Advertisement

Archives of Sexual Behavior

, Volume 47, Issue 1, pp 1–15 | Cite as

Fraternal Birth Order, Family Size, and Male Homosexuality: Meta-Analysis of Studies Spanning 25 Years

  • Ray Blanchard
Target Article

Abstract

The fraternal birth order effect is the tendency for older brothers to increase the odds of homosexuality in later-born males. This study compared the strength of the effect in subjects from small versus large families and in homosexual subjects with masculine versus feminine gender identities. Meta-analyses were conducted on 30 homosexual and 30 heterosexual groups from 26 studies, totaling 7140 homosexual and 12,837 heterosexual males. The magnitude of the fraternal birth order effect was measured with a novel variable, the Older Brothers Odds Ratio, computed as (homosexuals’ older brothers ÷ homosexuals’ other siblings) ÷ (heterosexuals’ older brothers ÷ heterosexuals’ other siblings), where other siblings = older sisters + younger brothers + younger sisters. An Older Brothers Odds Ratio of 1.00 represents no effect of sexual orientation; values over 1.00 are positive evidence for the fraternal birth order effect. Evidence for the reliability of the effect was consistent. The Older Brothers Odds Ratio was significantly >1.00 in 20 instances, >1.00 although not significantly in nine instances, and nonsignificantly <1.00 in 1 instance. The pooled Older Brothers Odds Ratio for all samples was 1.47, p < .00001. Subgroups analyses showed that the magnitude of the effect was significantly greater in the 12 feminine or transgender homosexual groups than in the other 18 homosexual groups. There was no evidence that the magnitude of the effect differs according to family size.

Keywords

Birth order Homosexuality Maternal immune hypothesis Meta-analysis RevMan Sexual orientation Transgender 

Introduction

It has now been 25 years since I published my first study on birth order and sexual orientation in males (Blanchard & Sheridan, 1992). That study began a gradual revival of research interest in a topic sporadically revisited for decades with no particular conclusion on whether or why homosexual men have a different average birth order than heterosexual men (see historical review by Blanchard, 1997).

Research in the succeeding years established several important facts. First, homosexual men do have a higher birth order (i.e., later births) than heterosexual men, and this difference can be attributed to their greater number of older brothers (Blanchard & Bogaert, 1996b). Older brothers increase the odds of homosexuality in later-born males, whereas older sisters, younger brothers, and younger sisters have no effect on those odds. This phenomenon has therefore been termed the fraternal birth order effect (FBOE).

The second finding was that biological brothers increase the odds of homosexuality in later-born males even if they were reared in different households, whereas stepbrothers or adoptive brothers have no effect on sexual orientation (Bogaert, 2006). Thus, the most direct evidence currently available indicates that the effect of older brothers on younger brothers is mediated by some long-term change in their mother’s uterine environment. To my knowledge, no one has ever attempted to repeat Bogaert’s (2006) study. However, its implication that the FBOE operates prenatally rather than postnatally is indirectly supported by evidence that homosexual males with older brothers have lower than expected birth weights (Blanchard & Ellis, 2001; Blanchard et al., 2002; VanderLaan, Blanchard, Wood, Garzon, & Zucker, 2015).

The third finding was that the FBOE occurs in different cultures and in widely separated geographic regions. The FBOE has been demonstrated in Brazil (VanderLaan et al., 2016), Canada (Blanchard & Bogaert, 1996b), Iran (Khorashad et al., 2017), Italy (Camperio-Ciani, Corna, & Capiluppi, 2004), The Netherlands (Schagen, Delemarre-van de Waal, Blanchard, & Cohen-Kettenis, 2012), Independent Samoa (VanderLaan & Vasey, 2011), Spain (Gómez-Gil et al., 2011), Turkey (Bozkurt, Bozkurt, & Sonmez, 2015), the UK (King et al., 2005), and the U.S. (Schwartz, Kim, Kolundzija, Rieger, & Sanders, 2010).

I have previously authored or co-authored three meta-analyses of studies on the FBOE. Each of these meta-analyses addressed a specific question or targeted a specific subset of the literature, and each used a completely different methodology. Jones and Blanchard (1998) addressed the narrow question of whether older sisters have no effect on sexual orientation in later-born males, or simply a weaker effect than older brothers. To this end, we developed competing mathematical models of the two possibilities and compared the fit of these models to the empirical data available at that time. Jones and Blanchard concluded that any tendency for homosexual males to be born later among their sisters is, in effect, a statistical artifact of their tendency to be born later among their brothers.

In Blanchard (2004), I reported a meta-analysis of studies from my own research program. These studies included 10,143 male subjects (3181 homosexuals and 6962 heterosexuals). The results reinforced the conclusion that homosexual men have more older brothers than do heterosexual men and the conclusion that differences in all other sibship parameters (older sisters, younger brothers, or younger sisters) are secondary consequences of the difference in older brothers.

Blanchard and VanderLaan (2015) performed a meta-analysis of studies in which neither we nor any of our customary collaborators carried out the collection or the analysis of the original data. The purpose of this selection criterion was to produce results that would be as free as possible from our own potential unconscious biases. Although the data-analytic procedure was different from that used by Blanchard (2004), the conclusions were closely similar.

Among the remaining questions that might be answered using meta-analysis of published or archived data is whether the magnitude of the FBOE is greater in male samples with larger or with smaller family sizes.1 Previous research focused on mathematical statistics does not generate consistent predictions for this precise question. Suarez and Przybeck (1980) based their mathematical modeling on the assumption of a causal relation between excess brothers and male homosexuality. Although their equations assume that all brothers rather than older brothers influence a subject’s sexual orientation, the form of their reasoning seems to hold if the relevant parameters concern older siblings instead of all siblings. They developed two different models, both of which predicted that the proportion of brothers in the sibships of homosexual men should decrease as family size increases. Both of their models provided a good fit to the archival data they examined. Suarez and Przybeck did not offer any opinion on whether comparisons of homosexual and heterosexual groups would be more likely to find significant differences if the compared groups had large family sizes or small family sizes. The thrust of their analysis, however, seems to imply that the FBOE would be easier to detect if the compared groups had small family sizes. That is because homosexual men from small families would have the greatest proportion of older brothers (see Suarez & Przybeck, 1980; Fig. 2).

The studies by Cantor, Blanchard, Paterson, and Bogaert (2002) and Bogaert (2004) were based on the assumption of a causal relation between older brothers and male homosexuality. The parameters of their equations were estimated directly from empirical comparisons of homosexual and heterosexual men. Cantor et al. concluded that the proportion of homosexual men who can attribute their sexual orientation to older brothers increases with the number of older brothers, rising from 0% in groups of homosexual men with no older brothers to over 50% in groups with three or more older brothers. Cantor et al. also, however, offered no opinion on the implications for between-groups comparisons. In contrast to the analyses by Suarez and Przybeck, the analyses by Cantor et al. and Bogaert (2004) seem to imply that the FBOE would be easier to detect if the compared groups had large family sizes. That is because homosexual men who owe their sexual orientation to the FBOE would be most likely to come from large families.

Adding to the difficulty of making theory-based predictions about the optimal family size for studying the FBOE is the possibility that the relative prevalence of other etiologic factors in sexual orientation may also vary with family size. One such hypothetical factor would be a genetic variant that predisposes males to homosexuality but predisposes their female relatives to greater fertility, perhaps by increasing sexual attraction to men in both groups. The hypothesized existence of one or more such genetic variants is sometimes associated with the balancing selection hypothesis of homosexuality, that is, the proposition that the smaller number of direct descendants of homosexual men, compared with heterosexual men, is offset by a larger number of other close biological relatives. Several studies have, in fact, confirmed the prediction of the balancing selection hypothesis that homosexual research subjects will report larger family sizes than heterosexual control subjects (e.g., Iemmola & Camperio Ciani, 2009; King et al., 2005; Schwartz et al., 2010; VanderLaan & Vasey, 2011; but see Blanchard, 2012; Kishida & Rahman, 2015). If “gay genes” with these effects are one (but not the only) cause of homosexuality in males, we might expect that homosexual men from larger families would be more likely to have acquired their sexual orientation through this direct genetic route. Such men would have average birth orders, both in relation to brothers and in relation to all siblings, and thus would dilute evidence of the FBOE. This possibility further complicates predictions about the relative magnitude of the FBOE in large versus small families. For these reasons, the most certain way to establish the relation (if any) between family size and the FBOE is through empirical rather than theoretical or deductive means.

