Prof. Herek's blog

   Subscribe to
Prof. Herek's blog

Open bibliography
in a separate
window

In a similar fashion, gay people have often been portrayed as a threat to children. Back in 1977, when Anita Bryant campaigned successfully to repeal a Dade County (FL) ordinance prohibiting anti-gay discrimination, she named her organization "Save Our Children," and warned that "a particularly deviant-minded [gay] teacher could sexually molest children" (Bryant, 1977, p. 114). [Bibliographic references are on a different web page]

In recent years, antigay activists have routinely asserted that gay people are child molesters. This argument was often made in debates about the Boy Scouts of America's policy to exclude gay scouts and scoutmasters. More recently, in the wake of Rep. Mark Foley's resignation from the US House of Representatives in 2006, antigay activists and their supporters seized on the scandal to revive this canard.

It has also been raised in connection with scandals about the Catholic church's attempts to cover up the abuse of young males by priests. Indeed, the Vatican's early response to the 2002 revelations of widespread Church cover-ups of sexual abuse by priests was to declare that gay men should not be ordained.

By contrast, in a 1999 national poll, the belief that most gay men are likely to molest or abuse children was endorsed by only 19% of heterosexual men and 10% of heterosexual women. Even fewer – 9% of men and 6% of women – regarded most lesbians as child molesters.

Consistent with these findings, Gallup polls have found that an increasing number of Americans would allow gay people to be elementary school teachers. For example, the proportion was 54% in 2005, compared to 27% in 1977.
 

One problem is that none of the studies in this area have obtained data from a probability sample, that is, a sample that can be assumed to be representative of the population of all child molesters. Rather, most research has been conducted only with convicted perpetrators or with pedophiles who sought professional help. Consequently, they may not accurately describe child molesters who have never been caught or have not sought treatment.
 

Pedophilia and child molestation are used in different ways, even by professionals. Pedophilia usually refers to an adult psychological disorder characterized by a preference for prepubescent children as sexual partners; this preference may or may not be acted upon. The term hebephilia is sometimes used to describe adult sexual attractions to adolescents or children who have reached puberty.

Whereas pedophilia and hebephilia refer to psychological propensities, child molestation and child sexual abuse are used to describe actual sexual contact between an adult and someone who has not reached the legal age of consent. In this context, the latter individual is referred to as a child, even though he or she may be a teenager.

Although the terms are not always applied consistently, it is useful to distinguish between pedophiles/hebephiles and child molesters/abusers. Pedophilia and hebephilia are diagnostic labels that refer to psychological attractions. Not all pedophiles and hebephiles actually molest children; an adult can be attracted to children or adolescents without ever actually engaging in sexual contact with them.

Child molestation and child sexual abuse refer to actions, and don't imply a particular psychological makeup or motive on the part of the perpetrator. Not all incidents of child sexual abuse are perpetrated by pedophiles or hebephiles; in some cases, the perpetrator has other motives for his or her actions and does not manifest an ongoing pattern of sexual attraction to children.

Thus, not all child sexual abuse is perpetrated by pedophiles (or hebephiles) and not all pedophiles and hebephiles actually commit abuse. Consequently, it is important to use terminology carefully.

Another problem related to terminology arises because sexual abuse of male children by adult men² is often referred to as "homosexual molestation." The adjective "homosexual" (or "heterosexual" when a man abuses a female child) refers to the victim's gender in relation to that of the perpetrator. Unfortunately, people sometimes mistakenly interpret it as referring to the perpetrator's sexual orientation.

As an expert panel of researchers convened by the National Academy of Sciences noted in a 1993 report: "The distinction between homosexual and heterosexual child molesters relies on the premise that male molesters of male victims are homosexual in orientation. Most molesters of boys do not report sexual interest in adult men, however" (National Research Council, 1993, p. 143, citation omitted).

To avoid this confusion, it is preferable to refer to men's sexual abuse of boys with the more accurate label of male-male molestation. Similarly, it is preferable to refer to men's abuse of girls as male-female molestation. These labels are more accurate because they describe the sex of the individuals involved but don't implicitly convey unwarranted assumptions about the perpetrator's sexual orientation.
 

