In 2000 a good review article from Canadian scientists appeared on lefthandedness and homosexuality.1 Following selection and collation of many studies it concluded that male homosexuals are about one third (31%) more likely than heterosexuals to be left-handed2and lesbians almost twice as likely (91%).
This appears to suggest an overwhelmingly strong (biological) factor underlying both left-handedness and homosexuality.
I argue below that there is possibly some link, but it is weak, because the vast majority of left-handed persons are not homosexual and most homosexual people are not left-handed.
The Canadian paper is a meta-analysis; an increasingly popular way of combining data from multiple studies to overcome the problems of slightly different approaches, and pin-point small effects which individual studies do not have enough power to detect.
There are many traps in meta-analyses, but the present paper has avoided most of them, as one would expect given the prominence and expertise of its best-known authors, Blanchard and Zucker. A criticism might rather be that it was published in the Psychological Bulletin, certainly correct for reaching its target audience, but given that the paper is extremely technical and uses odds ratios (statistics from the field of epidemiology), and refers to fetal masculinization (endocrinology), the Major Histocompatibility Complex (immunology), and fluctuating asymmetry (developmental biology), it is quite doubtful it got a rigorous refereeing, since this number of specialties exceeds the usual number of referees!
The main caution must be about the significance of the effect. Neither the authors nor I can quantify its error, because it was done using a meta-analysis. A known epidemiological rule of thumb for individual surveys states that, in a test-population, a prevalence of twice that in the control group is intriguing but inconclusive, and a prevalence of three times is probably real. This means that if this were an individual survey, the result of 31% above normal would be insignificant, and that of 91% above normal (about twice the control group) intriguing but not final. The point here is that the rule of thumb applies to individual surveys, but a meta-analysis combines many. Shouldn't that make the grand average more reliable? In theory, but we need a new rule of thumb for meta-studies and don't have one yet. (In a footnote the authors erroneously but conservatively imply the rule for individual studies applies to theirs.)
The authors conclude by discussing possible biological reasons for their results, and prefer an explanation in which both left-handedness and homosexuality result from biological developmental error.
Using their results it is possible to derive a number, which shows the extent of any link there may be. Here's how it's done. Given that 2.7% of adult Western males are homosexuals and 1.7% of adult Western women female homosexuals, (both figures including bisexuals, and defined as activity in the last 12 months3 we can calculate by standard methods that only 3.9% of left-handed males are homosexual and only 3.3% of left-handed females are lesbian.4
In other words the overwhelming majority of those who are left-handed when young do not become homosexual. The strength of any underlying factor producing both is weak. Something unusual happens to a small number of left-handed people and their sexual-orientation development is atypical.
What might that be? Are there alternative psychological explanations of the effect described in the paper?
I suggest at least one - everyone knows that left-handed people feel in a minority, and different. Very many homosexual people attest to feeling different and not belonging when they were growing up. Left-handedness could increase feelings of difference and be one important factor in the complex leading to homosexuality. This could apply to both men and women and the authors quote a study in which non-right-handed women rated themselves as less feminine.5 Additionally feeding into this could be altered brain hemisphere dominance and the psychological effects of atypical gender-specific skills.
Most left-handed children do not become homosexual. Conversely and importantly, most homosexual people are not left-handed. So the main routes to homosexuality are not via whatever causes being left-handed. There is a link or correlation to homosexuality but it is weak.
1. Lalumière, M.L.; Blanchard, R.; Zucker, K.L. (2000): Sexual orientation and handedness in Men and Women: a meta-analysis. Psychological Bulletin 126, 575-592.
2. Although for conciseness the term "left-handed" is used in this article, the original definition in the paper is "all those who are not exclusively right-handed".
3.Whitehead, N.E.; Whitehead, B.K. (2000): My Genes Made Me Do It!.
4. The calculation is quite easy and useful to put in perspective other alleged links. F1 = 0.0277 * R1 (for men) and F2 = 0.0173 * R2 (for women), where F1 and F2, the results, give the fraction of left-handers who are homosexual (men and women respectively), and R1 is the ratio of the percentage of left-handed homosexual men to the percentage of left-handed heterosexual men. R2 is similar but for women. The numerical factors are the ratio of homosexuals to non-homosexuals in a population, thus it is 2.7/97.3 for males and 1.7/98.3 for females. Other figures could be substituted if you think them more accurate. The formulae apply similarly to other biological factors that are alleged to link with homosexuality; only R1 and R2 change.
5. Casey, M.B.; Nuttall, R.B. (1990): Differences in feminine and masculine characteristics in women as a function of handedness: Support for the Geschwind/Galaburda theory of brain organization. Neuropsychologia 29, 749-754.