Skip to main content
Log in

New Data on Birth Order in Homosexual Men and Women and a Reply to Vilsmeier et al. (2021a, 2021b)

  • Original Paper
  • Published:
Archives of Sexual Behavior Aims and scope Submit manuscript

Abstract

The fraternal birth order effect (FBOE) is the repeated finding that older brothers increase the odds of homosexuality in later-born males. It has been our working assumption, based on the majority of previous studies, that a similar FBOE does not occur in females. In an elaborate quantitative review posted last year to a preprint server, Vilsmeier et al. (2021a) concluded that there is no valid evidence for an FBOE in men or women. Ablaza et al. (2022) subsequently published a study of population-level data from the Netherlands with conclusions completely opposite to those of Vilsmeier et al., namely, that there is robust evidence of an FBOE in both men and women. The present research was initially undertaken to refute the assertion of Vilsmeier et al. that there is no proof of an FBOE in men and to investigate how they obtained such a discrepant conclusion. We found evidence that the discrepancy may relate to Vilsmeier et al.’s use of the large and demonstrably unreliable sample published by Frisch and Hviid (2006). After the publication by Ablaza et al., we expanded our article to address their finding of an FBOE in women. We argue that our preferred explanation of the FBOE in men—that it reflects the progressive immunization of some mothers to Y-linked antigen by each succeeding male fetus and the concomitantly increasing effects of anti-male antibody on sexual differentiation in the brain in each succeeding male fetus—could plausibly be extended to female homosexuality.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Notes

  1. This is our formulation of Vilsmeier et al.’s assertion; they do not express their point in exactly these words. It is unclear what else they could mean, however. Vilsmeier et al. do not appear to claim that homosexual men have no excess of older brothers; rather, they argue that the excess of older brothers is no greater than the excess of older sisters. In other words, they seem to state that homosexual men have an excess of older siblings, not an excess of older brothers.

    Our interpretation of Vilsmeier et al.’s conclusion is based on statements such as these, all from Vilsmeier et al.’s (2021a) Introduction: “researchers are bound to frequently draw incorrect conclusions about the presence of an excess of older brothers but not sisters”.... “For now, we show that the OBOR, MROB, and MPOB [derived measures of the FBOE used in previous studies] are measures of the general excess of older siblings (i.e., brothers and sisters) among homosexual men, rather than measures of the more specific excess of older brothers only”.... “this estimated OBOR appears compatible with a model of an excess of older siblings, rather than a genuine excess of older brothers among homosexual men”.... “Consequently any positive association between the probability of homosexual orientation and the odds (or probability) of observing an older brother among all of the siblings of a participant may be compatible with both scenarios, i.e., a genuine excess of older brothers but also with an excess of older brothers and sisters.”

  2. Again, this is our formulation of Vilsmeier et al.’s assertion. The following sentences, from the Introduction to Vilsmeier et al. (2021a), are one expression of their view of the similarity of homosexual men and women: “Suppose it turned out that an excess of older brothers among men was much more inconsistent than previously believed by researchers, due to the previous evidence for such an excess being due to misapplications of statistical methods. Suppose further that there was a similarly-sized excess of older brothers among homosexual women. These observations combined would strip the FBOE and the MIH of its (sic) verisimilitude. In what follows, we provide converging evidence for both of these suppositions.”

    The notion that the sibship composition of homosexual women is the same as the sibship composition of homosexual men may also be inferred from the statement that their findings are “consistent with both an excess and a lack of older brothers among homosexual men and women” (p. 18). The notion that homosexual men have an excess of older siblings rather than an excess of older brothers may be inferred from the statements quoted in Footnote 1. If homosexual women are like homosexual men, and if homosexual men have an excess of older siblings (rather than older brothers), then homosexual women must also have an excess of older siblings (rather than older brothers).

  3. We decided to respond to Vilsmeier et al.’s preprint rather than wait for their published version because we were uncomfortable with allowing the preprint to circulate unchallenged for an indefinite amount of time. As of January 7, 2022, the preprint had been downloaded 277 times from the PsyArXiv preprint server and viewed 50 times on Vilsmeier’s ResearchGate page. It has already been cited in at least one published article (Fořt & Kaňková, 2021). This is more attention than some published articles receive; see Halevi and Moed (2014).

  4. This change was made in order to increase the number of subjects that could be extracted from the 14 smaller samples.

  5. In Table 1 and elsewhere throughout this article, the word “proband” is used interchangeably with “subject.” The word “proband”—the index case who brings his or her family into the study—is commonly used instead of “subject” in the FBOE literature.

    There is some continuing rationale for this usage beyond historical tradition. Many studies in this research area examine the sibling sex ratio, the ratio of brothers to sisters collectively reported by a group of homosexual or heterosexual individuals. In these analyses, the “cases” (in the statistical sense) are usually the individual’s siblings rather than the individual him- or herself. This could potentially lead readers to some confusion: Who are the subjects in the analysis—the examined individual or the examined individual’s brothers and sisters? The use of the proband–sibling terminology might help to forestall such potential confusion.