Over the years, this fact has been incorporated into various systems for categorizing child molesters. For example, Finkelhor and Araji (1986) proposed that perpetrators' sexual attractions should be conceptualized as ranging along a continuum – from exclusive interest in children at one extreme, to exclusive interest in adult partners at the other end.

Typologies of offenders have often included a distinction between those with an enduring primary preference for children as sexual partners and those who have established age-appropriate relationships but become sexually involved with children under unusual circumstances of extreme stress. Perpetrators in the first category – those with a more or less exclusive interest in children – have been labeled fixated. Fixation means "a temporary or permanent arrestment of psychological maturation resulting from unresolved formative issues which persist and underlie the organization of subsequent phases of development" (Groth & Birnbaum, 1978, p. 176). Many clinicians view fixated offenders as being "stuck" at an early stage of psychological development.

By contrast, other molesters are described as regressed. Regression is "a temporary or permanent appearance of primitive behavior after more mature forms of expression had been attained, regardless of whether the immature behavior was actually manifested earlier in the individual's development" (Groth & Birnbaum, 1978, p. 177). Regressed offenders have developed an adult sexual orientation but under certain conditions (such as extreme stress) they return to an earlier, less mature psychological state and engage in sexual contact with children.

Some typologies of child molesters divide the fixation-regression distinction into multiple categories, and some include additional categories as well (e.g., Knight, 1989).

For the present discussion, the important point is that many child molesters cannot be meaningfully described as homosexuals, heterosexuals, or bisexuals (in the usual sense of those terms) because they are not really capable of a relationship with an adult man or woman. Instead of gender, their sexual attractions are based primarily on age. These individuals – who are often characterized as fixated – are attracted to children, not to men or women.

Using the fixated-regressed distinction, Groth and Birnbaum (1978) studied 175 adult males who were convicted in Massachusetts of sexual assault against a child. None of the men had an exclusively homosexual adult sexual orientation. 83 (47%) were classified as "fixated;" 70 others (40%) were classified as regressed adult heterosexuals; the remaining 22 (13%) were classified as regressed adult bisexuals. Of the last group, Groth and Birnbaum observed that "in their adult relationships they engaged in sex on occasion with men as well as with women. However, in no case did this attraction to men exceed their preference for women....There were no men who were primarily sexually attracted to other adult males..." (p.180).
 

In yet another approach to studying adult sexual attraction to children, some Canadian researchers observed how homosexual and heterosexual adult men responded to slides of males and females of various ages (child, pubescent, and mature adult). All of the research subjects were first screened to ensure that they preferred physically mature sexual partners. In some of the slides shown to subjects, the model was clothed; in others, he or she was nude. The slides were accompanied by audio recordings. The recordings paired with the nude models described an imaginary sexual interaction between the model and the subject. The recordings paired with the pictures of clothed models described the model engaging in neutral activities (e.g., swimming). To measure sexual arousal, changes in the subjects' penis volume were monitored while they watched the slides and listened to the audiotapes. The researchers found that homosexual males responded no more to male children than heterosexual males responded to female children (Freund et al., 1989).

In summary, each of these studies failed to support the hypothesis that homosexual males are more likely than heterosexual men to molest children or to be sexually attracted to children or adolescents.
 

Are homosexual adults in general sexually attracted to children and are preadolescent children at greater risk of molestation from homosexual adults than from heterosexual adults? There is no reason to believe so. The research to date all points to there being no significant relationship between a homosexual lifestyle and child molestation. There appears to be practically no reportage of sexual molestation of girls by lesbian adults, and the adult male who sexually molests young boys is not likely to be homosexual (Groth & Gary, 1982, p. 147).

In a later literature review, Dr. Nathaniel McConaghy (1998) similarly cautioned against confusing homosexuality with pedophilia. He noted, "The man who offends against prepubertal or immediately postpubertal boys is typically not sexually interested in older men or in women" (p. 259).

This well known lack of a linkage between homosexuality and child molestation accounts for why relatively little research has directly addressed the issue. For example, a 1998 comprehensive review of published empirical research on the sexual abuse of boys reported only one study (the 1994 study by Jenny and colleagues, cited above) that included data about the self-reported sexual orientation of perpetrators (Holmes & Slap, 1998).

