Unbiased Study shows Importance of Environment in Sexual Orientation Development?
Dr Whitehead has a precis of this new research that has incidentally shown what many of us have been batting on about for years.
Update – Please note the comment below by Warren Throckmorton who casts doubt on the usefulness of this study. I have adjusted the title accordingly!
Dean Hamer (whose name is particularly associated with “gay gene” studies) has an interview segment on the U-tube ex-ex-gay website in which he says that upbringing has nothing at all to do with the development of homosexuality. The reason he says that goes back basically to studies from 1981. But a recent paper from Taiwan, (Lung and Shu, 2007) shows for the first time in a modern sociological survey that in some places, in some cultures, the influence of mothers and fathers and upbringing can be extremely strong, in fact accounting for most of the influences. This also shows that cultural factors are important, because they cause the relative importance of genetic and environmental factors to shift. In this paper I review the intellectual history of this argument, to put the Taiwanese paper in context.
A little history
In the West there have been two main sources of material on the importance of parents, one backing their importance, the other not. The first was reports from psychiatrists and therapists of what clients had found important. These went back to the mid twentieth century and even earlier. They could hardly be disputed as important to the particular clients, but they did not enable us to make statements about the population as a whole. For that, sociological surveys were necessary. The basic impression from the papers published by psychiatrists and therapists was that “smothering mothers” could be to blame, and perhaps emotionally or physically absent fathers. Sissiness, perhaps resulting from maternal over-protection was another facet of the same family configuration.
The second source was Bell, Weinberg and Hammersmith (1981). They published the results of a large sociological survey on a sample gathered by the Kinsey Institute before 1970, which contained a high percentage of homosexual people, and hence allowed statistically reliable conclusions, though it wasn’t a random sample, so we have to be a bit careful about the conclusions. They tried to make this study definitive and assembled a list of almost every social factor asserted by someone at some time that possibly influenced the development of homosexuality, and checked if they did correlate with later homosexuality. The results were at odds with the previous anecdotal evidence gathered by the clinicians. Each of the family factors led to homosexuality in only a small minority of cases. Other unknown factors were more important.
One possibility of course is that researchers asked the wrong questions. But in that case the therapists had also been wrong since all their suggested possibilities had apparently failed the test. Another possibility was that biological factors were predominant, not social ones, and the authors speculated that might indeed be so. A third possibility was not even considered – that random reactions to common- environment factors predominated. (Whitehead, 2007) . The evidence points fairly strongly to the latter being the case.
Combining all the apparently relevant social factors the authors were able to explain 30% of homosexuality using their mathematical model ( (Bell, Weinberg & Hammersmith 1981) (or download chapter 11 from this website). However at one part of their book they said 30% was “Significant, and at another part “Not significant”. The contradiction between these two statements led to many subsequent writers simply stating no social factors produced homosexuality. Van Wyk and Geist, (1984) pointed out this was incorrect, but were ignored. The truth is that explaining 30% in any study of this type, using this type of statistical method is common, and 30% is significant. But as to explaining most homosexuality it was indeed “not significant”. Most homosexuality was not explained. (Bell, Weinberg & Hammersmith 1981) .
No more studies of this extent or on this scale were done until now, and the literature, deferring to Bell/Weinberg/Hammersmith’s paper, perpetuated the untruth that “family factors have no effect (at all)”.
Twin studies show no social factors ? Really?
Twin studies, from the year 2000 on, particularly, seemed to support the idea that social factors had no effect on homosexuality. Twin studies subdivide influences into genetic factors, environmental factors experienced in common, and environmental factors experienced by one but not the other. For homosexuality twin studies could not detect significant influence from common environmental factors. (Kendler, Thornton, Gilman, & Kessler, 2000; Bailey, Dunne, & Martin, 2000; Bearman & Bruckner, 2002; Santtila et al., 2008).
But this conclusion was not as clear as it seemed, because (Whitehead & Whitehead, 2007) the twin study methodology for homosexuality tends to overestimate the genetic percentage at the expense of the common environmental percentage. As said by Visscher et al. (Visscher, Gordon, & Neale, 2008) ”…the twin literature based upon the classical twin design and model selection procedures could be severely biased…” i.e. twin studies will simply not detect common influences unless sample sizes are very large, and common influences are very strong.
But more importantly twin studies actually conceal the level of common environmental influence. It shows up instead as non-common environmental influence, that is, people reacting in a very individualistic way to the same influence. A case could even be made that these individualistic erratic reactions are predominantly to the common environmental influences, and that the individualistic erratic reactions to biological factors are minor. This point has already been made in Whitehead ( 2007) and in a much fuller way in another paper submitted for publication. See also chapter 10 on this website.