The relation between mean family sizes and the magnitude of the FBOE is of practical relevance both for evaluating the results of published studies and for planning future ones. I therefore conducted the present meta-analysis partly as a step toward answering this question.

This study had a second goal, namely to investigate whether the FBOE is greater in feminine or transgender homosexual males than in relatively masculine homosexual males. Wampold (2013) conjectured that fraternal birth order correlates positively with the preference for the receptive role in anal intercourse and that the preference for the receptive role in anal intercourse correlates positively with femininity in homosexual males (see, for example, Zheng, Hart, & Zheng, 2012), from which it should follow that fraternal birth order correlates positively with femininity in homosexual males. Swift-Gallant, Coome, Monks, and VanderLaan (2017), in an empirical study, did find a positive correlation between fraternal birth order and the preference for the receptive role in anal intercourse as well as a positive correlation between the preference for the receptive role in anal intercourse and femininity in homosexual males. They did not, however, find a significant correlation between fraternal birth order and femininity. Several other studies have investigated whether the FBOE is greater in feminine homosexual males, and none of them found evidence for such a relation (Bogaert, 2003, 2005a; Kishida & Rahman, 2015; Semenya, VanderLaan, & Vasey, 2017). However, the meta-analytic methodology of the present study offered a different and potentially more powerful approach to the question, and the answer to this question is still pertinent to interpreting the findings of past studies and to planning future ones.

In order to establish confidence in the answers to my first two questions, I first needed to demonstrate that the FBOE was clearly present in the material assembled for investigation. This was accomplished by adding a third goal to the study, namely conducting an updated meta-analysis on the magnitude and reliability of the FBOE itself.

Method

Studies and Samples Analyzed

The analysis consisted of 60 groups, 30 homosexual2 and 30 heterosexual, from 26 studies. My second meta-analysis (Blanchard, 2004) contributed 28 of these. My third meta-analysis (Blanchard & VanderLaan, 2015) contributed 10 more. I could not use all the groups from that meta-analysis because several studies did not report the mean older brothers, older sisters, and so on, for their groups,3 and because 10 groups from Rahman and his colleagues were included in a later study (Kishida & Rahman, 2015), which I used for the present analysis.

After I retrieved the usable data from studies in my two previous meta-analyses, I searched the published literature and my own archival data for studies that met the following criteria:
  1. 1.

    The study sample consisted entirely of singleton births, or else the number of multiple births was simply the (small) number expected by chance.

     
  2. 2.

    Homosexual orientation was classified according to the subject’s self-report or from phallometric testing and not from proxy variables like same-sex marriage.

     
  3. 3.

    Means for all four types of siblings (older brothers, older sisters, younger brothers, and younger sisters) were reported by the publishing authors or else were computable from archival data in my files.

     
  4. 4.

    The reported means were computed directly from raw data and not estimated from incomplete data.

     
  5. 5.

    The means were not known to be distorted by stoppage rules (e.g., Blanchard & Lippa, 2007).

     
This search produced an additional 22 groups.4 The descriptions of the 60 groups are given in Table 1. The number of subjects in each group is also reported in this table. All data used in this study are presented in the Appendix.
Table 1

Studies in the meta-analysis

Authors

Description of the sample

N homosexual

N heterosexual

Blanchard and Bogaert (1996a)

American volunteers over age 18 interviewed by Alfred Kinsey and associates from 1938 to 1963 (see Gebhard & Johnson, 1979)

799

3807

Blanchard and Bogaert (1996b)

Canadian volunteers

302

434

Blanchard and Bogaert (1998)

American sex offenders against adults, from earlier study by Gebhard, Gagnon, Pomeroy, and Christenson (1965)

156

173

Blanchard and Bogaert (1998)

American sex offenders against pubescents, from earlier study by Gebhard et al. (1965)

69

127

Blanchard and Bogaert (1998)

American sex offenders against children, from earlier study by Gebhard et al. (1965)

42

143

Blanchard et al. (2000)

Canadian pedophilic and hebephilic patients (men attracted to prepubescent and pubescent children, respectively)

65

152

Blanchard et al. (2006, Table 1)

Canadian patients referred to a specialty clinic for assessment of their sexual feelings or behaviors: “Blanchard” subsample

92

672

Blanchard et al. (2006, Table 1)

Canadian men (e.g., adoptees) reared in environments other than biological families: “Bogaert (non-biological families)” subsample

280

222

Blanchard et al. (2006, Table 1)

Canadian homosexual community volunteers and heterosexual university students: “Bogaert (other)” subsample

267

148

Blanchard and Sheridan (1992)

Canadian outpatients referred for assessment of gender dysphoria (roughly, transsexualism)

193

273

Blanchard and Zucker (1994)

American volunteers from earlier study by Bell, Weinberg, and Hammersmith (1981)

569

281

Blanchard et al. (1995)

Canadian homosexual or prehomosexual gender-dysphoric child and adolescent patients, matched clinical control boysa

156

156

Blanchard et al. (1996)

Dutch gender-dysphoric patients, adult and adolescent samples combined

104

79

Blanchard et al. (1998)

British and American volunteers, from earlier studies by Siegelman (1972, 1973, 1974, 1978, 1981)

385

225

Bogaert et al. (1997)

Canadian pedophilic patients

68

57

Bozkurt et al. (2015, Table 1)

Turkish homosexual transsexuals and heterosexual cissexuals psychiatrically evaluated for compulsory military service

60

61

Currin et al. (2015a, Table 1; 2015b)

American (or presumably American) volunteers

118

500

Ellis and Blanchard (2001)

American and Canadian volunteers

175

971

Gómez-Gil et al. (2011, p. 507)

Spanish male-to-female transsexuals, dichotomously classified, according to their natal sex, as homosexual or non-homosexual (heterosexual, bisexual, and asexual)b

287

38

Green (2000, Table 3, p. 792)

UK male-to-female transsexuals, dichotomously classified, according to their natal sex, as homosexual or non-homosexual (heterosexual, bisexual, and asexual)b

106

336

Khorashad et al. (2017)

Iranian homosexual male-to-female transsexuals and heterosexual cissexual psychiatric patients

92

72

King et al. (2005, pp. 119–121)

UK men attending clinics for sexually transmitted infections.

301

404

Kishida and Rahman (2015, Table 1)

UK volunteers

905

999

Schagen et al. (2012, Table 4)

Dutch biologically male, peripubertal gender-dysphoric patients and presumably cissexual heterosexual adolescent controlsa

94

875

Schwartz et al. (2010, pp. 101–103)

American and Canadian volunteers

677

873

VanderLaan et al. (2014, Table 2)

Canadian children and adolescents referred to a child and adolescent gender identity clinica

346

210

VanderLaan et al. (2016)

Brazilian homosexual male-to-female transsexuals and heterosexual cissexual men recruited at same hospital

118

143

VanderLaan and Vasey (2011, Table 1)

Samoan transgender homosexual males (fa’afafine) and cisgender heterosexual males

133

208

Vasey and VanderLaan (2007, Table 1)

Samoan transgender homosexual males (fa’afafine) and cisgender heterosexual males

83

114

Zucker and Blanchard (1994)

American psychoanalytic patients from earlier study by Bieber et al. (1962)

98

84

aThe sample included highly feminine boys who were too young to self-report their sexual orientation. These were classed as homosexual in this study because that is the most common adult outcome in the great majority of cases (Green, 1987; Singh, 2012; Steensma, 2013; Wallien & Cohen-Kettenis, 2008)

bThe rationale for this classification of male-to-female transsexuals is given in Blanchard (1989) and Lawrence (2010)

Measurement of Family Size

I used a special index of family size, Other Siblings, defined as the total number of siblings other than older brothers. In other words, Other Siblings = older sisters + younger brothers + younger sisters. This measure, first used in Blanchard et al. (2000), was specifically designed for comparing homosexual male study groups with heterosexual male control groups. The computing formula was originally developed for application to the sibling counts reported by individual subjects; in the present study, it was applied to the means for each group.