Proving something that is already widely known simply isn't a priority for scientists. Indeed, a commentary that accompanied publication of the study by Jenny et al. in Pediatrics noted that debates about gay people as molesters "have little to do with everyday child abuse" and lamented that they distract lawmakers and the public from dealing with the real problem of children's sexual mistreatment (Krugman, 1994).
 

These are cases in which the term pedophilia – referring as it does to attractions to prepubescent children – can cause confusion. Rather than pedophilia, the accusations stemming from these scandals raised the question of whether gay people shouldn't be trusted in positions of authority where there is any opportunity for sexually harassing or abusing others.

Here again, there is no inherent connection between an adult's sexual orientation and her or his propensity for endangering others. Scientific research provides no evidence that homosexual people are less likely than heterosexuals to exercise good judgment and appropriate discretion in their employment settings. There are no data, for example, showing that gay men and lesbians are more likely than heterosexual men and women to sexually harass their subordinates in the workplace. Data from studies using a variety of psychological measures do not indicate that gay people are more likely than heterosexuals to possess any psychological characteristics that would make them less capable of controlling their sexual urges, refraining from the abuse of power, obeying rules and laws, interacting effectively with others, or exercising good judgment in handling authority. As explained elsewhere on this site, sexual orientation is not a mental illness nor is it inherently associated with impaired psychological functioning.

Gay men and lesbians function effectively in a wide variety of employment settings. The research literature doesn't reveal any differences between heterosexuals, bisexuals, and homosexuals in job performance or ability to properly exercise authority in supervisory roles. As indicated by workplace policies around the United States, a large and growing number of private and public employers do not perceive a problem with hiring gay and bisexual people as employees or managers. Many corporations, educational institutions, and local governments have adopted policies that prohibit discrimination against employees on the basis of sexual orientation. Many of those organizations provide benefits such as health insurance for employees' same-sex partners. Indeed, one widely cited reason for offering such benefits is that they enable a company to remain competitive by attracting high quality employees who happen to be gay, lesbian, or bisexual.

Thus, there is no factual basis for organizations to avoid hiring homosexual or bisexual people, simply on the basis of their sexual orientation, for positions that involve responsibility for or supervision of others, whether children, adolescents, or adults.
 

Some conservative groups have argued that scientific research strongly supports their claims that homosexuality and pedophilia are linked. The Family Research Council has produced what is perhaps the most extensive attempt to document this claim. It is an article by Timothy J. Dailey titled Homosexuality and Child Abuse.

With 76 footnotes, many of them referring to papers in scientific journals, it appears at first glance to be a thorough and scholarly discussion of the issue. On further examination, however, its central argument – that "the evidence indicates that homosexual men molest boys at rates grossly disproportionate to the rates at which heterosexual men molest girls" – doesn't hold up.

In the following section, the main sources cited by Dailey and the FRC to support their claim are reviewed. The papers are listed in the same order in which they are first cited by the FRC article.

1. Freund et al. (1989). Heterosexuality, homosexuality, and erotic age preference. Journal of Sex Research, 26, 107-117.

   This article is discussed above in the "Other Approaches" section. As the FRC concedes, it contradicts their argument. The abstract summarizes the authors' conclusion: "Findings indicate that homosexual males who preferred mature partners responded no more to male children than heterosexual males who preferred mature partners responded to female children."

    

2. Silverthorne & Quinsey. (2000). Sexual partner age preferences of homosexual and heterosexual men and women. Archives of Sexual Behavior, 29, 67-76.

   The FRC cites this study to challenge the Freund et al. data (see the previous paper above). However, the methodologies were quite different. Freund and his colleagues used a sample that included sex offenders and they assessed sexual arousal with a physiological measure similar to that described below for the 1988 Marshall et al. study. Silverthorne and Quinsey used a sample of community volunteers who were asked to view pictures of human faces and use a 7-point scale to rate their sexual attractiveness. The apparent ages of the people portrayed in the pictures was originally estimated by Dr. Silverthorne to range from 15 to 50. However, a group of independent raters perceived the male faces to range in age from 18 to 58, and the female faces to range from 19 to 60.

