Birth order and sibling sex ratio in homosexual transsexual South Korean men: Effects of the male-preference stopping rule


Kenneth J. Zucker, PhD, Gender Identity Service, Child, Youth, and Family Program, Center for Addiction and Mental Health, 250 College Street, Toronto, Ontario, Canada M5T 1R8. Email:


Abstract  Two biodemographic variables – birth order and sibling sex ratio – have been examined in several Western samples of homosexual transsexual men. The results have consistently shown that homosexual transsexuals have a later birth order and come from sibships with an excess of brothers to sisters; the excess of brothers has been largely driven by the number of older brothers and hence has been termed the fraternal birth order effect. In the present study the birth order and sibling sex ratio were examined in an Asian sample of 43 homosexual transsexual men and 49 heterosexual control men from South Korea. Although the transsexual men had a significantly late birth order, so did the control men. Unlike Western samples, the Korean transsexuals had a significant excess of sisters, not brothers, as did the control men, and this was largely accounted for by older sisters. It is concluded that a male-preference stopping rule governing parental reproductive behavior had a strong impact on these two biodemographic variables. Future studies that examine birth order and sibling sex ratio in non-Western samples of transsexuals need to be vigilant for the influential role of stopping rules, including the one identified in the present study.


In Western culture the phenomenology of transsexualism (in the DSM-IV, gender identity disorder) has been well-described now for many years in both biological male and female subjects. By adulthood, most transsexuals report a severe discordance between their cross-gender behavior and feelings and their anatomic sex. In biological men, for example, there is often a strong desire to receive feminizing hormone treatment (e.g. to induce breast development) and surgical sex change (e.g. removal of the penis and testes, creation of a neovagina etc.), to align the body with the felt psychological state of a woman.

In biological men with transsexualism it is well-established that there are two major subtypes: those who are erotically attracted to other biological men (homosexual transsexualism) and those who are erotically attracted to biological women and/or to the thought of oneself as a woman (heterosexual transsexualism/autogynephilia).1,2 Apart from sexual orientation, a number of other parameters also distinguish these two subtypes of transsexualism in men. One such parameter pertains to an early childhood onset of persistent and marked cross-gender behavior (e.g. peer affiliation preference, toy preference, roles in fantasy play, cross-dressing etc.); such indicators are more common in homosexual than in heterosexual transsexual men.1,3,4

Two demographic variables – sibling sex ratio and birth order – have also been shown to be reliably associated with homosexual, but not heterosexual, transsexualism. Homosexual transsexuals come from sibships with an excess number of brothers to sisters and are later born (for review, see Blanchard5). Moreover, the late birth order appears to be accounted for predominantly by number of older brothers, not older sisters, and thus has been termed the fraternal birth order effect. The fraternal birth order effect has also been reliably established in homosexual men, unselected for their degree of recalled childhood cross-gender behavior (for review, see Blanchard5).

In recent years sexological, cross-cultural ethnographic research (both historical and in field work) has identified many instances of apparent equivalents (or, at least, similarities) to the contemporary Western form of transsexualism in men. These include studies of the native North American berdache,6 the hijras from India,7 the mustergil from southern Iraq,8 the acault from Burma,9 the jôgappa from south India,10 the xanîth from Oman,11 the kathoey from Thailand,12 the bayot from the Philippines,13 the fa'afafine from Samoa,14 the mak-nyahs from Malaysia,15 and transgendered Brazilian prostitutes.16 When the information is available, one commonality that links the phenomenology across these various “third-gender” categories is the childhood onset of cross-gender behavior and feelings.17 Of course, a great source of debate is whether or not this apparent surface similarity in behavior reflects common or distinct underlying mechanisms.18 As noted by Herdt, the “special conditions” required to create these diverse third-gender categories are poorly understood.19 If it were possible to establish that transgendered people from cultural backgrounds quite distinct from Western transsexuals shared associated features that represent, at least in theory, underlying causal processes, it would give support to the possibility of a common underlying mechanism.

To date, the study of birth order and sibling sex ratio of transsexuals has been predominantly limited to Western samples, consisting of Canadian,20 Dutch,21 and English22 patients. To our knowledge only two studies have examined birth order in transsexuals in non-Western samples. Tsoi et al. reported a late birth order in homosexual transsexual men from Singapore but they did not apply a test to confirm its statistical significance or utilize a control group.23 In a small sample of Samoan fakafefine, Poasa et al. also found a later than expected birth order but, unfortunately, a control group was not available.17,24 In addition, the sample was too small to reliably test for an altered sibling sex ratio favoring brothers. Thus, at present, it is not clear if the sibling sex ratio and birth order patterns are unique to contemporary Westernized transsexuals or have a broader generality.

In a recent article in this Journal, Kim et al. reported data on psychologic distress in a sample of male homosexual transsexuals and non-transsexual heterosexual men from South Korea.25 Because Kim et al. collected data on the sex composition and birth order of their probands and controls, this afforded us an opportunity to examine birth order and sibling sex ratio in a third non-Western sample to test the hypothesis that, like their Western counterparts, Korean transsexual men come from families with an excess of brothers to sisters and are later born largely in relation to number of older brothers.