  6. These were feminine or transgender samples: Blanchard and Sheridan (1992), Blanchard et al. (1995), Blanchard et al. (1996), Bozkurt et al. (2015), Gómez-Gil et al. (2011), Green (2000), Khorashad et al. (2020), Schagen et al. (2012), VanderLaan et al. (2014, 2017), VanderLaan and Vasey (2011), Vasey and VanderLaan (2007).

  7. The selection criteria for these study groups (numbers 5, 6, and 7) were generated using the same algorithm as in Blanchard and Lippa (2021). The criteria for study groups 1 and 2 were modified to create groups 5 and 6 by adding “ ≥ 1” to each sibling category that was not required to be a zero. Thus, 1 older brother, 0 older sisters, and 0 younger siblings (group 1) became ≥ 2 older brothers, 0 older sisters, and ≥ 1 younger sibling (group 5). In order to keep family size constant in the comparisons involving larger families, group 7 had to have ≥ 3 younger siblings.

  8. https://www.ukbiobank.ac.uk/.

References

Download references

Acknowledgements

The authors thank Anthony F. Bogaert, Adam J. MacNeil, Doug P. VanderLaan, and Paul L. Vasey for their input on earlier versions of the manuscript.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Ray Blanchard.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical Approval

This study was carried out on archived data. Institutional ethics approvals were obtained by the original investigators.

Informed Consent

Informed consent, where relevant, was obtained from individual participants by the authors of the original studies.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix 1: Reanalysis of Vilsmeier et al.’s (2021a) Tables 5 and 6

Vilsmeier et al.’s conclusions regarding male subjects are based on the 64 samples listed in their Table 5. Our conclusions are based on the 43 samples collectively employed in the present Studies 1–3. The various reasons for the difference of 21 samples are presented in Table

Table 9 Reasons for the difference between the number of samples in Vilsmeier et al.’s Table 5 and the combined number of samples in the present Studies 1–3

9. Almost half the discrepancy (nine instances) were samples that we excluded because they overlapped with other samples. It is noteworthy that all four samples that we excluded because of probable stopping rule distortion were East Asian (two Chinese, one South Korean, and one Thai).

In “Appendix 1,” we report analyses of Vilsmeier et al.’s own choice of samples to support our contention that their discrepant conclusion regarding the FBOE in males can be traced to their use of the problematic Frisch and Hviid (2006) sample. To conduct these analyses, we converted the means presented in Vilsmeier et al.’s Tables 5 and 6 into sums of older brothers and older sisters by multiplying each mean by the number of subjects on which it was computed and then rounding up or down to the nearest integer.

Before performing this conversion, we corrected data for three samples in Table 5 that were obviously decimal errors (e.g., a mean of 957.00 older brothers for Nila et al.’s homosexual males). After performing the conversion, we randomly spot-checked data for five studies to ensure that the sums of older brothers and older sisters that we computed were identical or close to the original data. The sums matched or else were off by an amount that seemed consistent with rounding error. We therefore did no further checking and used the data from Tables 5 and 6 exactly as is, except for the three samples mentioned above. For the purpose of making additional comparisons, we included data from Ablaza et al. (2022), which was provided by that study’s first author at our request (C. Ablaza, personal communication, August 30, 2021).

The data are presented in Table

Table 10 Numbers of probands (subjects) and their siblings from Vilsmeier et al. (2021a), full sample and two subsamples, and from Ablaza et al. (2022)

10. The cell entries are the grand totals of probands (i.e., subjects examined in the original studies) and the grand totals of older brothers and older sisters that those subjects reported. We then carried out a series of meta-analyses, based on these data, using the same procedure that we used in Studies 2 and 3.

The results are presented in Figs.

Fig. 5
figure 5

Proportions of older brothers, calculated from Vilsmeier et al. (2021a) Table 5 (male subjects) and Table 6 (female subjects). HET = heterosexual subjects. HOM = homosexual subjects. ALL = including all subjects in Vilsmeier et al.’s Tables 5 and 6. W/O F&H = without the subjects from Frisch and Hviid (2006)

5 and

Fig. 6
figure 6

Proportions of older brothers from Vilsmeier et al. (2021a) and Ablaza et al. (2022). W/O B&L = Vilsmeier et al. (2006) minus the subjects from Frisch and Hviid (2006) and minus the subjects from Blanchard and Lippa (2007)

6, which are interpreted the same way as Figs. 2, 3, and 4 in the main text. The upper half of Fig. 5 shows the proportions of brothers among the older siblings of the 64 male and 17 female samples studied by Vilsmeier et al. (2021a). Each proportion is bracketed by its 95% confidence interval. The vertical reference line at 51.5% marks the proportion of brothers expected for a sample of probands drawn at random from the general population. If the confidence interval for a given proportion crosses that reference line, then it does not differ significantly from the human sex ratio at birth (that is, 0.515, or 106 male births per 100 female births). The lower half of Fig. 5 shows the proportions of brothers when the Frisch and Hviid (2006) sample is omitted from Vilsmeier et al.’s data.

