelihood that children will enter prostitution or be forced into it by others, but the international data have huge gaps.

ECPAT International is an international nonprofit organization concerned with commercial child sexual exploitation, including child pornography, child prostitution, and the trafficking of children for sexual purposes ( http://www.ecpat.net ). The organization is busy on the international policy landscape with regional and country reports and collaborations, but it is very difficult to drill down because of the few substantiated numbers. The reports tend to rely on anecdotes and expert estimates that are necessarily noisy, which is acknowledged, for example, in ECPAT’s (2016) report, “Offenders on the Move.” Advances in this area face many challenges, including an absence of systematic research data collection infrastructure in the poorest places, the fact that the poorest and most vulnerable children are often “invisible,” and government and other interest groups may minimize or exaggerate the extent of the problem. The ECPAT reports indicate some consensus about general trends, such as which countries are source versus destination countries (or both) and the influence of organized crime, but there is very little to draw upon regarding the estimated numbers of adults and minors involved, the ages of minors, or the characteristics of these interactions.

Half of Bernard’s (1979/1985) sample of self-identified pedophilic individuals, surveyed in the Netherlands, reported that they had traveled outside of the country to have sexual contact with a child, elsewhere in Europe or in northern Africa. Similarly, three decades later, some of Houtepen et al.’s (2016) 15 respondents reported interest in child sex tourism. This in turn may be related to a public tendency to view sexual crimes committed elsewhere as less serious than the same crimes committed in one’s own country (Kosuri & Jeglic, 2017).

In summary, reports of men who seek sexual contacts with minors through prostitution are still mostly anecdotal and estimates are of unknown validity, with insufficient data about prevalence, client characteristics, and the age and maturity status of the youth who are involved. What little researchers do know suggests the problem of child sex tourism and prostitution is mostly one that involves underage adolescents rather than prepubescent or pubescent children.

Unfortunately, little has changed about our understanding of female pedophilia since the first edition of this book. Only a few case examples or series have been described in the peer-reviewed literature; these cases involve female sex offenders meeting the diagnostic criteria for pedophilia. Fedoroff et al. (1999) described several women who reported having sexual fantasies about children or reported feeling sexually aroused while engaging in sexual acts with children. For example, one woman admitted feeling sexually aroused while play-wrestling with her 8-year-old nephew, and another woman admitted she sexually fantasized about her 1- and 3-year-old sons, reporting herself to a child protection service after she fellated them. Chow and Choy (2002) described a woman who offended against two girls she was babysitting, both age 4, when she was 18 and 23 years old, respectively. This woman admitted having sexual fantasies about girls who were ages 3 or 4 since she was a child herself (as well as sexual fantasies about adult women and men) and admitted feeling sexually aroused while bathing the girls. She denied any sexual interest, however, in her two young sons.

Even less is known about female persons with pedophilia who are not already known to the criminal justice system. Fromuth and Conn (1997) conducted a survey of 546 female college students regarding their lifetime sexual experiences with children who were at least 5 years younger than they were at the time. Four percent acknowledged having at least one such experience with a child, usually (92%) involving kissing and sexual touching. The average age of the respondent at the time was 12, and the average age of the child was 6. Contrary to the data available for identified female sex offenders, which finds girls are more likely to be victims, the majority (70%) of the children these women described were boys. None of these incidents were detected by the police or by child protection services. Only 18% reported any sexual attraction or sexual fantasies about children, however, which can be compared with 5% of the women who reported no sexual contacts with children. Last, of the 799 women who responded to Joyal, Cossette, and Lapierre’s (2015) survey, 0.8% responded yes to the statement “I have fantasized about having sex with a child under the age of 12 years”; 1.8% of 717 men responded yes to this item. The mean intensity of interest rating was also less than half that of men.

In contrast to the literature on pedophilia in women, the literature on female sexual offending has grown substantially in the past decade. Indeed, as Gannon and Cortoni (2010) noted in their introduction to an edited volume on female sex offenders, their book would not have been possible a decade earlier. This literature has focused on two central questions: (a) To what extent do female sex offenders differ from male sex offenders, with implications for assessment, treatment, and management? and (b) To what extent do female sex offenders differ from other women, with implications for etiology and for developmental trajectories? D. A. Harris (2010) noted that a common starting point is to apply male models of sexual offending to women, which has not been particularly successful because research suggests different factors are involved. For example, Schatzel-Murphy, Harris, Knight, and Milburn (2009) found that male perpetration of sexually coercive behavior was explained well by hypermasculinity and impersonal sex, in line with Malamuth’s (2003) confluence model to explain sexual offending by men against women, whereas female perpetration was not explained by these factors but by sexual compulsivity.

