Lars Eighner’s Adult Home Page

Nearly all of the bad science about human homosexuality in the last sixty years come from the gay-friendly side for the simple reason that homophobes rarely attempt to be scientific. The opposite is true for supposedly scientific studies much more than sixty years old for the simple reason that there were hardly any gay-friendly scientists in those times.

1. Bad Sample: The number one technical problem with studies about human homosexuality is that the study sample was not adequate to support the study’s conclusions. Bad samples come in (at least) two flavors: inadequate sample (too small) and biased sample. In studies of homosexuality, usually both flaws are found.

The reason these problems are so prevalent — all but universal — is that however a study defines homosexuality (another problem I will get to), it is relatively rare. The percentage of “homosexuals” in the population is estimated to be in the low single digits. So finding a large, truly random sample of “homosexuals” requires winnowing them out from a random sample of the population that is twenty, thirty, or more times larger. This would be extremely expensive to do right, which is why no one tries to do it right.

There have been numerous twin studies, supposedly showing a genetic predisposition to homosexuality. Studies up to the 1950s had tiny samples and generally relied on finding twins from among those identified as homosexuals in prisons and institutions. Some of these showed that as much as 100% of identical twins would both be homosexual if one was. Of course people in prisons or mental health institutions are not representative of the gay population (or the straight population for that matter). And the incident of homosexual behavior is much higher in such institutions because of the situational homosexuality among straight men.

The twin study nowadays usually means Bailey & Pillard, who had a very much larger study group than previous twin studies. The problem with their sample was that they got it by advertising in gay newspapers. This would bias the sample for younger people (as in the following item) and also would tend to over-represent gay men who were willing to involve their brothers in the study — presumably because they knew their brothers were gay or might be. As it was, the study found that in only about half the cases did a gay man with an identical twin have a gay twin, or in other word, whatever genetic component might be in homosexuality, it was defeated about half the time. This leaves aside the question of whether being a twin itself causes a greater likelihood of homosexuality, a question that might be answered by studying twins reared apart without knowing they had a twin.

Dr. Paul and Kirk Cameron of the homophobic American Family Association issued several papers which reported studies supposedly showing homosexual men die (significantly) younger than heterosexual men. The problem: the sample was drawn from obituaries appearing in supposed gay newspapers, many of which were essentially clubbing guides. These obituaries were much more likely to include people of clubbing age or people active in the gay community and to systematically exclude people who were retired from or who never participated in gay nightlife or the activist community. Naturally people in a disco who do die will be younger than average — but if you do not know how many people are in the disco, you do not know whether their death rate is abnormally high. A more recent study from the same culprits compared the age at death of married homosexuals to that of married heterosexuals in Denmark and Norway. But gay marriage (or civil unions) had only been legal for a few years in those places. Both heterosexuals and homosexuals (where it is legal) tend to marry at younger ages. Older heterosexuals who died were much more likely to be married because they were able to marry when they were younger and more likely to get married. Older homosexuals were less likely to be married because they could not get married at the age when people are most likely to marry.

A brain study, by Simon LeVay, supposedly showed that a certain tiny part of the brains of gay men were smaller (more the size of the part in women's brains) than was that part in heterosexual men. The problems here: the sample was very small (19 gay men, 16 straight men, 6 straight women) and the sample may have been biased by the fact that all of the subjects were dead. In particular many of the supposedly gay subjects had died of AIDS, which is known to affect the brain in various ways at various stages, and none of the subjects had been interviewed while they were living. There is an even more serious problem, acknowledged by the author, which I mention in a while.

2. Retrospection: The process of identifying homosexual subjects involves sample bias, but the problem of retrospective studies involves more bias. For example, almost every little boy will try on his mother's high-heeled shoes. Boys who develop into straight men may attach no significance to this and may forget it. Boys who develop into gay men may see it as significant and be more likely to remember it. Likewise, mothers and families of gay men, may know their relative is gay or at least is different in some way, and this will influence their memories.

Many of these ”studies“ have been from the field of psychoanalysis and would better be called “collected anecdotes,‘ as psychoanalysis has never bothered to put itself on a scientific footing. An exception was an East German study that hypothesized that male homosexuality might be caused by differences in intra-uterine hormones when the mother is stressed (thus it would be congenital, but not genetic). The methodology was retrospective: mothers of men identified as homosexual were interviewed. This may also illustrate some sampling bias since any woman in East Germany might remember something stressful from the Nazi or Soviet period if asked to do so.

3. Definitions: The principal meta-scientific problem with studies of homosexuality is failure to find agreement on what constitutes being “a homosexual.” There are three notable candidate criteria: behavior, preference, and identity.

Preference is a bit dicey because some people think preference implies choice, so those people insist on orientation instead. The classical compromise in the free-will versus determinism dilemma is: “Man does as he pleases, but he does not please as he please.” In this light, there really is no difference between preference and orientation.

