|
Article Excerpt
Introduction
Recent research on family and sexuality concludes that the only biodemographic correlate of male sexual orientation is the number of older brothers (Blanchard & Bogaert, 1996; Bogaert, 2006). Nearly all studies find that older sisters have no effect on male homosexuality (Blanchard, 2001; Blanchard & Bogaert, 1996; Blanchard, Zucker, Siegelman, Dickey, & Klassen, 1998; Ellis & Blanchard, 2001). There is little to no evidence that any biodemographic variables correlate with female sexual orientation (Blanchard, 1997; Bogaert, 1997). These studies complement earlier studies that relate male homosexuality to birth order and intrafamily sex ratios (Blanchard & Sheridan, 1992; Blanchard & Zucker, 1994; Hare & Moran, 1979; Slater, 1962; Zucker & Blanchard, 1994).
It is hypothesized that the causal mechanism is purely biological (Blanchard, 2001; Blanchard & Bogaert, 1996; Bogaert, 2006). The maternal immune hypothesis states that a woman's immune system "remembers" the number of male, but not female, fetuses. A mother's immune system responds differently to each successive male child in that each child receives different levels of maternal antibodies, which determine the sexual differentiation of the brain. The maternal immune hypothesis is generally consistent with biological evidence on the influence of prenatal hormones on sexual orientation (Gladue, Green, & Hellman, 1984; LeVay, 1991; Swaab & Hofman, 1990).
In this article, 1 empirically identify the family-demographic correlates of sexual orientation in men and women and, thereby, reevaluate the maternal immune hypothesis. Using a nationally representative sample of men and women mostly ages 20 to 24, I regress several measures of sexual orientation on a set of sibling variables with and without additional controls. For a number of reasons, the empirical strategy in this article improves upon the literature. Previous studies do not use representative datasets. Moreover, the sample size in this article is about five times larger than the sample size associated with the best-known studies of family and sexual orientation. (1) I allow for nonlinear effects in the number of siblings, while other studies assume linearity. Also, I control for potential omitted variables, that is, family structure, ethnicity, and education, which other studies ignore.
For men, I find that having one older brother does not raise the likelihood of homosexual behavior, desire, or identity. While having multiple older brothers has a positive effect, it is not significant at the 5% level. In contrast, having any older sisters lowers the likelihood of homosexual or bisexual identity. Growing up without a biological parent is positively associated with homosexuality. Being Black and having less than a high school education are each positively related to homosexual behavior and desire, whereas having a college education is positively related to homosexual identity. For women, I find that having an older brother, as well as having any sisters, is negatively associated with homosexuality. Growing up without a biological parent is also positively correlated with homosexuality. Being Black or other race is negatively related to female homosexual desire and identity, while not graduating from high school is positively related.
Therefore, the maternal immune hypothesis is unable to explain the entire pattern of family-demographic correlates reported in this article. Contrary to the theory, having an older brother does not have a significant effect on male homosexuality, whereas having an older sister does. The theory does not account for the results on homosexuality in women. Plus, several background characteristics, that is, family structure, race, and education, also may relate to sexual orientation. Either biological or social theories of sexual orientation may help to explain and interpret the findings. Given the complexity of the empirical relationship between sexual orientation and the biodemographic and other correlates, it is likely that both biological and social mechanisms may play a role.
The remainder of the article is organized as follows. The next section describes the data, variables, and empirical strategy. Empirical results are presented in the Results Section. The final section discusses the results and concludes this article.
Data and Methods
In the empirical analysis, I linearly regress measures of sexual orientation on a set of sibling variables both with and without controls, which include respondent age and dummy variables for family structure, ethnicity, and education. The tables display the regression coefficients, robust standard errors associated with the coefficients, and asterisks indicating whether each coefficient is significantly different from zero at the 5% level. I now introduce the micro dataset, describe the dependent and independent variables, and display summary statistics.
I use the National Longitudinal Study of Adolescent Health (Add Health), a nationally representative study of adolescent health in the United States (Udry, 2003). Adolescents in grades 7 through 12 were initially interviewed in 1995 and 1996 (Waves I and II) and were reinterviewed in 2001 (Wave III). The sample size of male respondents is about 5,000, and the sample size of female respondents is about 5,600. Table 1 displays summary statistics. At Wave III, all respondents in the sample were 18 years old or older. About 88% were between the ages of 20 and 24. It is possible that some young adults may not yet have recognized or acknowledged same-gender attraction. Nevertheless, the evidence suggests that underreporting of homosexuality is not a problematic issue in Add Health. (2)
I construct four measures of homosexuality, including behavior, desire, and identity. (3) The first measure (behavior) is...
|
