Birth Order and Family Size of UK Biobank Subjects Identified as Asexual, Bisexual, Heterosexual, or Homosexual According to Self-Reported Sexual Histories

Abstract

This study used a recently developed statistical technique to investigate the relations between various elements of a subject's family background and the odds of that subject reporting a sexual history indicative of a minority sexual orientation. The subjects were 78,983 men and 92,150 women who completed relevant questionnaire items in the UK Biobank, a large-scale biomedical database of volunteers aged 40-69 years. The men were divided into three sexual minority groups-homosexual, bisexual, and asexual-and a comparison group of heterosexual men. The same was done for the women. The analytic procedure consisted of logistic regressions specifically designed to disentangle the effects of birth order and family size. The results showed that older brothers increased the odds of homosexuality in both men and women, and that older sisters increased the odds in men. In contrast, neither older brothers or older sisters affected the odds of bisexuality or asexuality in men or in women. These results suggest that birth order effects may be specific to homosexuality and not common to all minority orientations. The only family size finding was the negative association between family size and the odds of asexuality in both men and women. The outcomes of this study indicate that the maternal immune hypothesis, which was advanced to explain the relation between older brothers and homosexuality in later-born males, might have to be abandoned or else expanded to explain the findings concerning females. A few such modifications are considered.

Keywords: Asexuality; Birth order; Bisexuality; Family size; Homosexuality; Maternal immune hypothesis.

Fig. 1

Calculation of sibship composition. Younger = number of younger siblings, older = number of older siblings, brothers = total number of brothers, sisters = total number of sisters, OB = older brothers, OS = older sisters, YB = younger brothers, YS = younger sisters. The “sample shares” are the percentages of subjects whose sibship composition (i.e., number of older brothers, older sisters, and so on) was computed by the corresponding path within the flow chart. The sample share data are given purely to help describe the subjects; sample share is not a variable that is used in any data analysis. The sample shares were computed on the subset of UK Biobank participants who responded to each of the sibling questions used in the assignment protocol, and who reported the number of their sexual partners

Fig. 2

Predicted changes in the odds of a sexual minority preference associated with the explanatory variables of the preferred logistic regression model (see Table 7). Data sourced from the UK Biobank. OBSE (older brother swap effect) = effect of older brothers on the odds of a sexual minority orientation. OSSE (older sister swap effect) = effect of older sisters. BSSE (brother–sister swap effect) = difference in magnitude between the effect of older brothers and the effect of older sisters. YSAE (younger sibling addition effect) = effect of family size. Whiskers denote 95% robust confidence intervals. Estimates that are significantly different from the no-effect value (1.0) at 5% confidence level are denoted by shaded bars. To maintain consistent scaling across charts, the upper part of the confidence interval corresponding to BSSE on male asexuality has been truncated

Fig. 3

Predicted changes in the odds of homosexual preference associated with the explanatory variables of the conventional logistic regression model. Data sourced from the UK Biobank. OB (older brothers) = effect of adding an older brother to the sibship. OS (older sisters) = Effect of adding an older sister. YB (younger brothers) = Effect of adding a younger brother. YS (younger sisters) = Effect of adding a younger sister. Whiskers denote 95% robust confidence intervals. Estimates that are significantly different from the no-effect value (1.0) at 5% confidence level are denoted by shaded bars

Fig. 4

Comparison of the coefficient estimates corresponding to the logistic regression model of homosexual preference (UK Biobank sample) and the logistic regression model of same-sex union entry (Dutch population, as reported by Ablaza et al., 2022). Data sourced from the UK Biobank. OBSE (older brother swap effect) = effect of older brothers on the odds of a sexual minority orientation. OSSE (older sister swap effect) = effect of older sisters. BSSE (brother–sister swap effect) = difference in magnitude between the effect of older brothers and the effect of older sisters. YSAE (younger sibling addition effect) = effect of family size. Whiskers denote 95% robust confidence intervals. Estimates that prove significantly different from the no-effect value (1.0) at 5% confidence level are denoted by shaded bars