Results

Analyses of Unpaired Groups

The first part of the data analysis concerned the third goal of the study, namely the reliability of the FBOE itself. Of equal or greater importance, it also demonstrated the need to control for family size in FBOE research, and it laid the groundwork for the outcome variable used in the second part of the data analysis. This first set of analyses ignored the fact that the groups were originally collected in pairs and treated them as if they were completely independent entities. The groups’ summary variables were treated as single data points with no variance of their own. These analyses did not weight the individual data points in any way.

Each group’s mean number of Older Brothers is plotted against its mean number of Other Siblings in Fig. 1. The homosexual groups are indicated by black dots, and the heterosexual groups are indicated by white dots. I fitted separate regression lines through the homosexual and heterosexual groups using quadratic equations.5 I added the lines primarily as an aid to visual interpretation of the figure; these regressions have no role in the central analyses presented later.6
Fig. 1

Older brothers as a function of Other Siblings (older sisters + younger brothers + younger sisters) for the 30 homosexual and 30 heterosexual groups

Figure 1 illustrates two generalizable aspects of this data set. (1) For any narrow range of values along the Other Siblings dimension (e.g., 1.50–2.00 siblings), the homosexual groups tend to have more older brothers than the heterosexual groups. Thus, for example, in the air space above the 1.50–2.00 stretch of the X-axis, the black dots are mostly higher than the white dots. Stated generally, if a homosexual group and a heterosexual group have equal totals of older sisters, younger brothers, and younger sisters, then the homosexual group will probably have more older brothers. (2) If two groups differ substantially in Other Siblings, it can be misleading to compare them with regard to Older Brothers, at least when one is working in raw data. For example, the expected number of Older Brothers for a heterosexual group with 3 Other Siblings is about equal to the expected number of Older Brothers for a homosexual group with 2 Other Siblings and higher than the expected number for a homosexual group with 1 other sibling.

In order to adjust Older Brothers for family size, I divided this variable by Other Siblings, thus producing a new variable: the Older Brothers Ratio. Because the present research used published means from other investigators, I calculated the Older Brothers Ratio for each group by dividing its mean number of Older Brothers by its mean number of Other Siblings. Note that the identical value would result from dividing a group’s total number of Older Brothers by its total number of Other Siblings. The latter method of computing the Older Brothers Ratio points up the formal similarity between this statistic and the traditionally computed sibling sex ratio (e.g., Jensch, 1941a, 1941b; Lang, 1936, 1940, 1960), which is the ratio of brothers to sisters collectively reported by a given group of subjects.7

I replotted the new variable, Older Brothers Ratio, against Other Siblings in Fig. 2. The new variable had essentially no correlation with Other Siblings. I again plotted regression lines, using quadratic equations, simply to aid visual interpretation of the figure.8 The FBOE is again apparent in this figure from the generally greater height of the black dots relative to the white dots.
Fig. 2

Older Brothers Ratio (older brothers ÷ Other Siblings) as a function of Other Siblings

The overall classification accuracy obtainable with the Older Brothers Ratio was assessed with binary logistic regression. All 60 groups served as the cases. The criterion variable was Sexual Orientation, and the single predictor variable was the Older Brothers Ratio. This predictor classified 93% of groups (56 of 60) correctly. Obviously, this result was statistically significant, χ 2(1) = 49.73, p < .0001.9 Two homosexual groups were misclassified as heterosexual (Blanchard & Zucker, 1994; Currin, Gibson, & Hubach, 2015a, 2015b), and two heterosexual groups were misclassified as homosexual (Blanchard, Cantor, Bogaert, Breedlove, & Ellis, 2006, “Bogaert—Other” sample; Vasey & VanderLaan, 2007).

The location of the four misclassified groups is shown in Fig. 3. Most of the misclassified groups fell at the low end of the Other Siblings dimension; however, so did most of the correctly classified groups. A t test comparing the four misclassified and 56 correctly classified groups found no tendency for the misclassified groups to have lower values for Other Siblings, t(58) = .16, n.s.
Fig. 3

Misclassification of sexual orientation by Older Brothers Ratio as a function of Other Siblings

There was some additional useful information that could be extracted from the data depicted in Figs. 2 and 3. For the homosexual groups, the mean Older Brothers Ratio was .50 (SD = .21). When the one possible outlier (Khorashad et al., 2017, homosexual group) was excluded, the mean of the remaining 29 homosexual groups was .46 (SD = .08). For the heterosexual groups, the mean was .32 (SD = .06). These values could be helpful to future FBO researchers in evaluating the typicality of their samples’ results.

The variable, Older Brothers Ratio, can be viewed as the odds that any randomly selected sibling of a research subject will be an older brother, or more concisely, the odds of an older brother. For readers unfamiliar with the distinction between odds and probability, odds = probability ÷ (1 − probability). In this context, it is noteworthy that the heterosexual groups’ Older Brothers Ratio of .32 is close to the theoretical odds of .33 that would obtain if all four categories of siblings were equally probable at .25; that is, .33 = .25 ÷ (1 − .25). See also Blanchard (2014, Footnote 1).

Analyses of Paired Groups

As previously stated, the analyses up to this point ignored the information that the groups were originally collected in pairs. In contrast, the analyses in the second part of the investigation made full use of this information. These analyses addressed all three of the study’s main goals, namely the updated meta-analysis of the FBOE, the potential difference between feminine and non-feminine homosexual groups, and the relation between family size and FBOE magnitude, in that order.

The analyses used total numbers of older brothers, older sisters, and so on, rather than means, because totals (or counts) open the door to inferential statistics for individual studies. The mean numbers of siblings from published articles were converted to total numbers of siblings simply by multiplying each mean by the number of subjects on which it was computed and then rounding up or down to the nearest integer.

Updated Meta-analysis of the FBOE

On the view of the Older Brothers Ratio as a type of odds, dividing the Older Brothers Ratio of a homosexual group by the Older Brothers Ratio of its original heterosexual control group yields an Older Brothers Odds Ratio for that pair. The Older Brothers Odds Ratio is a ratio of ratios: (homosexuals’ older brothers ÷ homosexuals’ other siblings) ÷ (heterosexuals’ older brothers ÷ heterosexuals’ other siblings). Values >1.00 indicate that a homosexual group has a higher proportion of older brothers among its siblings than does its heterosexual control group. In other words, values >1.00 are positive evidence of the FBOE. The value of 1.00 represents the point of zero effect, and values <1.00 would represent direct evidence against the FBOE. It should be noted that this odds ratio is not the same as the usually calculated multiplicative change in the odds of homosexuality produced by each older brother (e.g., Blanchard & Bogaert, 1996b).

Figure 4 shows the Older Brothers Odds Ratios for the 30 pairs of groups plotted against the mean number of Other Siblings for the homosexual group in each pair. Figure 4 also shows a reference line at 1.00, representing no effect of sexual orientation. The first thing that one notices is that the Older Brothers Odds Ratio is >1.00 for 29 of the 30 groups. The binomial probability of this result under the null hypothesis of no sexual orientation effect is p < .000001. The one discrepant result was from the study by Currin et al. (2015a, 2015b); as previously indicated, this result concerned their homosexual group, not their heterosexual control group.
Fig. 4

Older Brothers Odds Ratio (homosexuals’ older brothers ratio ÷ heterosexuals’ older brothers ratio) as a function of Other Siblings for the homosexual group in each pair. The dashed reference line represents no effect of sexual orientation. The Older Brothers Odds Ratios of the feminine (n = 12) and non-feminine (n = 18) groups are identified by black and white dots, respectively

The Older Brothers Odds Ratio data were further investigated in two parallel meta-analyses. Both meta-analyses were performed with the Review Manager (RevMan) computer program (version 5.3. Copenhagen, Denmark: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Both used the inverse-variance weighting method; one used a random effects model, and the other used a fixed effects model. I will report the results obtained under the random effects model. This model was more conservative in the sense that it consistently produced higher (i.e., less significant) p values; however, there was no instance in which one model produced a p value below the traditional cutoff of .05 and the other did not. In subgroups analyses, the two models produced practically identical pooled estimates of the Older Brothers Odds Ratio when the subgroup in question showed no significant evidence of heterogeneity, and they produced quite similar estimates even when heterogeneity was detected.