Enter Lung and Shu
At this point in the debate , the paper by Lung and Shu (2007) appeared. It showed a very strong influence of parental style in Taiwan on the development of homosexuality. This was not a marginal result like so many tend to be, but unequivocal. It showed that parental influences were predominant.
This seemed to contradict much of what had gone before (interestingly Lung and Shu don’t seem to have heard of Bell, Weinberg and Hammersmith, or twin studies!). Incidentally they quote a 1986 paper of Elizabeth Moberly with an additional author Phil, M.A.D. whose name appears to be constructed from her degrees MA and D.Phil……But how can the papers be so contradictory? I believe there are good reasons, and they are cultural, and shed light on the results obtained in the West.
Lung and Shu seem to be associated with the military in Taiwan and their subjects were from the annual intake of 140,000 young recruits. In that country military service is compulsory, hence the recruits represent the whole population of men. There are inevitably those who find military training almost unbearable, and many in Taiwan are diagnosed with adjustment disorder (a DSM mental health category). From these the authors selected 51 homosexuals, and 100 non-homosexuals. The controls were 124 recruits without adjustment disorder. Recruits with mental health issues other than these were eliminated from the study. It seems obvious that this study was possible because Lung and Shu were told by the authorities to study recruits who could not cope. However it means that the study of homosexuality is complicated by the adjustment disorder, which the authors had to take into account in the interpretation of their results. No other study has involved those with an accompanying mental condition like this. However it seems to me that the authors allowed for the adjustment disorder quite adequately. Overall the sample is much more representative than many in the West.
To give the flavour of the extraordinarily clear-cut results from the well known Parental Bonding Instrument the authors used, (a questionnaire which measures relationships to parents), I present here the results for Paternal Care.
These are numbers on a scale, but you don’t need to know how that compares with the realities of the family – merely compare the numbers. The results were 13.65 ±1.00 (standard deviation) for the homosexual group, 18.07 ±0.53 for the non-homosexual group and 19.02 ±0.44 for the controls. This means the heterosexual adjustment disorder group had parental issues indistinguishable statistically from the controls (In Fig 1 the curves for both almost completely overlap.) But the parental issues for the homosexual group were much, much more important. (There is no overlap at all between the curves.) In fact the homosexual group is so far separated from the non-homosexual that it represents some kind of record – a Taiwanese man (with adjustment disorder) is classed as homosexual or non-homosexual depending (almost entirely) only on the absence of care he received from his father, i.e. a distant father. There would be a very low error rate because the standard deviations are so relatively small. I know of no other indicator with such a power to discriminate. Similarly the homosexuals were found to be highly significantly deprived of maternal care, but also there was a very high degree of protection by both mothers and fathers. This seems paradoxical, but the general picture is of parents who are psychologically very distant, but performing their parental duties, and over-protecting the proto-homosexual by keeping him a little Daddy’s or Mommy’s boy, and not exposing him to the difficulties of life. Allowing for the adjustment disorder, other results showed the homosexual men were also much more introverted and neurotic.
Social factors explain 62% of homosexuality
In their statistical model to explain homosexuality, Lung and Shu managed to explain 62% of the variance by parental factors, neuroticism level, and introversion level. ie 62% of homosexuality of their sample can be explained by parental factors, and higher than normal levels of neuroticism and introversion. It is quite rare to get a figure as high as this when a sociological survey is involved. This is twice the effect that Bell, Weinberg and Hammersmith found, and has an extra fascinating implication: for the first time a careful modern study shows social factors predominate, and hence other factors, such as genetics, (at least in Taiwan), must be minor!
These results support those who talk about over-close mothers and distant fathers as causes of homosexuality. Why were Lung and Shu’s results so clear compared with results from the West which were much less clear? Could the authors have manufactured their results? Are they too good to be true? I think not, because the authors seem largely unaware of the details of the controversy there has been in the West, though they understand it existed. They do not appreciate the deep significance of their results.
They do mention a cultural difference. They say that Taiwanese society is very traditional and conservative. This presupposes a high degree of social control, and a suppression of any genetic predispositions there might be. The role of any family factors is likely to be highly magnified under such circumstances.
I conclude that the results reflect one extreme – what happens in a society in which family influences are very strong. That itself is useful because it gives a picture of what would happen at one extreme even in the West, in social groups where only family factors are involved (?the Amish?). However generally in the West things are very different. Why? Because in comparison with Taiwan we are hyperindividualistic. Our high divorce rate and extreme diversity of belief and custom are evidence of this. Hyper-individualism seems to be one of our most prized and politically correct values. Therefore in the West even two genetically identical twins are likely to react in a different way to circumstances which might tend to trigger homosexuality.
