Family and Sexual Orientation: The Family-Demographic Correlates of Homosexuality in Men and Women

Abstract

Using a nationally representative sample of young adults, I identify the family-demographic correlates of sexual orientation in men and women. Hence, I test the maternal immune hypothesis, which posits that the only biodemographic correlate of male homosexuality is the number of older brothers, and there are no biodemographic correlates of female homosexuality. For men, I find that having one older brother does not raise the likelihood of homosexuality. Although having multiple older brothers has a positive coefficient, it is not significant. Moreover, having any older sisters lowers the likelihood of homosexual or bisexual identity. For women, I find that having an older brother or having any sisters decreases the likelihood of homosexuality. Family structure, ethnicity, and education are also significantly correlated with male and female sexual orientation. Therefore, the maternal immune hypothesis cannot explain the entire pattern of family-demographic correlates. The findings are consistent with either biological or social theories of sexual orientation.

Notes

¹Blanchard and Bogaert (1996) analyzed 302 homosexual men matched with 302 heterosexual men. Homosexual subjects were recruited at a meeting of a homosexual community organization and at a Gay Pride parade. Bogaert (2006) examined about 900 homosexual and heterosexual men. Some of the subjects were (undergraduates) recruited at a university. The remaining subjects were recruited using advertisements in gay or other publications. All of the subjects in the two studies were Canadian.

Note. M/F indicates whether the summary statistics are calculated for male respondents, female respondents, or both. Sibling and sexuality variables come from Waves I and III, respectively. Data source: Add Health (Waves I and III).

²In the National Health and Social Life Survey (NHSLS), which is nationally representative of people ages 18–59, 4.5% of men rated having sex with someone of the same gender as appealing, whereas 5.8% of men in Add Health reported having had a romantic attraction to another man (Laumann, Gagnon, Michael, & Michaels, 1994, 1995). About 5.6% of women in the NHSLS rated having sex with someone of the same gender as appealing, whereas 13.1% of women in Add Health reported having had a romantic attraction to another woman. Moreover, in the NHSLS, younger men tended to express more same-gender attraction than older men. About 7.7% of male respondents ages 20–24 and 5.2% of respondents ages 30–39 rated same-gender sex as appealing. This pattern reverses for women. About 4.3% of female respondents ages 20–24 and 7.0% of respondents ages 30–39 rated same-gender sex as appealing. Since same-gender attraction for women in Add Health was considerably bigger than that in the NHSLS, it is doubtful that female homosexuality is significantly underreported in Add Health.

³To measure homosexuality, Blanchard and Bogaert (1996) used self-reported sexual identity, pooling homosexual and bisexual subjects. Bogaert (2006) used a combination of self-reported measures of sexual attraction and behavior.

⁴ The survey item is, “Choose the description that best fits how you think about yourself: 100% heterosexual; mostly heterosexual, but somewhat attracted to people of your own sex; bisexual, that is, attracted to men and women equally; mostly homosexual, but somewhat attracted to people of the opposite sex; or 100% homosexual.”

⁵To estimate what fraction of respondents age 13, 14, etc., had a younger sibling born between Waves I and III, I examine respondents age 18 in Wave I. 1.78%, 1.28%, 1.00%, 0.50%, and 0.39% of respondents age 18 had a younger sibling at least 14, 15, 16, 17, and 18 years younger, respectively. This suggests that less than 1% of respondents in the sample had a younger sibling who is not counted.

⁶Specifically, 8,536 (79.5%) respondents had no older brothers, 1,923 (17.9%) had one, 250 (2.3%) had two, 25 (0.2%) had three, 0 had four, and 1 (0.0%) had five.

Note. Numbers in parentheses are robust standard errors. An asterisk indicates significance at the 5% level. The sample size associated with the first two columns is 5,003. The sample size associated with the last two columns is 4,981. Data source: Add Health (Waves I and III).

⁷By far, most respondents who have an older brother had just one (87%).

Note. Numbers in parentheses are robust standard errors. An asterisk indicates significance at the 5% level. The sample size associated with the first two columns is 5,656. The sample size associated with the last two columns is 5,617. Data source: Add Health (Waves I and III).

Note. Numbers in parentheses are robust standard errors. An asterisk indicates significance at the 5% level. The sample size associated with the first two columns is 4,990. The sample size associated with the last two columns is 4,968. Data source: Add Health (Waves I and III).

Note. Numbers in parentheses are robust standard errors. An asterisk indicates significance at the 5% level. The sample size associated with the first two columns is 5,646. The sample size associated with the last two columns is 5,606. Data source: Add Health (Waves I and III).