Figure 5 shows a forest plot for the meta-analysis of all 30 samples along with inferential statistics for the pooled and individual samples. The 95% confidence intervals for individual samples extend the previously mentioned finding that 29 of the 30 groups had Older Brothers Odds Ratios >1.00. For 20 of these 29 groups, the difference was statistically significant, and for 9 it was not. In the 1 discrepant case that had an Older Brothers Odds Ratio <1.00, the difference was not statistically significant.
Fig. 5

Meta-analysis of Older Brothers Odds Ratio for all 30 samples. “Events” refers to older brothers, and “Total” refers to all siblings, including older brothers. The lozenge-shaped object at the bottom of the forest plot represents the pooled estimate of the Older Brothers Odds Ratio and its 95% confidence interval. See text for additional explanation. Symbols identify multiple samples from the same study (see Table 1): §Offenders against pubescents; Offenders against children; *Offenders against adults; §§Bogaert non-biological families subsample; ¶¶Bogaert “other” subsample; **Blanchard subsample

In the meta-analysis proper, the pooled Older Brothers Odds Ratio was 1.47, 95% CI [1.33, 1.62], which was significantly greater than the no-effect value of 1.00, z = 7.59, p < .00001. There was also, however, evidence of substantial heterogeneity among the samples, as reflected in the multiple non-overlapping confidence intervals shown in Fig. 5. The statistical test for heterogeneity was highly significant, χ 2(29) = 116.18, p < .00001, indicating that the variation in effect estimates was beyond chance. The quantitative estimate of inconsistency, I 2 (Deeks, Higgins, & Altman, 2011), indicated that 75% of the variability in effect estimates was due to differences among the individual samples rather than to sampling error.

Femininity in Homosexuals and Fraternal Birth Order

The heterogeneity in the full set of 30 samples was investigated with subgroup analysis. The subgroups chosen for examination corresponded to the variables already identified in the Introduction as potentially relevant (i.e., gender identity and family size); thus, the investigation of heterogeneity necessitated by the foregoing results was essentially the same as the original research plan. I began with the investigation of feminine and non-feminine samples.

As shown in Table 1, 12 of the homosexual groups were characterized by feminine behavior or transgender feelings,10 and 18 were not so characterized. The Older Brothers Odds Ratios of the feminine and non-feminine groups are identified by black and white dots in Fig. 4.

The pooled Older Brothers Odds Ratio for the feminine/transgender groups was 1.85, 95% CI [1.50, 2.29], which was significantly >1.00, z = 5.65, p < .00001. These groups still showed evidence of substantial heterogeneity, χ 2(11) = 55.53, p < .00001, I 2 = 80%. The pooled estimate for the non-feminine/cisgender groups was 1.27, 95% CI [1.20, 1.35], which was also significantly >1.00, z = 7.71, p < .00001. These groups, however, did not show evidence of heterogeneity, χ 2(17) = 20.01, p = .27, I 2 = 15%. The 95% confidence intervals for the pooled feminine and non-feminine groups did not overlap, and the test for subgroup differences (Deeks et al., 2011) confirmed that the feminine groups had significantly higher Older Brothers Odds Ratios (i.e., higher ratios of older brothers) than did the non-feminine groups, χ 2(1) = 10.97, p = .0009.

Family Size and Fraternal Birth Order

My strategy for investigating the relation between family size and the magnitude of the FBOE was to divide the feminine subgroup into two sub-subgroups with larger and smaller family sizes and to do the same with the non-feminine subgroup. The measure of family size was the mean number of Other Siblings for the homosexual group in each pair (I ignored the mean numbers of Other Siblings for the heterosexual control groups). Because this measure is the same variable defining the X-axis in Fig. 4, that figure serves to illustrate the relations (or lack thereof) examined in this analysis. The criterion for classifying a group as having a larger or smaller family size was a mean number of Other Siblings ≥2.00 or <2.00, respectively. There were 7 feminine/transgender groups with smaller family sizes and 5 with larger family sizes, and there were 13 non-feminine/cisgender groups with smaller family sizes and 5 with larger family sizes.

The pooled Older Brothers Odds Ratio for the feminine groups with smaller families was 1.96, 95% CI [1.39, 2.77], which was significantly >1.00, z = 3.81, p = .0001. These groups showed evidence of substantial heterogeneity, χ 2(6) = 37.02, p < .00001, I 2 = 84%. The pooled estimate for the feminine groups with larger families was 1.72, 95% CI [1.31, 2.25], which was also significantly >1.00, z = 3.94, p < .0001. These groups also showed evidence of heterogeneity, χ 2(4) = 17.01, p = .002, I 2 = 76%. The test for subgroup differences showed that the Older Brothers Odds Ratios of these subgroups did not differ, χ 2(1) = .34, p = .56. Thus, there was no evidence that the feminine groups with smaller families and those with larger families differed in regard to their Older Brothers Odds Ratios.

The findings for the non-feminine groups yielded similar conclusions regarding the effect of family size but not regarding heterogeneity among the samples. The pooled Older Brothers Odds Ratio for the non-feminine groups with smaller families was 1.28, 95% CI [1.18, 1.38], which was significantly >1.00, z = 6.09, p < .00001. These groups showed no evidence of heterogeneity, χ 2(12) = 16.55, p = .17, I 2 = 28%. The pooled estimate for the non-feminine groups with larger families was 1.24, 95% CI [1.12, 1.38], which was also significantly >1.00, z = 4.00, p < .0001. These groups also showed no evidence of heterogeneity, χ 2(4) = 3.31, p = .51, I 2 = 0%. The test for subgroup differences showed that the Older Brothers Odds Ratios of these subgroups did not differ, χ 2(1) = .15, p = .70. Therefore, as in the previous analysis, there was no evidence that the groups with smaller families and those with larger families differed in their estimated Older Brothers Odds Ratios.

Discussion

The results, no matter how they were analyzed, confirmed the reliability of the FBOE in the studies selected for investigation. The measure of fraternal birth order adjusted for family size, older brothers ÷ other siblings, classified over 93% of the 60 groups correctly. There was no apparent pattern in the misclassified groups. Four misclassifications out of 60 seems consistent with random error, especially when only about 15–29% of the men in any given homosexual group can attribute their sexual orientation to the FBOE (Blanchard & Bogaert, 2004; Cantor et al., 2002).

When considered pairwise, the homosexual group had a higher (adjusted) fraternal birth order than its original heterosexual control group 97% of the time, that is, in 29 of 30 pairs. For 20 of these 29 pairs, the difference was statistically significant. In the one discrepant case in which the homosexual group had a slightly lower fraternal birth order, the difference was not statistically significant.

Additional, more quantitative investigation of the data was carried out in a series of hierarchically organized meta-analyses. As in the other pairwise analyses, the outcome variable was a novel measure of the FBOE, the Older Brothers Odds Ratio, computed as (homosexuals’ older brothers ÷ homosexuals’ other siblings) ÷ (heterosexuals’ older brothers ÷ heterosexuals’ other siblings), where other siblings = older sisters + younger brothers + younger sisters. Values >1.00 on this new measure represent positive evidence for the fraternal birth order effect.

The meta-analysis found that, estimated from all 30 samples, the Older Brothers Odds Ratio was 1.47. This means that the ratio of older brothers to other siblings was 47% greater for homosexuals than it was for heterosexuals. This result can be re-cast in different terms, namely the proportion of older brothers (older brothers ÷ all siblings) rather than the odds of older brothers (older brothers ÷ other siblings). Recomputed in these terms, the result was that the proportion of older brothers was 31% greater in the sibships of homosexuals than in the sibships of heterosexuals. The latter value is the so-called risk ratio. (The statistic has this name for historical reasons; it can be computed for good things, bad things, or neutral things). In the context of this study, the risk ratio was always lower than the odds ratio—in this instance, 1.31 versus 1.47.

For most people, an effect measure calculated on a mix of feminine/transgender and non-feminine/cisgender homosexual males will be less useful or interesting than effects measures calculated separately for these groups. The pooled Older Brothers Odds Ratio for the feminine groups was 1.85, and the value for the non-feminine groups was 1.27. The corresponding risk ratios were 1.52 and 1.19. The differences between groups were highly significant. To sum up the results so far in common language: Feminine homosexual males have more older brothers than non-feminine homosexual males, and non-feminine homosexual males, in turn, have more older brothers than heterosexual males.