Figure 5 shows that removing the Frisch and Hviid (2006) sample from the Vilsmeier et al. data eliminated the impossibly high older sibling sex ratio for the heterosexual males. It also lowered the impossibly high sibling sex ratio for the heterosexual females, but it still left a significant excess of older brothers for the heterosexual females. Inspection of the data indicated that this excess was primarily caused by the Blanchard and Lippa (2007) sample. This finding might relate to the operation of parental stopping rules in that sample, which produced a general tendency for the male subjects to have excess older sisters and the female subjects to have excess older brothers (Blanchard & Lippa, 2007).

The analyses depicted in Fig. 6 tested our conjecture that the excess of older brothers for heterosexual females that remained after the exclusion of the Frisch and Hviid (2006) sample could be attributed to the Blanchard and Lippa (2007) sample. The upper half of the figure shows the Vilsmeier et al. data after both the Frisch and Hviid (2006) and Blanchard and Lippa (2007) samples were omitted. The lower half of the figure presents the sex ratio data from Ablaza et al. (2022).

Figure 6 shows that removal of the Blanchard and Lippa (2007) sample from the Vilsmeier et al. (2021a) data lowered the older sibling sex ratio for the heterosexual females to the expected value for the general population. This also lowered the sibling sex ratio for the homosexual females to a value well below the expected population ratio, a result we had not anticipated. With both the Frisch and Hviid (2006) and Blanchard and Lippa (2007) samples removed from Vilsmeier et al.’s data, the findings for the heterosexual males, homosexual males, and heterosexual females resemble the findings of Ablaza et al. (2021), but the findings for the homosexual females are strikingly different.

Appendix 2: Summary of Recent (2021–2022) Findings

While the present article was under review, another major study of sibship composition and homosexuality was posted on a preprint server. Raymond et al. (2022) conducted analyses of male subjects using both aggregated and individual data. The addition of Raymond et al. brings to four the number of non-affiliated research teams or research networks who have recently studied this topic, each using its own, novel statistical methodology. We felt that this was a sufficient number to justify a box score summary of recent findings. Such a summary may be useful for evaluating the state of past research and for identifying which questions most need addressing in future research.

Male Subjects

Fraternal Birth Order Effect (FBOE). Ablaza et al. (2022), Blanchard and Lippa (2021), and Raymond et al. (2022) found evidence for an FBOE. Vilsmeier et al. (2021a) did not.

Sororal Birth Order Effect (SBOE). Ablaza et al. (2022) and Blanchard and Lippa (2021) found evidence for an SBOE. Raymond et al. (2022) and Vilsmeier et al. (2021a) did not.

Fraternal–Sororal Effect Differential (FSED). Ablaza et al. (2022) and Blanchard and Lippa (2021) found evidence that the magnitude of the FBOE is reliably larger than that of the SBOE. The point is moot for Raymond et al. (2022) and Vilsmeier et al. (2021a), because they did not find evidence of an SBOE.

Female Fecundity Effect (FFE). Ablaza et al. (2022), Blanchard and Lippa (2021), and Raymond et al. (2022) all investigated the fecundity of female relatives of their subjects and found that it was not greater for the relatives of the homosexual subjects. Vilsmeier et al. (2021a) did not investigate this.

Female Subjects

Fraternal Birth Order Effect (FBOE). Ablaza et al. (2022) found evidence for an FBOE in females. Blanchard (Blanchard, 2022; Blanchard & Skorska, present study) and Vilsmeier et al. (2021a) did not. As previously indicated, Raymond et al. (2022) did not study female subjects.

Sororal Birth Order Effect (SBOE). Ablaza et al. (2022) found evidence for an SBOE in females. Blanchard (Blanchard, 2022; Blanchard & Skorska, present study) and Vilsmeier et al. (2021a) did not.

Fraternal–Sororal Effect Differential (FSED). Ablaza et al. (2022) found evidence that the magnitude of the FBOE is reliably larger than that of the SBOE. The point is moot for Blanchard (Blanchard, 2022; Blanchard & Skorska, present study) and Vilsmeier et al. (2021a), because they did not find evidence of an SBOE.

Female Fecundity Effect (FFE). Ablaza et al. (2022) and Blanchard (Blanchard, 2022; Blanchard & Skorska, present study) investigated the fecundity of female relatives of their female subjects and found that it was not greater for the relatives of the homosexual subjects. Vilsmeier et al. (2021a) did not investigate this.

Our reading of the foregoing summary is that there is most agreement regarding the absence of an FFE in male or female subjects. The second largest agreement is that there is an FBOE among male subjects. The most urgent question for future research is whether the finding of an FBOE among female subjects can be confirmed in another large sample.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Blanchard, R., Skorska, M.N. New Data on Birth Order in Homosexual Men and Women and a Reply to Vilsmeier et al. (2021a, 2021b). Arch Sex Behav 51, 3319–3349 (2022). https://doi.org/10.1007/s10508-022-02362-z

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10508-022-02362-z

Keywords

Navigation