Birth order and male homosexuality: Extension of Slater's index

MARSHALL B. JONES AND RAY BLANCHARD2

Abstract Homosexual men tend to be later-born children. Slater's index, the ratio of older sibs to all sibs, is consistently higher for male homosexuals than for comparable heterosexuals. According to some explanations of this tendency, homosexual men are later born with respect to their brothers and later born with respect to their sisters only secondarily and less strongly. We show that if sisters have no direct bearing on a brother's sexual orientation and brothers do, then

These ratios are calculated and compared in nine samples of homosexual men and nine corresponding samples of control heterosexuals. The first equation holds for homosexual men, and the second equation holds for heterosexual men. The late birth order of homosexual men is sex specific. What matters is a boy's birth order relative to his brothers only. This effect may have its origins in an immune reaction or in behavioral contagion.

Quantitative research on birth order and sexual orientation was introduced by Eliot Slater during the 1950s. Slater's study of homosexual patients at the Maudsley Hospital produced the finding that homosexual men, on average, tend to be born late in their sibships (Slater 1958, 1962). Despite its confirmation in a second large series of Maudsley patients (Hare and Moran 1979), Slater's finding did not generate much research interest in the following three decades. It is likely that part of the reason for this neglect concerned methodological problems. The Maudsley studies were conducted without heterosexual control groups, which left open the possibility that the findings were due to variations (patient status, population trends, etc.) other than sexual orientation (Hare and Price 1969, 1974; Price and Hare 1969).

In addition to first noticing the effect, Slater (1958, 1962) also introduced an index for measuring it. Slater's index is calculated by dividing the number of a proband's older sibs by the number of all his sibs. This index is not defined for only children, but for all others it varies between 0 and 1. For the eldest child the index equals 0, and for the youngest child the index is 1. Under the assumption that all birth orders are equally likely, the expected value of Slater's index is 0.50, regardless of sibship size. In his 1962 study Slater reported a mean value for his index of 0.58 in a sample of 337 male homosexuals, all of whom had at least one sibling.

In recent years the question of birth order and sexual orientation has been revisited in a series of nine studies, covering nine samples of homosexual men and nine corresponding samples of heterosexual controls. These 18 samples are listed in Table 1 along with brief descriptions and sample sizes. The first columns of Tables 2 and 3 present the mean Slater index for homosexual and control men in each of the nine comparisons. In every one of the nine comparisons the homosexual probands have a higher (later) mean birth order than do the heterosexual controls. Three of these comparisons (1, 5, and 6) are significant at the 0.001 level and two other comparisons (4 and 7) are significant at the 0.02 level or better. For further information, see the original reports.

The samples include probands examined in Canada, the United States, England, and the Netherlands. The probands were examined over seven decades, with years of birth ranging over more than a century. The consistency of the findings and the diversity of the groups in which they have been obtained strongly reinforce and extend Slater's original conclusion. In discussing his result, Slater (1958, p. 81) suggested that "mutual sex play occurring with a brother during childhood might have a predisposing effect" toward homosexuality. If it is also supposed that this predisposing effect is greater on the younger than on the older boy, the result is a tendency for male homosexuals to be born late in their sibships. Such an explanation would be a special case of behavioral contagion, as analyzed by D.R. Jones and Jones (1992) and M.B. Jones and Jones (1994, 1995). This analysis has been developed primarily in relation to antisocial behavior but applies equally well to prosocial behaviors, for example, participation in community programs or volunteering.

Another possibility is one put forward originally by MacCulloch and Waddington (1981) but currently advocated mainly by Blanchard (Blanchard and Bogaert 1996b; Blanchard and Klassen 1997). According to this explanation, a male fetus provokes an immune reaction in the mother and this reaction alters the sexual differentiation of the brain in later-conceived male fetuses. The model for this theory is hemolytic disease of the newborn (HDN), with the obvious differences that HDN affects red blood cells, not the brain, and is provoked by and subsequently interacts with fetuses of both sexes. Blanchard also supposes that the effect is cumulative, so that the likelihood of an alteration in sexual orientation increases with the number of male fetuses that the mother has previously carried.