The difference between feminine and the non-feminine homosexual males was the first novel finding of this study. In contrast to previous studies (Bogaert, 2003, 2005a; Kishida & Rahman, 2015; Semenya et al., 2017), this study did find evidence that the FBOE was significantly stronger in feminine or transgender homosexual groups than in other homosexual groups. The discrepancy in results may be the result of a large methodological difference: Unlike previous studies, the present study contrasted entire groups of highly feminine homosexual males with entire groups of relatively typical homosexual males. The earlier studies examined either highly feminine homosexual males or typical homosexual males but not both, and within-group variance in femininity was assessed with questionnaire measures. In other words, the earlier studies may have suffered from a restriction-of-range problem.

This finding is important to theory for two related reasons. First, masculinity–femininity is one of the most salient behavioral variables within the male homosexual population (Bailey & Zucker, 1995), and second, the relation between FBO and masculinity–femininity represents one of the few demonstrated links between any established behavioral variable and a specific etiologic pathway.

The interpretation of the present finding is more uncertain. It may reflect a dosage effect of some maternal product that operates directly on the fetal brain during prenatal life. According to this hypothesis, a smaller number of older brothers simply increases the odds of sexual orientation toward males, whereas a larger number not only increases those odds but also leads to wider-ranging emotional and behavioral propensities typical of females.

A second possible hypothesis also assumes that the relation between number of older brothers and the odds of homosexuality concerns prenatal life, but it posits that any dosage effect pertains solely to sexual orientation itself. According to this hypothesis, it is the experience of growing up in a highly male-dominated sibship that somehow predisposes a homosexual boy to behave in a feminine manner or identify as a girl.

The distinction between the foregoing hypotheses can be illustrated with the example of gay men who were adopted at birth by biologically unrelated parents. These men might have biological older brothers (whom they have never met) through their birth mother or adoptive older brothers (with whom they were raised) or both. The first explanation predicts that the men’s femininity should vary as a function of their number of biological older brothers; the second explanation predicts that their femininity should vary as a function of their number of adoptive older brothers.

A third possibility is that the neurodevelopmental pathway triggered by older brothers is inherently more feminizing than pathways triggered by other etiologic factors (e.g., “gay” genes or prenatal hormone exposure). Thus, a group of homosexual males selected for generalized femininity is likely to contain a higher proportion of individuals who acquired their sexual orientation via the older brother pathway. Other hypotheses, equally speculative, are also possible. The present study could not address such questions.

The study’s second novel finding is that it does not matter, with regard to the magnitude of the FBOE, whether homosexual and heterosexual groups being compared on older brothers have large or small family sizes, so long as they have the same family size (or can be adjusted to simulate that condition). There are many research designs that might have been used to approach this question, and it is possible that some of them might have led to a different conclusion. It is also possible that the inclusion of several homosexual and heterosexual groups with very small or very large family sizes would have altered the results. I tend to doubt, however, that a strong relationship that would be apparent with another research design would be completely invisible in this study. The provisional conclusion that size does not matter (within limits) is relevant as fraternal birth order research approaches the next step in its evolution, namely movement into the laboratory.

Blanchard and Bogaert (1996b) proposed that the FBOE reflects the progressive immunization of some mothers to male-specific (i.e., Y-linked) antigens by each succeeding male fetus and the concomitantly increasing effects of anti-male antibodies on sexual differentiation of the brain in each succeeding male fetus. According to this maternal immune hypothesis, cells (or cell fragments) from male fetuses enter the maternal circulation during childbirth or perhaps earlier in pregnancy. These cells include substances that occur only on the surfaces of male cells, primarily male brain cells. The mother’s immune system recognizes these male-specific molecules as foreign and produces antibodies to them. When the mother later becomes pregnant with another male fetus, her antibodies cross the placental barrier and enter the fetal brain. Once in the brain, these antibodies bind to male-specific molecules on the surface of neurons. This prevents these neurons from “wiring-up” in the male-typical pattern, so that the individual will later be attracted to men rather than women.

At the time that Blanchard and Bogaert (1996b) advanced the maternal immune hypothesis as an explanation of the FBOE, relatively little was known about Y-linked antigens. As knowledge of Y-linked antigens accumulated in the following years, however, it became possible to identify candidate antigens for the maternal immune hypothesis according to certain obvious criteria (e.g., Blanchard & Klassen, 1997). These criteria included tissue localization (brain), timing of expression (prenatal), and cellular location (external cell membrane). When the relevant information became available, Blanchard et al. (2002) nominated PCDH11Y as one candidate gene/antigen, and Blanchard (2004) nominated NLGN4Y as another. It would not be possible at present to do a prospective study of these proteins starting with pregnant women, but it is already feasible to examine mothers of adult males for evidence of past immune reactions to them, using ELISA or ELISpot assays.

The present study was not focused on methodology per se; however, some of the graphical and data-analytic approaches developed in it can be applied more widely. The Older Brothers Ratio and Older Brothers Odds Ratio, in particular, are handy for comparing FBOE magnitude in different studies (including future studies) in a way that automatically takes family size into account. I will illustrate this with the particularly interesting example of Frisch and Hviid (2006). I did not include this study in my meta-analysis because there are notable problems with its raw data (Blanchard, 2007; Blanchard & VanderLaan, 2015), and because I was interested in generating quantitative estimates for the Older Brothers Ratio and the Older Brothers Odds Ratio and not simply in demonstrating greater-than/less-than relationships. This study is noteworthy, however, because it has been cited as an important failure to find the FBOE in a very large sample (Kishida & Rahman, 2015).

From Table 4 in Frisch and Hviid (2006), one can calculate the (approximate) total numbers of older brothers, older sisters, younger brothers, and younger sisters for the homosexual and heterosexual groups. From these sums, it is easy to calculate the Older Brothers Ratios and the Older Brothers Odds Ratio. The Older Brothers Ratio for Frisch and Hviid’s homosexual group was .35 (699 ÷ 1985) and the Older Brothers Ratio for their heterosexual group was .29 (147,704 ÷ 503,873). Thus, the Older Brothers Odds Ratio for this study was 1.20, that is (699 ÷ 1985) ÷ (147,704 ÷ 503,873). The 99% confidence interval for this value is 1.07–1.35.11 This interval includes the present study’s estimated mean Older Brothers Odds Ratio of 1.27 for non-feminine homosexual groups, but it does not include the no-effect value of 1.00. Thus, a study that was regarded both by the original investigators and by subsequent commentators as an important failure to detect the FBOE turns out to have results comparable to those of positive studies when family size is taken into account.

Limitations

The results of the meta-analysis showed that there was considerable heterogeneity among the feminine/transgender homosexual groups, that is, their observed Older Brothers Odds Ratios were more different from each other than one would expect from random error (chance) alone. In other words, there were substantial differences between effect estimates where similarity of effect estimates is the desirable outcome. When heterogeneity is present, the pooled effect estimate might not correspond to any single population. In this study, heterogeneity could be the result of differences in the subjects themselves, cultural or economic differences influencing their likelihood of attending a gender identity clinic, differences in the methods used to collect their sibship data, or any combination of these or other potential factors. There are at least two ways in which the feminine groups were certainly more variable than the non-feminine groups: They were more diverse geographically and culturally, and they included subjects examined in childhood. I did not attempt to investigate the unexplained heterogeneity using subgroups analysis, partly because I did not hypothesize geography/culture or age at examination as variables beforehand, and partly because it would be desirable to have a larger number of feminine/transgender samples (with appropriate heterosexual controls) before such an analysis is attempted. I have put all the data used in this study into the “Appendix,” so other researchers can easily revisit this question when more data become available.

Conclusion

Fraternal birth order is, by far, the most broadly established factor influencing sexual orientation in men. The subjects who contributed siblings to this meta-analysis came from Canada in the north to Brazil in the south, from Iran in the east to Samoa in the west. They were born over nearly a 150-year period, starting in 1861 (Blanchard & Bogaert, 1996a). Some were psychiatry or clinical psychology patients, some were convicted sex offenders, and some were research volunteers. Some were examined in childhood and some were examined in adulthood. The data came from studies expressly designed to study fraternal birth order and from archived studies in which sibship items were merely part of a broader survey. There are many remaining questions about the FBOE, especially the underlying mechanism it reflects and whether the same mechanism completely explains the difference between feminine/transgender and non-feminine/cisgender homosexual males. The simple fact of the FBOE’s existence, however, seems almost beyond doubt at this point.

Footnotes

  1. 1.

