review of older brother effect in homosexuality

Biological versus non-biological older brothers, and male gay orientation

Comments on Bogaert’s PNAS (1) paper, 2006:


N.E.Whitehead, Ph.D



This paper recruited from the gay community people with elder brothers, to compare with heterosexual people with elder brothers. It also recruited heterosexuals from blended families, and families involved in adoption, and also used archival samples. It confirmed a well known effect in the literature – the more elder brothers a boy has, the higher the chance of same-sex attraction (SSA). But the really interesting and novel point is that there was a similar prediction even if the subject had been brought up in a family isolated from those biological elder brothers. The author therefore argues that homosexuality cannot be a social effect peculiar to either family, and must be biological in origin. He then attempts to explain the effect using the maternal immune hypothesis (see below).

The study is not very likely to be affected by the volunteer effect which is the bane of much research in this field. The advertisements (not described in the paper) only asked for gay adoptees and made no mention of older brothers. This was good methodology.

Bogaert does not give enough information about the adoptions. It could be that adoptions were mostly extremely early before there could be any social influence of the biological family. That would strengthen the author’s case for sheerly biological influences. But it could also be that most were mid childhood or later (a result of long-term fostering or divorces, for example) after social effects in the family of origin had come into play. In a check for social factors the author found no connection between SSA and numbers of years spent with either non-biological or biological brothers. But looking only at numbers of years could easily overlook other crucial social data.

I am also unhappy statistically with some aspects of the modeling. Trying to use his model with so many (18) variables and so few (378) subjects, weakens the statistical power so much that the result will have a large error, and the author does not present an analysis of how bad this effect is.

The author goes on to strongly press the case not just for a biological origin of SSA but a maternal immune hypothesis. This hypothesis has at least four speculative layers to it, and requires much better evidence before it gains general acceptance. (Click here for a critical review of this hypothesis).

Importantly it should be remembered that most gays do not have elder brothers. The present theory tries to explain the SSA of perhaps only about 17% of the gay population. The SSA of the others must arise from other sources.

The paper in no way warrants the media frenzy it has generated, particularly the headlines which argue that SSA is already fixed in the womb.


(1)Proceedings of the National Academy of Sciences