Typologies of female sexual offending are similar across different schemes and clearly differ from typologies developed for male sexual offending (e.g., Sandler & Freeman, 2007; Vandiver & Kercher, 2004). One of these categories is women who sexually offend against adolescent boys and who tend not to consider their criminal sexual behavior to be exploitative or abusive; another is women who offend against young children, typically alone and typically in their own or the child’s residence. This latter group is purportedly the most likely to have paraphilic sexual fantasies and to use coercion or force. The third category is women who offend with a male perpetrator, either as co-offenders or because they were coerced by the male; scholars have characterized these women as unassertive, dependent, and subscribing to stereotypical gender roles. The last group is women who offend against other adults, usually women. This is the least common type, representing only 8% of the cases reviewed by Vandiver and Kercher (2004). Notably, none of these models offer prominent roles to paraphilic motivations such as pedophilia or hebephilia.

Sexual recidivism rates are low for female sexual offenders. In a meta-analysis of 10 studies comprising a total of 2,490 women offenders, followed for an average of 6.5 years, less than 3% committed a new sexual offense (Cortoni, Hanson, & Coache, 2010). Because of this very low base rate in conjunction with the smaller number of identified female sex offenders, it is methodologically difficult to empirically identify risk factors for sexual recidivism and to develop validated risk assessment measures (see Chapter 7 ). Nonetheless, some research indicates that criminal history matters, which is consistent with the literature showing that male and female offenders have many risk factors in common (Andrews & Bonta, 2010).

The gender difference in sexual offending is very large, with women accounting for less than 5% of sexual offenders, based on reviews of official police records across multiple countries (Cortoni, Babchishin, & Rat, 2017). Cortoni et al. (2017) examined 17 samples from 12 countries and found the gender difference was larger for official police reports than for self-report surveys (2.2% vs. 11.6%), for juvenile perpetration than for adult perpetration (5.2% vs. 3.3%), and for male versus female victims (40% vs. 4%). The proportions of female sex offenders may also be higher when focusing on those who have sexually offended against children, but this was not reported in the Cortoni et al. meta-analysis.

The gap in the prevalence of female sexual offending when estimated from victimization data versus known offender data, especially for offenses against males (see also Saradjian, 2010), is intriguing. The likeliest explanation is that younger adolescent boys who are victimized by women rarely report to police, or if they do, they are not believed or taken seriously. The gender difference is even starker for child pornography offending, with women representing 1% or fewer of all known cases (Babchishin et al., 2011). This large gender difference needs to be understood within the context of large gender differences in paraphilias, sexual risk taking, criminal behavior, and involvement with mainstream visual pornography.

Paraphilic motivations play a less prominent role in sexual offending by women, compared with men, and co-offending with a male perpetrator—sometimes as an accomplice, sometimes under coercion—plays a bigger role. Nonetheless, the majority of women who sexually offend act on their own, with most of their victims being children or young adolescents under their care. Girl victims are more common than boy victims, and girl victims tend to be younger than boy victims—the same patterns observed for male offenders against children (Lewis & Stanley, 2000).

Wiegel, Abel, and Jordan (2003) examined questionnaire data from an American sample of 242 women who admitted having committed a sexual offense. The majority (70%) had sexually offended against a child (ages unspecified); the rest had engaged in obscene telephone calls or acts involving bestiality, exhibitionism, or voyeurism. A slight majority of women acknowledged having a single victim, and a small minority (10%) of the women admitted to sexual offenses against five or more children. The average age of sexual offending onset was 19. Approximately half of the women had offended against a related child, either their own child (17%) or a niece or nephew (35%). Only 4% of the women had sexually offended against a stranger. Approximately one third of the women reported being sexually aroused by male or female children, with slightly more admitting to a sexual interest in boys than in girls; relative sexual arousal to children versus adults was not reported, however, and neither was intensity nor recurrence, so whether these women might have pedophilia is not known. More respondents were sexually aroused by male or female adolescents and would therefore not qualify for either pedophilia or hebephilia.