    The available evidence already indicates that the FBOE cannot be detected when family size is strongly affected by the various parental strategies (so-called stopping rules) of ceasing reproduction after one child, after one male child, or after a child of each sex (Blanchard & Lippa, 2007; Xu & Zheng, 2014, 2017; Zucker, Blanchard, Kim, Pae, & Lee, 2007). In these particular situations, which are not addressed in the present paper, neither homosexual nor heterosexual males have enough older brothers to make comparisons meaningful.

  2. 2.

    The groups included pedophiles with a sexual preference for boys and male-to-female transsexuals with a sexual preference for men. These groups can accurately be labeled as homosexual (according to biological sex), but they do not have a social identity as gay.

  3. 3.

    The studies that did not report all means were Francis (2008), Iemmola and Camperio Ciani (2009), Kangassalo, Pölkki, and Rantala (2011), and Robinson and Manning (2000).

  4. 4.

    I identified only two published studies that met the indispensable criterion 3 but failed some other criterion, and that are not mentioned in the Introduction or Discussion of this article. These were Bogaert (2005b, 2010), which did not satisfy criterion 4. I can mention, for the sake of completeness, four other studies that I did not use in my meta-analysis because they did not meet criterion 3 and sometimes also other criteria: Bearman and Brückner (2002), Bogaert (1998), McConaghy et al. (2006), and Zietsch et al. (2012).

  5. 5.

    One could also graph data this way in individual studies, plotting values for individual subjects rather than for groups. The scatterplot itself would not be as readable, because individual values could occupy only points defined by integers on the XY axes and would therefore frequently overlie one another. The regression lines might be useful, however, perhaps as diagnostics.

  6. 6.

    Full information about the regression lines, for interested readers, is as follows. Homosexual groups: R 2 = .56, F(2, 27) = 17.25, p < .0001, Constant = .32, b 1 = .18, b 2 = .07. Heterosexual groups: R 2 = .81, F(2, 27) = 55.89, p < .0001, Constant = .21, b 1 = .12, b 2 = .04.

  7. 7.

    The Older Brothers Ratio can be calculated for individual subjects, but this requires a slightly modified formula. See Blanchard (2014, Footnote 1), where this variable is labeled the Modified Ratio of Older Brothers.

  8. 8.

    Homosexual groups: R 2 = .02, F(2, 27) = .20, n.s., Constant = .67, b 1 = −.16, b 2 = .03. Heterosexual groups: R 2 = .10, F(2, 27) = 1.46, n.s., Constant = .47, b 1 = −.14, b 2 = .03.

  9. 9.

    Comparable results were obtained with an ROC analysis. The area under the curve (AUC) was .95, with a 95% confidence interval of .90–1.00. The value of the AUC may be interpreted as a 95% probability that a randomly chosen homosexual group will have a higher mean Older Brothers Ratio than will a randomly chosen heterosexual group.

  10. 10.

    Blanchard and Sheridan (1992), Blanchard, Zucker, Bradley, and Hume (1995), Blanchard, Zucker, Cohen-Kettenis, Gooren, and Bailey (1996), Bozkurt et al. (2015), Gómez-Gil et al. (2011), Green (2000), Khorashad et al. (2017), Schagen et al. (2012), VanderLaan, Blanchard, Wood, and Zucker (2014), VanderLaan et al. (2016), VanderLaan and Vasey (2011), Vasey and VanderLaan (2007).

  11. 11.

    There are at least two online interactive calculators for finding confidence intervals around odds ratios: https://select-statistics.co.uk/calculators/confidence-interval-calculator-odds-ratio/ and https://www.medcalc.org/calc/odds_ratio.php.

