I had to share this detail from the Family Research Council’s webpage for their book Getting It Straight:What the Research Shows About Homosexuality:
Chapter 6: Is There a Link between Homosexuality and Child Sexual Abuse?
Ã?Â· A study of 229 convicted child molesters in Archives of Sexual Behavior found that “eighty-six percent of offenders against males described themselves as homosexual or bisexual.” (emphasis added)
(W. D. Erickson, “Behavior Patterns of Child Molesters,” Archives of Sexual Behavior 17 (1988): 83.)
I’m reminded of Kieran’s classic question, “Why are so many of the closets I open full of my clothes?”
Here’s what I wrote in the comments:
This is actually not trivial; in fact, it’s almost certainly wrong. Most surveys of instances of child molestation have found that male molesters of male children are overwhelmingly not “homosexual” in any conventional sense (i.e. pursuing relationships with adult males). And that’s important because the proportion of instances of child molestation that involve male-male contact is much higher than the proportion of gay men in the population. That fact is used by conservatives to claim that gay men are more likely than other men to molest children; the reason they’re wrong about that is precisely because, in fact, the overwhelming majority of male molesters of male children are not “gay” in the colloquial sense.
The survey result is probably the result of men whose primary sexual attraction is to young boys (though they may be heterosexual or, more likely, totally asexual, in their relationships with adults) concluding that they must, therefore, be “gay” (since it is, after all, boys they’re attracted to—the conservative argument). But psychologists and social scientists tend to regard pedophilia as, in a sense, a separate orientation of its own. If someone’s primary sexual activity is screwing goats, you don’t ask whether the goats are male or female to determine whether the person is “straight” or “gay”. That’s a function of their relationships with adult humans (if they have any), and once you’ve determined that, you look to see whether it’s correlated in any way with the pathology.