Cooper, Swaminath, Baxter, and Poulin (1990) described the results of a psychophysiological assessment of a female sex offender with multiple child victims. She did not discriminate between sexual stimuli depicting children or adults using a measure of genital sexual arousal in women (vaginal photoplethysmography, based on relative color change because of vaginal vasocongestion). She also did not distinguish between depictions of coercive or consensual sex. It is unlikely that female sex offenders with child victims would show the same patterns of sexual response as their male counterparts, given the results of research by Chivers and colleagues demonstrating that vaginal responding is not category specific to audiovisual stimuli for exclusively heterosexual women (Chivers, 2003, 2017; Chivers & Bailey, 2005; Chivers, Rieger, Latty, & Bailey, 2004). Moreover, the gender difference in the correspondence of subjective and genital measures of sexual arousal is robust, with women showing lower (but still positive) concordance when compared with men (Chivers, Seto, Lalumière, Laan, & Grimbos, 2010). At the same time, category specificity is exhibited by women with any degree of gynephilia, with lower intensity sexual stimuli, such as stories, or with stimuli depicting purely masochistic interactions versus masochistic sex interactions (see Chivers, 2017).

There is a growing literature on adolescents who sexually offend. Like the adult literature, most attention has been paid to male youth. The majority of adolescents who sexually offend do so against younger children rather than peers or adults, and evidence suggests meaningful differences based on victim age that are similar to those found when comparing adult sex offenders with child victims to those with adult victims (Lalumière, Harris, Quinsey, & Rice, 2005; Leroux, Pullman, Motayne, & Seto, 2016; Seto, 2008).

A complication in considering this distinction of offenses based on victim age, which is relevant to the question of potential pedophilic or hebephilic motivations, is that definitions of children are inconsistent across studies, with some simply specifying a maximum child age, others defining a relative age difference between perpetrator and victim, and some using both a maximum age and age difference. For an adult offender, one can simply define child victims as any victim under a specified age, using a legally determined criterion or a criterion based on interest in identifying pedophilic or hebephilic offenders. But for a juvenile perpetrator, a 12-year-old victim would be a peer if he was 13, or a distinctly younger child if he was 17. Fortunately, Kemper and Kistner (2007) conducted a study to examine the impact of six different definitions of child versus peer/adult victims. They found that only 28% of juveniles who sexually offend changed categories (child vs. peer/adult) when definitions were changed and that this had little impact on overall results. In other words, offenders against children using one definition were often still classified as offenders against children using a different definition.

Keelan and Fremouw (2013) reviewed 21 studies that compared adolescents who offended against children and those who offended against peers or adults. Their impressions were that juveniles who sexually offended against younger children were less likely to use force, more likely to have related victims, more likely to have male victims, more likely to have experienced sexual abuse, and more likely to exhibit psychopathology. It was unclear if they differed in family background or in recidivism rates, which are quite low for juveniles who have sexually offended (Caldwell, 2002, 2010, 2016). Notably, the reviewed studies either did not include data on atypical sexual interests or did not do so in a way that allowed Keelan and Fremouw to draw any conclusions about this very important domain in the adult offender literature. Leroux et al. (2016) compared adolescents who offended against children, peers/adults, or both, and found that those who offended against both children and peers/adults were the most likely to admit atypical sexual fantasies or to be clinically identified as having a sexual problem.

Clinicians and researchers seem to be reluctant to consider paraphilic motivations for sexual offending as potentially important for some youth, perhaps out of concerns about the negative impacts of labeling, in conjunction with recognition that the current diagnostic schemes exclude younger adolescents. The available data suggest that some adolescents do show evidence of sexual interest in prepubescent or pubescent children. For example, Seto, Lalumière, and Blanchard (2000) showed that adolescents who sexually offended against any boys showed relatively greater sexual arousal to child stimuli than those who offended only against girls. Moreover, the Screening Scale for Pedophilic Interests (SSPI; Seto & Lalumière, 2001; see also Chapter 2 , this volume) operated as expected among male juveniles who have sexually offended, except that number of child victims was a better indicator of sexual response to child stimuli than having boy victims (Seto, Murphy, Page, & Ennis, 2003) and a smaller proportion of adolescents showed a preference for child stimuli in the phallometric laboratory than adult sex offenders against children (Seto & Lalumière, 2001). In their study of 485 male juveniles who had sexually offended, Zolondek, Abel, Northey, and Jordan (2001) found that many admitted to a range of paraphilic behaviors, suggesting again that a minority had paraphilia-like sexual interests, even if they would not meet criteria from the Diagnostic and Statistical Manual of Mental Disorders (DSM ; fifth ed.; American Psychiatric Association, 2013) because of their age. Last, Worling and Curwen (2000) found that self-reported sexual interest in children predicted sexual recidivism in their sample of treated juveniles who sexually offended.