References

  1. Bailey, J. M., & Zucker, K. J. (1995). Childhood sex-typed behavior and sexual orientation: A conceptual analysis and quantitative review. Developmental Psychology, 31, 43–55.CrossRefGoogle Scholar
  2. Bearman, P. S., & Brückner, H. (2002). Opposite-sex twins and adolescent same-sex attraction. American Journal of Sociology, 107, 1179–1205.CrossRefGoogle Scholar
  3. Bell, A. P., Weinberg, M. S., & Hammersmith, S. K. (1981). Sexual preference: Its development in men and women. Bloomington, IN: Indiana University Press.Google Scholar
  4. Bieber, I., Dain, H. J., Dince, P. R., Drellich, M. G., Grand, H. G., Gundlach, R. H., … Bieber, T. B. (1962). Homosexuality: A psychoanalytic study of male homosexuals. New York, NY: Basic Books.CrossRefGoogle Scholar
  5. Blanchard, R. (1989). The classification and labeling of nonhomosexual gender dysphorias. Archives of Sexual Behavior, 18, 315–334.CrossRefPubMedGoogle Scholar
  6. Blanchard, R. (1997). Birth order and sibling sex ratio in homosexual versus heterosexual males and females. Annual Review of Sex Research, 8, 27–67.PubMedGoogle Scholar
  7. Blanchard, R. (2004). Quantitative and theoretical analyses of the relation between older brothers and homosexuality in men. Journal of Theoretical Biology, 230, 173–187.CrossRefPubMedGoogle Scholar
  8. Blanchard, R. (2007). Older-sibling and younger-sibling sex ratios in Frisch and Hviid’s (2006) national cohort study of two million Danes. Archives of Sexual Behavior, 36, 860–863.CrossRefPubMedGoogle Scholar
  9. Blanchard, R. (2012). Fertility in the mothers of firstborn homosexual and heterosexual men. Archives of Sexual Behavior, 41, 551–556.CrossRefPubMedGoogle Scholar
  10. Blanchard, R. (2014). Detecting and correcting for family size differences in the study of sexual orientation and fraternal birth order. Archives of Sexual Behavior, 43, 845–852.CrossRefPubMedGoogle Scholar
  11. Blanchard, R., Barbaree, H. E., Bogaert, A. F., Dickey, R., Klassen, P., Kuban, M. E., & Zucker, K. J. (2000). Fraternal birth order and sexual orientation in pedophiles. Archives of Sexual Behavior, 29, 463–478.CrossRefPubMedGoogle Scholar
  12. Blanchard, R., & Bogaert, A. F. (1996a). Biodemographic comparisons of homosexual and heterosexual men in the Kinsey interview data. Archives of Sexual Behavior, 25, 551–579.CrossRefPubMedGoogle Scholar
  13. Blanchard, R., & Bogaert, A. F. (1996b). Homosexuality in men and number of older brothers. American Journal of Psychiatry, 153, 27–31.CrossRefPubMedGoogle Scholar
  14. Blanchard, R., & Bogaert, A. F. (1998). Birth order in homosexual versus heterosexual sex offenders against children, pubescents, and adults. Archives of Sexual Behavior, 27, 595–603.CrossRefPubMedGoogle Scholar
  15. Blanchard, R., & Bogaert, A. F. (2004). Proportion of homosexual men who owe their sexual orientation to fraternal birth order: An estimate based on two national probability samples. American Journal of Human Biology, 16, 151–157.CrossRefPubMedGoogle Scholar
  16. Blanchard, R., Cantor, J. M., Bogaert, A. F., Breedlove, S. M., & Ellis, L. (2006). Interaction of fraternal birth order and handedness in the development of male homosexuality. Hormones and Behavior, 49, 405–414.CrossRefPubMedGoogle Scholar
  17. Blanchard, R., & Ellis, L. (2001). Birth weight, sexual orientation and the sex of preceding siblings. Journal of Biosocial Science, 33, 451–467.CrossRefPubMedGoogle Scholar
  18. Blanchard, R., & Klassen, P. (1997). H-Y antigen and homosexuality in men. Journal of Theoretical Biology, 185, 373–378.CrossRefPubMedGoogle Scholar
  19. Blanchard, R., & Lippa, R. A. (2007). Birth order, sibling sex ratio, handedness, and sexual orientation of male and female participants in a BBC Internet research project. Archives of Sexual Behavior, 36, 163–176.CrossRefPubMedGoogle Scholar
  20. Blanchard, R., & Sheridan, P. M. (1992). Sibship size, sibling sex ratio, birth order, and parental age in homosexual and nonhomosexual gender dysphorics. Journal of Nervous and Mental Disease, 180, 40–47.CrossRefPubMedGoogle Scholar
  21. Blanchard, R., & VanderLaan, D. P. (2015). Commentary on Kishida and Rahman (2015), including a meta-analysis of relevant studies on fraternal birth order and sexual orientation in men. Archives of Sexual Behavior, 44, 1503–1509.CrossRefPubMedGoogle Scholar
  22. Blanchard, R., & Zucker, K. J. (1994). Reanalysis of Bell, Weinberg, and Hammersmith’s data on birth order, sibling sex ratio, and parental age in homosexual men. American Journal of Psychiatry, 151, 1375–1376.CrossRefPubMedGoogle Scholar
  23. Blanchard, R., Zucker, K. J., Bradley, S. J., & Hume, C. S. (1995). Birth order and sibling sex ratio in homosexual male adolescents and probably prehomosexual feminine boys. Developmental Psychology, 31, 22–30.CrossRefGoogle Scholar
  24. Blanchard, R., Zucker, K. J., Cavacas, A., Allin, S., Bradley, S. J., & Schachter, D. C. (2002). Fraternal birth order and birth weight in probably prehomosexual feminine boys. Hormones and Behavior, 41, 321–327.CrossRefPubMedGoogle Scholar
  25. Blanchard, R., Zucker, K. J., Cohen-Kettenis, P. T., Gooren, L. J. G., & Bailey, J. M. (1996). Birth order and sibling sex ratio in two samples of Dutch gender-dysphoric homosexual males. Archives of Sexual Behavior, 25, 495–514.CrossRefPubMedGoogle Scholar
  26. Blanchard, R., Zucker, K. J., Siegelman, M., Dickey, R., & Klassen, P. (1998). The relation of birth order to sexual orientation in men and women. Journal of Biosocial Science, 30, 511–519.CrossRefPubMedGoogle Scholar
  27. Bogaert, A. F. (1998). Birth order and sibling sex ratio in homosexual and heterosexual non-white men. Archives of Sexual Behavior, 27, 467–473.CrossRefPubMedGoogle Scholar
  28. Bogaert, A. F. (2003). Interaction of older brothers and sex-typing in the prediction of sexual orientation in men. Archives of Sexual Behavior, 32, 129–134.CrossRefPubMedGoogle Scholar
  29. Bogaert, A. F. (2004). The prevalence of male homosexuality: The effect of fraternal birth order and variations in family size. Journal of Theoretical Biology, 230, 33–37.CrossRefPubMedGoogle Scholar
  30. Bogaert, A. F. (2005a). Gender role/identity and sibling sex ratio in homosexual men. Journal of Sex and Marital Therapy, 31, 217–227.CrossRefPubMedGoogle Scholar
  31. Bogaert, A. F. (2005b). Sibling sex ratio and sexual orientation in men and women: New tests in two national probability samples. Archives of Sexual Behavior, 34, 111–116.CrossRefPubMedGoogle Scholar
  32. Bogaert, A. F. (2006). Biological versus nonbiological older brothers and men’s sexual orientation. Proceedings of the National Academy of Sciences of the USA, 103, 10771–10774.CrossRefPubMedPubMedCentralGoogle Scholar
  33. Bogaert, A. F. (2010). Physical development and sexual orientation in men and women: An analysis of NATSAL-2000. Archives of Sexual Behavior, 39, 110–116.CrossRefPubMedGoogle Scholar
  34. Bogaert, A. F., Bezeau, S., Kuban, M., & Blanchard, R. (1997). Pedophilia, sexual orientation, and birth order. Journal of Abnormal Psychology, 106, 331–335.CrossRefPubMedGoogle Scholar
  35. Bozkurt, A., Bozkurt, O. H., & Sonmez, I. (2015). Birth order and sibling sex ratio in a population with high fertility: Are Turkish male to female transsexuals different? Archives of Sexual Behavior, 44, 1331–1337.CrossRefPubMedGoogle Scholar
  36. Camperio-Ciani, A., Corna, F., & Capiluppi, C. (2004). Evidence for maternally inherited factors favoring male homosexuality and promoting female fecundity. Proceedings of the Royal Society of London, Series B: Biological Sciences, 271, 2217–2221.CrossRefGoogle Scholar
  37. Cantor, J. M., Blanchard, R., Paterson, A. D., & Bogaert, A. F. (2002). How many gay men owe their sexual orientation to fraternal birth order? Archives of Sexual Behavior, 31, 63–71.CrossRefPubMedGoogle Scholar
  38. Currin, J. M., Gibson, L., & Hubach, R. D. (2015a). Multidimensional assessment of sexual orientation and the fraternal birth order effect. Psychology of Sexual Orientation and Gender Diversity, 2, 113–122.CrossRefGoogle Scholar
  39. Currin, J. M., Gibson, L., & Hubach, R. D. (2015b). “Multidimensional assessment of sexual orientation and the fraternal birth order effect”: Correction to Currin, Gibson, and Hubach (2015). Psychology of Sexual Orientation and Gender Diversity, 2, 265.CrossRefGoogle Scholar
  40. Deeks, J. J., Higgins, J. P. T., & Altman, D. G. (Eds.). (2011). Chapter 9: Analysing data and undertaking meta-analyses. In J. P. T. Higgins & S. Green (Eds.), Cochrane handbook for systematic reviews of interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration, 2011. http://handbook.cochrane.org/.
  41. Ellis, L., & Blanchard, R. (2001). Birth order, sibling sex ratio, and maternal miscarriages in homosexual and heterosexual men and women. Personality and Individual Differences, 30, 543–552.CrossRefGoogle Scholar
  42. Francis, A. M. (2008). Family and sexual orientation: The family-demographic correlates of homosexuality in men and women. Journal of Sex Research, 45, 371–377.CrossRefPubMedGoogle Scholar
  43. Frisch, M., & Hviid, A. (2006). Childhood family correlates of heterosexual and homosexual marriages: A national cohort study of two million Danes. Archives of Sexual Behavior, 35, 533–547.CrossRefPubMedGoogle Scholar