It is easy to turn up numerous examples where these three criteria are not in agreement. Behavior would seem easy to measure, identity would be very difficult, and preference may be somewhere in between. In truth, all three are culturally loaded cans of worms. In a study by Laud Humphreys, using methods now considered unethical, men were observed having sex in public restrooms, the men were traced, and asked to respond to an apparently unrelated survey about their sexual behavior. Even men who seemed to the investigator to be entirely honest and candid somehow did not recognize that their observed behavior was a homosexual act. In some cultures, even the most forthcoming subject may not think being the insertive partner is doing anything homosexual or that anything other than anal or vaginal penetration by a penis is a sexual act (hence technical virginity and I did not have sex with that woman.)

I cannot think of any good way to determine sexuality as identity except to ask the subject. That too is fraught with peril, even assuming honest subjects. How many subjects will be sufficiently introspective? Will self-identifying as “gay” be the same thing as self-identifying as “homosexual&rdqou;?

Conceivably preference could be observed by observing reactions to various stimuli. The trouble with this approach is that the physically observable effects of sexual arousal are not easily distinguishable from other kinds of arousal.

One such attempt involving the famous “peter meter” (plethysmograph) purported to show that homophobic men are more likely to be aroused by gay pornography than non-homophobic heterosexual men. The reliably of the device involved has been the subject of court cases, because it is also used in testing sex offends, and so far courts have not found scientific opinion concerning its accuracy sufficiently coherent to make its results admissible.

4. Single cause: Nearly every study, however it defines homosexuality, includes the tacit assumption that homosexuality has a single cause. There is no justification whatsoever for this assumption.

Bell & Weinberg's study of gay people in pre-AIDS San Francisco probably was intended to suggest a genetic cause of homosexuality because as the title Homosexualities suggests, gay people and their personal histories are very diverse. The same evidence, however, would also support the thought that no single cause could adequately explain such diversity.

At any rate, the possibility of several — or many — roads to homosexuality is seldom raised. If it is so, then several of the studies may already have identified one of the right tracks, but may fail or may already have been rejected, because no one track can account for the preponderance of homosexuality.

Almost certainly there is no single Mendelian gene that accounts for most homosexuality. It would have revealed itself in number of lineages that have been studies so far. Some of the congenital explanations seem promising, but at best could only explain a small fraction. Psychological explanations also fail to explain why the majority of people with backgrounds supposedly associated with homosexuality do not develop into homosexuals or how people without such a background do become homosexual.

5. Cause and Effect: One of the problems in LeVay's brain study, which he freely acknowledged on first publication is that even if there are physical differences between the homosexual brain and the heterosexual brain, there is no information as to whether they would be the cause or the effect of homosexuality. We can identify a working man by the calluses on his hands, but we know the calluses did not make him a working man. Were the differences there at birth? Did those with “gay brains” develop into gay adult? Or did the differences become more distinct as the person grew into his ultimate sexuality? Obviously we can never determine this for parts that cannot be distinguished except by dissection.

Likewise, psychologists — at least for a time — were fond of explaining homosexuality by a family constellation involving a domineering mother and a distant father. But their theory of male heterosexual development said that the male heterosexual infant will develop to reject his mother and to identify with his father. He ends up with a smothering father and a distant mother, but is the situation of his own making and the is the cause that he is heterosexual to begin with? So why would a homosexual male infant not develop to contribute to distance from his father and identification with his mother? Their reasoning seems to be that the Oedipal process is imposed on a blank slate, so that the result if all goes as they think it should, a heterosexual: Oedipal process is the cause, and heterosexuality is the effect. But what if it is the other way around? Psychoanalysis, as I have mentioned, never attempted to put itself on a scientific footing, so it may duck the issue of determining which is the cause and which is the effect.

Fortunately we are now largely past the period of extreme science by press release. There were times, it seemed, that there was a new cause of homosexuality every week — a little like dietary advice in the present with a new magic food or dreadful poison always in the news. Some of the old discredited reports are still being stirred around by homophobic organizations. Unfortunately, that the wave has passed does not mean that good science has replace the old. In general the old whiz-bang science has not been replaced by anything.

• Adams HE, Wright LW Jr, Lohr BA. Is homophobia associated with homosexual arousal? J Abnorm Psychol. 1996 Aug;105(3):440-5. Abstract
• Bailey JM, Pillard RC. A genetic study of male sexual orientation. Arch Gen Psychiatry. 1991 Dec;48(12):1089-96. (Abstract).
• Bell, Alan P and Martin S. Weinberg. Homosexualities : a study of diversity among men and women Simon & Schuster, 1979 ISBN 0671251503, 9780671251505
• Cameron, P., Playfair, W. L., & Wellum, S. The longevity of homosexuals : before and after the AIDS epidemic. Omega, 29, 249-272. (1994) (Critique by Gregory M. Harek, PhD)
• Hamer, Dean et al. A linkage between DNA markers on the X chromosome and male sexual orientation. Science 1993 Jul 16; 261(5119):321-7.
• Humphreys, Laud. Tearoom trade : a study of homosexual encounters in public places London : Duckworth, 1970. ISBN/ISSN: 0715605518 9780715605516 OCLC:12118038
• LeVay Simon. A difference in hypothalamic structure between homosexual and heterosexual men. Science, 253, 1034-1037. (1991)
• Taylor, TJ. 1992 "Twin Studies of Homosexuality", Part II Experimental Psychology Dissertation (unpublished), University of Cambridge, UK. Author html, 1997.