   The article doesn't report the data in great detail (e.g., average ratings are depicted only in a graphic; the actual numbers aren't reported) and the authors provide contradictory information about the rating scale (they describe it as a 7-point scale but also say it ranged from 0 to 7, which constitutes an 8-point scale). In either case, it appears that none of the pictures was rated as "very sexually attractive" (a rating of 7). Rather, the highest average ratings were approximately 5.

   On average, gay men rated the 18-year old male faces the most attractive (average rating = about 5), with attractiveness ratings declining steadily for older faces. They rated the 58-year old male faces 2, on average. By contrast, heterosexual men rated the 25-year old female faces the most attractive (about 5), with the 18- and 28-year old female faces rated lower (between 2 and 3) and the 60-year old female faces rated the least attractive (about 1).

   A serious problem with this study is that the researchers didn't control for the possibility that some of the faces pictured in the photos might simply have been more or less physically attractive than the others, independent of their age or gender. The researchers explicitly acknowledged this shortcoming, speculating that the women's faces in the 25-year old group might have been more attractive than women's faces in the other age groups. But they didn't address the possibility that the attractiveness of the male and female faces may not have been comparable.

   This issue could have been addressed in various ways. For example, prior to collecting data, the researchers could have started with a large number of photographs and asked a group of independent raters to evaluate the general physical attractiveness of the face in each photo; these ratings could have been used to select photos for the experiment that were equivalent in attractiveness. Getting independent ratings of experimental stimuli in this way is a common procedure in social psychological research.

   Thus, even if one accepts the questionable assumption that this study is relevant, it doesn't support the FRC's contention that gay men are more likely than heterosexual men to be child molesters for several reasons:

   • the researchers failed to control for the varying attractiveness of the different photos;
   • all of the faces portrayed in the photos were perceived to be at least 18; and
   • the study merely assessed judgments of sexual attractiveness rather than the research participants' sexual arousal.

    

3. Blanchard et al. (2000). Fraternal birth order and sexual orientation in pedophiles. Archives of Sexual Behavior, 29, 463-478.

   This study categorized convicted sex offenders according to whether they molested or reported sexual attraction to boys only, girls only, or both boys and girls. These groups were labeled, respectively, homosexual pedophiles, heterosexual pedophiles, and bisexual pedophiles. This classification referred to their attractions to children. Adult sexual orientation (or even whether the men had an adult sexual orientation) wasn't assessed.

    

4. Elliott et al. (1995). Child sexual abuse prevention: What offenders tell us. Child Abuse & Neglect, 19, 579-594.

   In this study, child sex offenders were interviewed. Their sexual orientation (gay, heterosexual, bisexual) wasn't assessed. The authors drew from their findings to suggest strategies for how parents and children can prevent sexual victimization. It is noteworthy that none of those strategies involved avoiding gay men.

    

5. Jenny et al. (1994). Are children at risk for sexual abuse by homosexuals? Pediatrics, 94, 41-44.

   This study, described above in the section on "Other Approaches," contradicts the FRC's argument. The FRC faults the study because the researchers didn't directly interview perpetrators but instead relied on the victims' medical charts for information about the offender's sexual orientation. However, other studies cited favorably by the FRC (and summarized in this section) similarly relied on chart data (Erickson et al., 1988) or did not directly assess the sexual orientation of perpetrators (Blanchard et al. 2000; Elliott et al. 1995; Marshall et al., 1988). Thus, the FRC apparently considers this method a weakness only when it leads to results they dislike.

    

6. Marshall et al. (1988). Sexual offenders against male children: Sexual preference. Behaviour Research and Therapy, 26, 383-391.

   In this study, the researchers compared 21 men who had sexually molested a male under 16 years (and at least 5 years younger than themselves) to 18 unemployed men who were not known to have molested a child. Over a series of sessions, each man watched color slides of nude males and females of various ages and listened to audiotaped descriptions of both coercive and consensual sexual interactions between a man and a boy. During the sessions, each man sat in a private booth, where he was instructed to lower his trousers and underwear and attach a rubber tube to his penis. The tube detected any changes in penis circumference, with increases interpreted as indicating sexual arousal.

   The FRC cites this study as showing that "a homosexual and a heterosexual subgroup can be delineated among these offenders." This is true but hardly relevant to their claims.