In the present study we will show that the hypothesized effects were likely masked by a male-preference stopping rule governing parental reproductive behavior.26,27 In Korea it has long been shown that parents have a marked preference for sons.28–37 When there is a strong bias in favor of one sex over the other, there is considerable epidemiological evidence that parents continue to have children until the birth of a baby of the preferred sex. At that point they stop having children.38,39



The Kim et al. sample consisted of 43 male homosexual transsexuals and 49 male heterosexual controls.25 The mean age of the transsexual group was 20.4 ± 1.5 years and the mean age of the control group was 19.8 ± 1.1 years. All subjects were evaluated during a physical and psychiatric examination prior to compulsory military service at three Regional Military Manpower Administrations in Seoul, Suwon, and Kwangju from November 2002 to August 2004. The transsexual subjects consisted of those recruits who were referred for evaluation of gender identity disorder by a primary medical officer or who already carried a medical certificate for transsexualism from a psychiatrist in private practice. The diagnosis of gender identity disorder was based on the consensus of two board-certified psychiatrists using the DSM-IV criteria. The heterosexual controls were consecutive recruits seen during the same time period as the transsexual subjects. For further details on the original study, see Kim et al.25


Raw data on the number of older brothers, older sisters, younger brothers, and younger sisters for both groups were provided to the correspondence author (T.-S. Kim, pers. comm., 17 and 31 October 2006). From these data we computed measures of birth order and sibling sex ratio.


The mean number of older brothers, older sisters, younger brothers, and younger sisters for the transsexual and control subjects are shown in Table 1. Birth order was calculated using Slater's index.40,41 This measure is of the number of siblings older than the proband divided by the proband's total number of siblings, that is, older siblings/(older siblings + younger siblings). This index cannot be calculated for only children; for all other individuals, regardless of sibship size, it expresses birth order as a quantity between 0 and 1, where 0 corresponds to first-born and 1 corresponds to last-born. In theory, a proband drawn at random from a population that is neither increasing nor decreasing with regard to total number of births or mean family size would have an expected Slater's index of 0.50.

Table 1.  Mean number of sibling type as a function of group
Sibling typeGroup
Transsexuals (n = 43)Controls (n = 49)
Older brothers0.260.490–20.430.650–3
Older sisters0.841.130–41.001.290–6
Younger brothers0.140.350–10.220.420–1
Younger sisters0.070.260–10.180.440–2
Older siblings1.091.150–41.431.350–6
Younger siblings0.210.410–10.410.610–2

In the present study, Slater's index for the transsexuals (n = 36) was 0.76 ± 0.42, t(35) = 3.75, P = 0.001, and 0.69 ± 0.45 for the controls (n = 47), t(46) = 2.85, P = 0.007; thus, both groups had a significantly later birth order. The difference between the two scores was not statistically significant, t(81) < 1.

Sibling sex ratio was calculated using the formula number of brothers/number of sisters. Sibling sex ratio is often compared to the secondary sex ratio at birth, which typically yields a value close to 106:100 for male live births to female live births.42,43 (The sex ratio of live births is slightly but reliably lower for populations descended from sub-Saharan Africans.) The ratio of brothers to sisters reported by any group of (non-black) probands drawn at random from the general population should therefore approach 106 (brothers per 100 sisters).

To test the hypothesis that transsexuals have an excess of brothers compared to sisters (and that the controls would not), we expressed this value as the proportion of brothers rather than the ratio of brothers to sisters, that is, 0.515 (106/206). James has suggested that a test value of 0.518 should be used when a sample of probands consists entirely of male subjects, as in the present study.44,45 It should be noted that the calculation of the sibling sex ratio of an empiric sample, whether expressed as a ratio or as a proportion, does not include the proband.

For the transsexuals, the sibling sex ratio was 44:100 (0.30) and 55:100 (0.36) for the controls. A binomial probability test indicated that both ratios were significantly skewed in favor of sisters, not brothers (two-tailed P < 0.002 and <0.003, respectively).

To test the hypothesis that transsexuals have an excess of older brothers rather than older sisters (and that the controls would not), we calculated similar binomial probability tests. For the transsexuals, the sibling sex ratio for older siblings was 31:100 (0.23) and 43:100 (0.30) for the controls. A binomial probability test showed that both the transsexuals and controls had significantly more older sisters than older brothers (two-tailed P < 0.0001 and <0.001, respectively). Thus, working against our hypothesis, this finding shows the role of the male-preference stopping rule of parents continuing to bear children until they have a male child and that this rule appeared to be operative in both the transsexual and control groups.


In the present study we found that both the homosexual transsexual and heterosexual control men had a significantly late birth order. Thus, although we were able to show that the homosexual transsexuals had a late birth order, as previously shown in Western samples of homosexual transsexual men, the effect was not specific.

Unlike Western samples, in which homosexual transsexual men have been found to have an excess of brothers to sisters, the Korean homosexual transsexual men in the present study had a significant excess of sisters to brothers, and this was also the case for the heterosexual control men. We also found that the excess of sisters to brothers was largely accounted for by the number of older sisters compared to the number of older brothers. Thus, in the present study, the late birth order of both the transsexuals and the controls was driven largely by an excess of older sisters, not older brothers. One limitation of the present study was that we did not have a clinical group of heterosexual (or non-homosexual) transsexuals; but, based on findings from Western samples of heterosexual transsexual men, there is no evidence, at present, to indicate that such patients would have an altered birth order and sibling sex ratio.20,21

The results from the present study show clearly how parental reproductive behavior that appears to be governed by a male-preference stopping rule can distort both birth order and sibling sex ratio. The evidence for such a rule was strengthened by the finding that it was apparent in both the transsexual and in the control group. From a theoretical point of view, it remains desirable to try and replicate the findings on birth order and sibling sex ratio that have been found in samples of Western transsexual men. We would encourage researchers to identify potential data sets in which to examine these parameters, but would caution that close attention is paid to the possibility that male-preference stopping rules or other types of stopping rules governing parental reproductive behavior are not distorting the results.46,47