  44. Gebhard, P. H., Gagnon, J. H., Pomeroy, W. B., & Christenson, C. V. (1965). Sex offenders: An analysis of types. New York, NY: Harper & Row.Google Scholar
  45. Gebhard, P. H., & Johnson, A. B. (1979). The Kinsey data: Marginal tabulations of the 1938–1963 interviews conducted by the Institute for Sex Research. Philadelphia, PA: W. B. Saunders.Google Scholar
  46. Gómez-Gil, E., Esteva, I., Carrasco, R., Cruz Almaraz, M., Pasaro, E., Salamero, M., & Guillamon, A. (2011). Birth order and ratio of brothers to sisters in Spanish transsexuals. Archives of Sexual Behavior, 40, 505–510.CrossRefPubMedGoogle Scholar
  47. Green, R. (1987). The “sissy boy syndrome” and the development of homosexuality. New Haven, CT: Yale University Press.CrossRefGoogle Scholar
  48. Green, R. (2000). Birth order and ratio of brothers to sisters in transsexuals. Psychological Medicine, 30, 789–795.CrossRefPubMedGoogle Scholar
  49. Iemmola, F., & Camperio Ciani, A. (2009). New evidence of genetic factors influencing sexual orientation in men: Female fecundity increase in the maternal line. Archives of Sexual Behavior, 38, 393–399.CrossRefPubMedGoogle Scholar
  50. Jensch, K. (1941a). Zur Genealogie der Homosexualität [Regarding the genetics of homosexuality]. Archiv für Psychiatrie und Nervenkrankheiten, 112, 527–540.Google Scholar
  51. Jensch, K. (1941b). Weiterer Beitrag zur Genealogie der Homosexualität [A further contribution regarding the genetics of homosexuality]. Archiv für Psychiatrie und Nervenkrankheiten, 112, 679–696.Google Scholar
  52. Jones, M. B., & Blanchard, R. (1998). Birth order and male homosexuality: Extension of Slater’s index. Human Biology, 70, 775–787.PubMedGoogle Scholar
  53. Kangassalo, K., Pölkki, M., & Rantala, M. J. (2011). Prenatal influences on sexual orientation: Digit ratio (2D:4D) and number of older siblings. Evolutionary Psychology, 9, 496–508.CrossRefPubMedGoogle Scholar
  54. Khorashad, B. S., Zucker, K. J., Talaei, A., Farid, F., Fayyazi Bordbar, M. R., Rajabzadeh, F., … Blanchard, R. (2017). Birth order and sibling sex ratio in a sample of Iranian transsexuals. Manuscript in preparation.Google Scholar
  55. King, M., Green, J., Osborn, D. P. J., Arkell, J., Hetherton, J., & Pereira, E. (2005). Family size in white gay and heterosexual men. Archives of Sexual Behavior, 34, 117–122.CrossRefPubMedGoogle Scholar
  56. Kishida, M., & Rahman, Q. (2015). Fraternal birth order and extreme right-handedness as predictors of sexual orientation and gender nonconformity in men. Archives of Sexual Behavior, 44, 1493–1501.CrossRefPubMedGoogle Scholar
  57. Lang, T. (1936). Beitrag zur Frage nach der genetischen Bedingtheit der Homosexualität [A contribution regarding the question of the genetic determination of homosexuality]. Zeitschrift für die gesamte Neurologie und Psychiatrie, 155, 702–713.Google Scholar
  58. Lang, T. (1940). Studies on the genetic determination of homosexuality. Journal of Nervous and Mental Disease, 92, 55–64.CrossRefGoogle Scholar
  59. Lang, T. (1960). Die Homosexualität als genetisches Problem [Homosexuality as a genetic problem]. Acta Geneticae Medicae et Gemellologiae, 9, 370–381.Google Scholar
  60. Lawrence, A. A. (2010). Sexual orientation versus age of onset as bases for typologies (subtypes) for gender identity disorder in adolescents and adults. Archives of Sexual Behavior, 39, 514–545.CrossRefPubMedGoogle Scholar
  61. McConaghy, N., Hadzi-Pavlovic, D., Stevens, C., Manicavasagar, V., Buhrich, N., & Vollmer-Conna, U. (2006). Fraternal birth order and ratio of heterosexual/homosexual feelings in women and men. Journal of Homosexuality, 51, 161–174.CrossRefPubMedGoogle Scholar
  62. Robinson, S. J., & Manning, J. T. (2000). The ratio of 2nd to 4th digit length and male homosexuality. Evolution and Human Behavior, 21, 333–345.CrossRefPubMedGoogle Scholar
  63. Schagen, S. E. E., Delemarre-van de Waal, H. A., Blanchard, R., & Cohen-Kettenis, P. T. (2012). Sibling sex ratio and birth order in early-onset gender dysphoric adolescents. Archives of Sexual Behavior, 41, 541–549.CrossRefPubMedGoogle Scholar
  64. Schwartz, G., Kim, R. M., Kolundzija, A. B., Rieger, G., & Sanders, A. R. (2010). Biodemographic and physical correlates of sexual orientation in men. Archives of Sexual Behavior, 39, 93–109.CrossRefPubMedGoogle Scholar
  65. Semenya, S. W., VanderLaan, D. P., & Vasey, P. L. (2017). Birth order and recalled childhood gender nonconformity in Samoan men and fa’afafine. Developmental Psychobiology, 59, 338–347.CrossRefGoogle Scholar
  66. Siegelman, M. (1972). Adjustment of male homosexuals and heterosexuals. Archives of Sexual Behavior, 2, 9–25.CrossRefPubMedGoogle Scholar
  67. Siegelman, M. (1973). Birth order and family size of homosexual men and women. Journal of Consulting and Clinical Psychology, 41, 164.CrossRefPubMedGoogle Scholar
  68. Siegelman, M. (1974). Parental background of male homosexuals and heterosexuals. Archives of Sexual Behavior, 3, 3–18.CrossRefPubMedGoogle Scholar
  69. Siegelman, M. (1978). Psychological adjustment of homosexual and heterosexual men: A cross-national replication. Archives of Sexual Behavior, 7, 1–11.CrossRefPubMedGoogle Scholar
  70. Siegelman, M. (1981). Parental backgrounds of homosexual and heterosexual men: A cross national replication. Archives of Sexual Behavior, 10, 505–513.CrossRefPubMedGoogle Scholar
  71. Singh, D. (2012). A follow-up study of boys with gender identity disorder (Unpublished doctoral dissertation). University of Toronto, Toronto, Canada.Google Scholar
  72. Steensma, T. D. (2013). From gender variance to gender dysphoria: Psychosexual development of gender atypical children and adolescents (Unpublished doctoral dissertation). Vrije Universiteit, Amsterdam, the Netherlands.Google Scholar
  73. Suarez, B. K., & Przybeck, T. R. (1980). Sibling sex ratio and male homosexuality. Archives of Sexual Behavior, 9, 1–12.CrossRefPubMedGoogle Scholar
  74. Swift-Gallant, A., Coome, L. A., Monks, D., & VanderLaan, D. P. (2017). Gender nonconformity and birth order in relation to anal sex role among gay men. Archives of Sexual Behavior. doi: 10.1007/s10508-017-0980-y.Google Scholar
  75. VanderLaan, D. P., Blanchard, R., Wood, H., Garzon, L. C., & Zucker, K. J. (2015). Birth weight and two possible types of maternal effects on male sexual orientation: A clinical study of children and adolescents referred to a gender identity service. Developmental Psychobiology, 57, 25–34.CrossRefPubMedGoogle Scholar
  76. VanderLaan, D. P., Blanchard, R., Wood, H., & Zucker, K. J. (2014). Birth order and sibling sex ratio of children and adolescents referred to a gender identity service. PLoS ONE. doi: 10.1371/journal.pone.0090257.PubMedPubMedCentralGoogle Scholar
  77. VanderLaan, D. P., Blanchard, R., Zucker, K. J., Massuda, R., Fontanari, A. M. V., Borba, A. O., … Lobato, M. I. (2016). Birth order and androphilic male-to-female transsexualism in Brazil. Journal of Biosocial Science. doi: 10.1017/S0021932016000584.PubMedGoogle Scholar
  78. VanderLaan, D. P., & Vasey, P. L. (2011). Male sexual orientation in Independent Samoa: Evidence for fraternal birth order and maternal fecundity effects. Archives of Sexual Behavior, 40, 495–503.CrossRefPubMedGoogle Scholar
  79. Vasey, P. L., & VanderLaan, D. P. (2007). Birth order and male androphilia in Samoan fa’afafine. Proceedings of the Royal Society of London, Series B: Biological Sciences, 274, 1437–1442.CrossRefGoogle Scholar
  80. Wallien, M. S., & Cohen-Kettenis, P. T. (2008). Psychosexual outcome of gender dysphoric children. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 1413–1423.CrossRefPubMedGoogle Scholar
  81. Wampold, C. H. (2013). Redesigning fraternal birth order studies from top to bottom [Letter to the Editor]. Archives of Sexual Behavior, 42, 1387–1389.CrossRefPubMedGoogle Scholar
  82. Xu, Y., & Zheng, Y. (2014). Birth order and sibling sex ratio in relation to sexual orientation in China. Social Behavior and Personality, 42, 995–1002.CrossRefGoogle Scholar
  83. Xu, Y., & Zheng, Y. (2017). Fraternal birth order, handedness, and sexual orientation in a Chinese population. Journal of Sex Research, 54, 10–18.CrossRefPubMedGoogle Scholar
  84. Zheng, L., Hart, T. A., & Zheng, Y. (2012). The relationship between intercourse preference positions and personality traits among gay men in China. Archives of Sexual Behavior, 41, 683–689.CrossRefPubMedGoogle Scholar
  85. Zietsch, B. P., Verweij, K. J. H., Heath, A. C., Madden, P. A. F., Martin, N. G., Nelson, E. C., & Lynskey, M. T. (2012). Do shared etiological factors contribute to the relationship between sexual orientation and depression? Psychological Medicine, 42, 521–532.CrossRefPubMedGoogle Scholar
  86. Zucker, K. J., & Blanchard, R. (1994). Reanalysis of Bieber et al’.s 1962 data on sibling sex ratio and birth order in male homosexuals. Journal of Nervous and Mental Disease, 182, 528–530.CrossRefPubMedGoogle Scholar
  87. Zucker, K. J., Blanchard, R., Kim, T.-S., Pae, C.-U., & Lee, C. (2007). Birth order and sibling sex ratio in homosexual transsexual South Korean males: Effects of the male-preference stopping rule. Psychiatry and Clinical Neurosciences, 61, 529–533.CrossRefPubMedGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2017

Authors and Affiliations

  1. 1.Department of PsychiatryUniversity of TorontoTorontoCanada

Personalised recommendations