   The researchers categorized 7 offenders who were more aroused overall by the male nudes than the female nudes as the homosexual subgroup. They categorized 14 offenders who were more aroused overall by the female nudes as the heterosexual subgroup. The offenders were not asked their sexual orientation (gay, straight, bisexual) and the paper does not report any information about the nature of the offenders' adult sexual relationships, or even if they had any such relationships.

    

7. Bickley & Beech. (2001). Classifying child abusers: Its relevance to theory and clinical practice. International Journal Of Offender Therapy And Comparative Criminology, 45, 51-69.

   This is a literature review and theoretical paper that discusses the strengths and weaknesses of various systems for classifying child molesters. In citing this study, the FRC says it:

   refers to homosexual pedophiles as a "distinct group." The victims of homosexual pedophiles "were more likely to be strangers, that they were more likely to have engaged in paraphiliac behavior separate from that involved in the offence, and that they were more likely to have past convictions for sexual offences.... Other studies [showed a] greater risk of reoffending than those who had offended against girls" and that the "recidivism rate for male-victim offenders is approximately twice that for female-victim offenders."

   In reality, however, the paper was summarizing the findings of other studies, not reporting new data. In the passage excerpted by the FRC, the authors were discussing published papers that used a classification system focusing entirely on the sex of victims (not whether the perpetrator is straight or gay). Here is the complete text (the passages that FRC omitted are highlighted):

   "Grubin and Kennedy (1991) reported that when dividing sex offenders based simply on the sex of their victims, offenders against boys stood out as a distinct group. They noted that their victims were more likely to be strangers, that they were more likely to have engaged in paraphiliac behavior separate from that involved in the offence, and they were more likely to have past convictions for sexual offences. Other studies have employed the sex-of-victim approach in the prediction of future risk, with offenders who have sexually abused boys or both boys and girls reported as having more victims and being at greater risk of reoffending than those who had offended against girls only [bibliographic references omitted]. In the nondiagnostic remarks, DSM-IV (APA, 1994) claims that the recidivism rate for male-victim offenders is approximately twice that for female-victim offenders, and although not demonstrating such a marked difference, Furby, Weinrott, and Blackshaw (1989), in an extensive review of recidivism rates, found that reoffending was higher for male victim offenders. [¶] However, the sex-of-victim distinction has not been consistently found, and contrasting findings have been reported in studies that have demonstrated no differences in recidivism rates between the groups [bibliographic references omitted]. Furthermore, Abel, Becker, Murphy, and Flanagan (1981) found that those child molesters who offended against girls reported more than twice as many victims as those who had offended against boys, a finding contrary to the hypothesized outcome." (p. 56)

    

8. Jay & Young. (1977). The gay report: Lesbians and gay men speak out about sexual experiences and lifestyles. New York: Summit.

   This book, published more than 30 years ago by a team of writer-activists, is not a scientific study. The authors' survey methodology is not reported in detail and, because it was a journalistic work, the survey was never subjected to scientific peer review.

    

9. Erickson et al. (1988). Behavior patterns of child molesters. Archives of Sexual Behavior, 17, 77-86.

   This study was based on a retrospective review of the medical records of male sex offenders admitted to the Minnesota Security Hospital between 1975 and 1984. Apparently, 70% of the men abused girls, 26% abused boys, and 4% abused children of both sexes. (The paper is unclear in that it doesn't explain how perpetrators with multiple victims were counted.) The paper asserts in passing that "Eighty-six percent of offenders against males described themselves as homosexual or bisexual" (p. 83). However, no details are provided about how this information was ascertained, making it difficult to interpret or evaluate. Nor did the authors report the number of homosexual versus bisexual offenders, a distinction that the Groth and Birnbaum study (described above) indicates is relevant.

 

One individual has claimed to have data that prove homosexuals to be child molesters at a higher rate than heterosexuals. That person is Paul Cameron. As detailed elsewhere on this site, Cameron's survey data are subject to so many methodological flaws as to be virtually meaningless. Even so, his assertions are sometimes quoted by antigay organizations in their attempts to link homosexuality with child sexual abuse.

