Socialization Effect

Male adolescents who have a female twin are more likely to report a same-sex attraction than any other group in these data (16.8 %, table 1, row 1)14. Female adolescents with a male twin, while not different from others, are significantly less likely to report a same-sex attraction than their male counterparts. This result points toward gender-specific differences in socialization.

Specifically, negative sanctioning of behavior that suggests femininity and/or homosexuality is a stronger component of male socialization than comparable  sanctioning of masculinity for female socialization. Girls wear pants but boys don.t wear skirts.

Table 3 reports significance tests for selected comparisons of the proportions reported in table 2.

13 The literature reviewed above dealt mostly with self-identified homosexuality. In the light of cultural and social variation in identity formation, focusing on same-sex romantic attraction rather than self-identification, should help separating biological effects from social influence. In any event, same-sex romantic attraction is a strong predictor of subsequent behavior. In our sample, adolescents who report same-sex attraction in wave 2 are much more likely than others to report same-sex dating, romantic, and sexual contact in the third wave (for males, 8.8 % compared to 1.6 %, p=0.0001, N=7174; for females, 9.5 % compared to 1.7 %, p=0.0001, N=7490).

14 When including respondents with more than one sibling, results do not change (for females, P=0.481; for males, P=0.027). For this analysis, respondents were classified hierarchically in the order shown (thus, a respondent who was in a pair of OS twins and a pair of SS full siblings is classified as an OS twin).

Strana 10

shows the probability that the proportion of teens with same-sex romantic attraction among females with a male twin equals the proportion of teens with same-sex romantic attraction among females with a SSDZ twin (.567).

The observed difference in same-sex romantic attraction between male SSDZ and OS twins (table 2, row 1 (16.8%), table 2, row 2 (9.8%), a 7 point percentage difference, P=0.027) cannot be explained by the genetic model. Males with a female twin are more than twice as likely to report a same-sex romantic attraction than males with a full sister (table 2, rows 1 (16.8%), and 4 (7.3%), P<0.000). Males in OS twin pairs are more likely to report same-sex attraction than males in the full non-genetic sample, and males with one-sibling in the non-genetic sample (rows 8 and 9 in table 2, P<0.000 for both comparisons).

Still, these results are compatible with both the social influence and intrauterine hormone transfer hypotheses, although prior evidence for the effect of shared intrauterine environment suggests masculinization of females rather than feminization of males. Here, we observe the obverse. We now disentangle social influence from hormonal influence.

Older siblings and same-sex romantic attraction

Hormonal transfer should be insensitive to birth order. If the observed prevalence of same-sex orientation among males in OS twin pairs is an outcome of a socialization process, the presence of older siblings should have an effect on OS twins. sexual orientation. Specifically, equality norms put constraints on the extent to which parents and others engage in gender-socializing behavior towards opposite-sex twins. If opposite-sex twins have older same-sex siblings, gendersocializing  mechanisms in the family may be locked-in. Parents will be more likely to negatively sanction gender-atypical behavior among OS twins if those twins have older siblings who are discouraged from gender-atypical behavior. Similarly, gender markers (clothes, toys, and rituals) may already exist in the repertoire of such families and be applied to OS twins. Consequently, older siblings should reduce the prevalence of homosexual attraction for opposite-sex twins under a social influence model. If the hormonal transfer hypothesis holds, no such reduction should be observed.

Among male opposite-sex twins, the proportion reporting a same-sex romantic attraction is twice as high among those without older brothers (18.7 %) than among those with older brothers (8.8 %). No such difference obtains for female opposite-sex twins, who are unlikely to report a samesex attraction whether they have older sisters or not (5.1 % versus 5.7 %). If differences between OS twins and others were based on prenatal hormonal transfers, older brothers should not decrease the likelihood of reporting same-sex romantic attraction. Based on the evolutionary dynamics model, in contrast, individuals with older siblings should be more likely to report samesex attraction. We now turn to this hypothesis.

________________________________________________

15 Arguably, a chi-square test is inappropriate because the data violate the assumption of independent observations

(siblings are not sampled independent from each other). If same-sex romantic attraction was determined by genes, and

observations are paired with respect to their genes, one set of genes is .duplicated. in the proportions calculated for

same-sex pairs in table 2. Thus, the comparison of OS twins with SSMZ twins is .conservative. with respect to the

social influence hypothesis, since each gene set enters the calculations only once for OS pairs but twice for the SSMZ

pairs, thus concordance should be higher for SSMZ pairs.

Strana 11

Birth Order Effect

As noted earlier, a speculative evolutionary theory suggests that that homosexuality increases fitness for individuals with many older siblings. The idea is that individuals sacrifice their specific interests in order to maximize group success. In this case, the theory suggests that younger brothers, unable to compete with their older bothers for access to women, help the family unit by engaging in cooperative raising of their elder brothers. children, at the same time, shifting the focus of their erotic interests to men. Under this model, engagement in co-operative raising of older siblings. offspring may be more likely to succeed (in increasing group fitness) than attempts at procreation. Thus, the proportion of individuals reporting homosexual orientation should increase with number of older (full) siblings. This hypothesis is addressed in table 4. As expected,

we find no association between same-sex attraction and number of older siblings, older brothers, or older sisters16.