We concluded from our meta-analysis that paraphilic sexual interests are a distinctive feature of male youths who sexually offend, although they are present for only a minority of juveniles (Seto & Lalumière, 2010). Some adolescents discover that they are sexually attracted to children with the onset of puberty, just as many other adolescents are discovering they are sexually attracted to their peers or adults, and realizing as well that they are attracted to the same gender, a different gender, nonbinary persons or no gender (see Seto, 2012, 2017b).

The public is beginning to recognize the import of this discovery: There was a strong reaction to the story of Adam, a 16-year-old male who realized he was sexually attracted to children and talked about it with reporter Luke Malone (2014a). Others have talked about becoming aware of their sexual interest in children as adolescents on websites such as Virtuous Pedophiles. This is an exciting new area for discovery and prevention, including the Help Wanted Project for youth from the Moore Center for the Prevention of Child Sexual Abuse at Johns Hopkins University and a pilot Dunkelfeld Project for juveniles who are concerned about their sexual interest in children (Beier et al., 2016; Shields & Ruzicka, 2016).

Frey (2010) reviewed what is known about girls who sexually offend. Girls account for 8% of juvenile arrests for sexual offending and approximately 20% of all known female sexual offending. It is difficult to say much about girls who sexually offend because most research consists of case reports and small descriptive studies. It is also hard to know what is distinctive about this group because of a lack of comparison studies, such as comparisons of girls who sexually versus nonsexually offend. Frey (2010) found that the median age of offending youth is 11 or 12, and victims are usually children (boys or girls) under the age of 6. The most common sexual acts involve touching.

Because each study group—self-identified persons with pedophilia, persons with pedophilia seen in clinical settings, sex offenders in criminal justice settings—has its advantages and disadvantages, this book draws from all the relevant research on pedophilia. This work now spans the past half century, with seminal work by Kurt Freund and others in the 1960s, preceded by data on adult–child sex from the Kinsey reports and by case studies from von Krafft-Ebing (1906/1999) a century ago. Convergence in findings would increase researchers’ confidence that the knowledge that has been gained is generalizable and not an artifact of clinical or criminal status, or of self-report bias, and not confined to a particular study group. Focusing on the phenomenology of pedophilia, evidence is converging regarding its age of onset, stability over time, and role in orienting sexual, romantic, and social attraction.

The study groups described in this chapter are quite different in some important ways. They differ in the likelihood that group members have pedophilia or hebephilia, from all self-identified pedophilic or hebephilic individuals, to majorities of clinically seen individuals and child pornography offenders, to approximately half of contact sex offenders, and finally to minorities of solicitation and statutory sex offenders. They differ in the likelihood they have committed sexual offenses involving children, in the reverse order they were just listed, from some self-identified pedophilic individuals and clinical cases to all identified sex offenders. And they differ in their antisociality, probably in the same order as for detected sexual offending, which is highest for contact sex offenders, then solicitation and statutory offenders, then child pornography offenders. Many self-identified pedophilic and hebephilic individuals have committed undetected sexual offenses; it is unclear how antisocial they are, on average, because studies with relevant measures have not been reported yet.

These group differences have major implications for policies and practices, which can be misguided to the extent that groups are confused with each other or combined. For example, policies and practices that view all sex offenders against children as having pedophilia will overestimate the risk posed by nonpedophilic sex offenders, all other things being equal, and might require nonpedophilic sex offenders to participate in interventions that do not make sense (e.g., aversive conditioning of pedophilic sexual arousal, sex-drive-reducing medication). As a second example, child pornography offenders do not pose the same risk of sexual recidivism as contact offenders, on average, so treating all child pornography offenders as either undetected or soon-to-be–contact offenders misuses scarce resources.