My belief is that if the twin studies done in the West were done in a society like Taiwan, the results would be totally different, and show a very high contribution from common factors, and much diminished contributions from genes and non-shared environment. (This is already known for various traits – in Japan, twin studies mostly give a much lower genetic contribution than in the West). This has significant research implications. We should deliberately sponsor studies in societies with extreme conditions to get a clearer picture of influences in our own!
So for the Western situation, if some male client seems to have the pattern of a distant father, or a family dynamic which has enforced conformity and discouraged initiative and rough-and-tumble play, these factors are inherently credible, and should not be dismissed, though they will not be politically correct. However reactions to these factors will be very variable, and individualistic.
Bailey, J.M., Dunne, M.P., & Martin, N.G. (2000). Genetic and Environmental influences on sexual orientation and its correlates in an Australian twin sample. Journal of Personality and Social Psychology , 78, 524-536.
Bearman, P.S., & Bruckner, H. (2002). Opposite-sex twins and adolescent same-sex attraction. American Journal of Sociology , 107, 1179-1205.
Bell , A.P., Weinberg, M.S., & Hammersmith, S.K. (1981). Sexual Preference: Its Development In Men and Women. Bloomington , Indiana : Indiana University Press.
Kendler, K.S., Thornton , L.M., Gilman, S.E., & Kessler, R.C. (2000). Sexual orientation in a U.S. national sample of twin and nontwin sibling pairs. American Journal of Psychiatry, 157, 1843-1846.
Lung, F.W., & Shu, B.C. (2007). Father-son attachment and sexual partner orientation in Taiwan . Comprehensive Psychiatry, 48, 20-26.
Santtila, P., Sandnabba, N.K., Harlaar, N., Varjonen, M., Alanko, K., & von der Pahlen, B. (2008). Potential for homosexual response is prevalent and genetic. Biological Psychology, 77( 1 ), 102-105.
Van Wyk, P.H., & Geist, C.S. (1984). Psychosocial development of heterosexual, bisexual and homosexual behavior. Archives of Sexual Behavior, 13(6), 505-544.
Visscher, P.M., Gordon, S., & Neale, M.C. (2008). Power of the classical twin design revisited: II Detection of common environmental variance. Twin Research and Human Genetics, 11(1), 48-54.
Whitehead, N.E., & Whitehead, B.K. (1999). My Genes Made Me Do It! Layfayette , Louisiana : Huntington House.
Whitehead, N. E. and Whitehead, B. K. (2007) My Genes Made Me Do It! 2nd Edition. Whitehead, N. (2007). An antiboy antibody? Re-examination of the maternal immune hypothesis. Journal of Biosocial Science , 39(6), 905-921.
I’m still trying to get hold of the original paper. If anyone can help with that let me know.
Peter – I will see if can get you and electronic copy of it. I am surprised that Neil has given this study much weight. There is indication of bias in the study in that homosexuality is still considered a disorder in this society. Here is my take on it. Note the much higher neuroticism scores among the gay group. These were not matched samples; the deck was stacked against the gay group.
Thanks Warren. That’s exactly why I want to get hold of the full paper so I can read Whitehead’s claims in context. Also, I didn’t realise that the paper had been out for at least a year.
thank you to you and to Warren for the post and the comments. I think it’s worth quoting one of Warren T’s paragraphs:
“UPDATE – 2/15/07: In further review of the study, it appears that the flaws in design outweigh any conclusions that could be drawn from it. Not only was the homosexual group a clinical group (they were diagnosed with an adjustment disorder), they had much higher neuroticism scores than the other two groups which stacked the deck against them. In essence, the authors varied two independent variables (sexual orientation, & neuroticism) simultaneously and attributed all the effects to the homosexual variable. You canâ€™t tell what predicts what and certainly not causes what. Without adequate controls, you cannot say how much of the variance in perceived parenting was related to being gay and how much was due to higher neuroticism.”
sorry to double post but have just thought – adding the question mark to the title seems a good move, but what about also removing the word ‘unbiased’, given Warren Throckmorton’s comments?
in friendship, Blair
I think I’m going to leave the “Unbiased” in there because the original research had nothing to do with homosexuality. In that sense the sample was unbiased, though the interpretation of the results may not have been so.
I’ve got hold of the paper and I’m going to go through it with a tooth-comb to see whether what Warren’s saying is true. Might not get round to it until next week though.