In a 1985 article published in Psychological Reports, Cameron purported to review published data to answer the question, "Do those who commit homosexual acts disproportionately incorporate children into their sexual practices?" (p. 1227). He concluded that "at least one-third of the sexual attacks upon youth are homosexual" (p. 1228) and that "those who are bi- to homosexual are proportionately much more apt to molest youth" than are heterosexuals (p. 1231).

Cameron's claims hinge on the fallacious assumption that all male-male molestations are committed by homosexuals. Moreover, a careful reading of Cameron's paper reveals several false statements about the literature he claimed to have reviewed.

For example, he cited the Groth and Birnbaum (1978) study mentioned previously as evidencing a 3:2 ratio of "heterosexual" (i.e., female victim) to "homosexual" (i.e., male victim) molestations, and he noted that "54% of all the molestations in this study were performed by bisexual or homosexual practitioners" (p. 1231). However, Groth and Birnbaum reported that none of the men in their sample had an exclusively homosexual adult sexual orientation, and that none of the 22 bisexual men were more attracted to adult males than to adult females. The "54%" statistic reported by Cameron doesn't appear anywhere in the Groth and Birnbaum (1978) article, nor does Cameron explain its derivation.

It is also noteworthy that, although Cameron assumed that the perpetrators of male-male molestations were all homosexual, he assumed that not all male-female molestations were committed by heterosexuals. He incorporated a "bisexual correction" into his data manipulations to increase further his estimate of the risk posed to children by homosexual/bisexual men.

In the latter half of his paper, Cameron considered whether "homosexual teachers have more frequent sexual interaction with their pupils" (p. 1231). Based on 30 instances of sexual contact between a teacher and pupil reported in ten different sources published between 1920 and 1982, Cameron concluded that "a pupil would appear about 90 times more likely to be sexually assaulted by a homosexual practitioner" (p.1232); the ratio rose to 100 times when Cameron added his bisexual correction.

This ratio is meaningless because no data were obtained concerning the actual sexual orientation of the teachers involved; as before, Cameron assumed that male-male contacts were perpetrated by homosexuals. Furthermore, Cameron's rationale for selecting particular sources appears to have been completely arbitrary. He described no systematic method for reviewing the literature, and apparently never reviewed the voluminous literature on the sexual development of children and adolescents. His final choice of sources appears to have slanted his findings toward what Cameron described as "the relative absence in the scientific literature of heterosexual teacher-pupil sexual events coupled with persistent, albeit infrequent, homosexual teacher-pupil sexual interactions" (p. 1232).

A subsequent paper by Cameron and others (Cameron, Proctor, Coburn, Forde, Larson, & Cameron, 1986) described data collected in a door-to-door survey in seven U.S. cities and towns, and generally repeated the conclusions reached in Cameron (1985). Even Cameron himself admitted that his conclusions in this study are "based upon small numbers of data points" (Cameron, 2005, p. 230). As before, male-male sexual assaults were referred to as "homosexual" molestations (e.g., Abstract, p.327) and the perpetrators' sexual orientation apparently was not assessed. This study also suffers from fatal methodological problems, which are detailed elsewhere on this site.

In yet another article published in Psychological Reports, Cameron claimed to have reviewed data about sexual abuse by foster parents in Illinois and found that 34% of the perpetrators had abused a foster child of their own sex, that is, female-female or male-male abuse (Cameron, 2005). Not only did Cameron again make the fallacious claim that all male-male molestations are committed by homosexuals, he also made the same claim about female-female molestations. Once again, he had no data about the actual sexual orientations of the molesters.

Cameron continues to produce reports that essentially repeat the same inaccurate claims. Perhaps one of the best indicators of his diminished credibility in this area is that his work was not even cited in the 2004 FRC report discussed in detail above.

 
Conclusion

The empirical research does not show that gay or bisexual men are any more likely than heterosexual men to molest children. This is not to argue that homosexual and bisexual men never molest children. But there is no scientific basis for asserting that they are more likely than heterosexual men to do so. And, as explained above, many child molesters cannot be characterized as having an adult sexual orientation at all; they are fixated on children.

Notes
 

Blog   |   Contents   |   Publications   |   Library   |   Site Search   |   Contact Us   |   Useful Links   |   Social Psych Net