An unintended (and counterproductive) consequence of conflating pedophilia and sexual offending against children is that the understandable fear and anger elicited by men who commit sexual crimes against children is also directed at pedophilic or hebephilic individuals who have never acted upon their sexual interest in children. As I discuss in Chapter 9 , the powerful stigma associated with pedophilia prevents many from seeking support from their family or friends, and from seeking professional help. As a result, pedophilic or hebephilic individuals who might not offend if they received social and professional support are left to sink or swim, with potentially terrible consequences for themselves, for victimized children, and their families.

Each of the study groups described in this chapter has its advantages and disadvantages for empirical research on pedophilia and on sexual offending against children as well. Self-identified pedophilic or hebephilic individuals have not been selected for psychological distress in the way that clinical samples are or for antisocial behavior in the way that criminal samples are. On the other hand, self-identified pedohebephilic individuals may select themselves for impression management, such that participants in anonymous online surveys may represent themselves in an overly positive light. Because many offenses involving child pornography, online solicitation, or child sexual abuse are not reported and, thus, not all offenders against children are detected, research relying on identified samples of offenders is partially confounded by factors associated with detection. Perhaps the most striking finding in this regard is the big gap between interest in boys, as reported by a majority of self-identified pedophilic individuals, and interest in girls, as reflected in victimization data and identified offenders.

Each of the study groups and assessment methods described in the previous chapter reveals new knowledge. However, given the complexities that different recruitment strategies and target populations bring, I am particularly interested in the convergence of findings. Despite the great deal of heterogeneity within groups and differences between groups, they also have many similarities. Most known pedophilic individuals and most sex offenders are men, consistent with a generally greater prevalence of paraphilias among men and higher involvement in criminal behavior by men, respectively. Although the majority of child victims (exploited for child pornography or sexually abused in contact offenses) are girls, one can conclude that pedophilia is more strongly associated with an interest in boys. The features of children that are sexually attractive to pedophilic individuals—as reported by self-identified persons, revealed by descriptive analyses of child pornography images, and realized in the stimuli that elicit sexual arousal in phallometric and other laboratory testing—appear to be similar across study groups. The preferred physical features include small body shape and size, smooth skin, and other indicators of youthfulness; the psychological features include gentleness, innocence, playfulness, and openness (Freund, McKnight, Langevin, & Cibiri, 1972; Taylor, Holland, & Quayle, 2001; G. D. Wilson & Cox, 1983). Finally, one can conclude that although many pedophilic individuals have acted upon their sexual interest in children, some have not.

A promising approach to studying pedophilia is to use the Internet for participant recruitment and/or data collection. This approach has the strength of providing some assurance of anonymity, especially if combined with the use of technical options that anonymize Internet protocol addresses, the use of free web-based email accounts for communication, and file encryption. The main limitation of Internet research is the possibility of self-report biases, because pedophilic respondents might alter their reports of problems to present themselves in a socially desirable manner, and anonymity precludes verifying the self-reports. For example, Riegel (2004) found that 87% of his respondents reported that their interest in mentoring a boy was equal or more important than their interest in sex. This may be a true reflection of their interests, but these statements cannot be fully accepted without speaking to the boys as well. Another methodological limitation is that some respondents may not actually be persons with pedophilia and may instead participate out of curiosity or mischief. Nonetheless, researchers have opportunities to test hypotheses about pedophilia that are less susceptible to socially desirable responding. For example, as I discuss again in Chapter 5 , Blanchard et al. (2003) demonstrated that pedophilic sex offenders are significantly more likely than nonpedophilic sex offenders to report having a head injury resulting in unconsciousness before the age of 13 but not after the age of 13. It is unlikely that either reporting or denying a childhood head injury is socially desirable, so a future study of self-identified persons with pedophilia could be conducted to replicate this finding, which has implications for the understanding of the etiology of pedophilia.

Whatever research approach is used, comparing the findings obtained from different study groups is a productive strategy in identifying common and unique factors. Garber and Hollon (1991) discussed how group comparisons of this kind can be very informative. I discuss the results of group comparisons in the next two chapters, first on the origins of sexual offending against children (by comparing sex offenders with other offenders and then sex offenders against children with sex offenders against adults) and then on the etiology of pedophilia (by comparing male persons with pedophilia with other men).