Personality differences between birth order categories and across sibship sizes

Significance

An enduring question in psychology is whether adult personality is related to birth order and number of siblings. Previous studies have generally found no associations except that firstborns average slightly higher than later-borns in intellect-related traits. Our large-sample study supports the findings for intellect-related traits, but unlike previous research, it also shows considerable differences for other traits: On the Honesty-Humility and Agreeableness dimensions, scores were somewhat higher on average for persons having more siblings. Among persons having the same number of siblings, middle-borns and last-borns both averaged slightly higher than firstborns. We discuss the possibility that growing up with more siblings promotes the development of a more cooperative personality.

Abstract

We examined associations of self-reports on the HEXACO Personality Inventory—Revised (HEXACO-PI-R) with birth order category and sibship size, controlling for participant sex and age. In a first sample (N > 700,000 online adults, mainly from English-speaking countries), Honesty-Humility and Agreeableness both showed the highest means for middle-borns, followed in order by last-borns (youngests), firstborns (oldests), and only children, with differences between middles and onlys of d ≥ 0.20. The same result was replicated in a similar but smaller second sample (N > 70,000) in which sibship size was also assessed, thereby allowing birth order differences to be separated from sibship size differences. In that sample, Honesty-Humility and Agreeableness showed higher means with larger sibship sizes, with differences between sibship sizes of 1 and 6+ of d = 0.30 and d = 0.36, respectively. Controls for upbringing religiousness and current religiousness reduced these differences by about 25%. Within sibship sizes, birth order differences in these dimensions were considerably smaller but oldests remained up to d = 0.10 lower than middles and youngests. Openness was d ≈ 0.10 higher for onlys than for non-onlys collectively, and within sibship sizes, Openness was d ≈ 0.10 higher for oldests than for middles and youngests.

Do siblings matter in personality development? More specifically, do adults’ personality trait levels depend on how many siblings they had and on whether those siblings were older or younger? Here we investigate mean levels of personality traits in adults as a function of their birth order and their sibship size.

Birth Order

The idea that birth order influences personality has a long history (e.g., 1, 2), but recent research findings suggest that persons of different birth order positions have nearly identical mean levels of personality traits (3, 4).

Rohrer, Egloff, and Schmukle examined self-reports on the Big Five personality factors from a total of about 20,000 participants in national panel studies in Germany, the United Kingdom, and the United States of America (3, 5). Those authors used information from the participants about their sibship size and birth order, both of which were defined in terms of the household rather than biologically. They compared personality trait levels for persons of various birth orders (1st, 2nd, 3rd, 4th) within sibship sizes of 2, 3, and 4. Their results showed no differences across birth ranks within sibship sizes except that the “Intellect” element of Big Five Openness to Experience showed slightly lower levels for persons having later birth order positions.

Damian and Roberts examined personality self-reports in about 272,000 US high school students who participated in the Project Talent sample in the year 1960 (4). The personality measures were developed on an ad hoc basis for Project Talent. Damian and Roberts found that birth order (operationalized as firstborn versus nonfirstborn) showed small positive correlations with the various personality scale scores, none exceeding 0.06. Among participants having a sibship size of three, firstborns tended to be higher than lastborns not only on an Openness-related trait [as later found by Rohrer et al. (3)] but also on several other socially desirable traits, with ds of 0.10 or higher for six of the 10 scales of the Project Talent Personality Inventory.

Despite the results of these large-sample studies, it could be worthwhile to revisit the possibility of links of personality trait levels with birth order and with sibship size. One reason to reexamine this issue involves limitations of the personality measures used in those investigations.

In the national panel samples examined by Rohrer et al. (3), the Big Five personality measures administered to members of those samples were generally short or with relatively little breadth of content. The Big Five scales consisted of two items in the US sample, three or four items in the German sample, and 10 items in the UK sample. Even in the UK sample, some of the scales were relatively homogeneous; for example, the Agreeableness scale consists heavily of items describing concern for others versus lack thereof.

In the dataset examined by Damian and Roberts (4), the Project Talent Personality Inventory consisted of 10 self-report scales. However, nine of those scales (all except Impulsiveness) were substantially positively intercorrelated within the Project Talent sample, with correlations ranging from 0.22 to 0.61 and averaging 0.44. These nine scales were later found to correlate negatively with Neuroticism and positively with the other four Big Five dimensions (6). Thus, the Project Talent scales appear to sample personality traits from the same region of the personality factor space, one that contrasts highly socially desirable with undesirable traits.

Given these features of the personality measures in these studies (3, 4), it seems at least possible that a broader sampling of personality traits, using factor-level scales having weaker intercorrelations and each containing more items spanning a wider array of narrow traits, would allow the detection of some birth order effects not previously observed.

Only Children and Children with Siblings—and Sibship Size More Generally

Previous research on birth order and personality has typically not included comparisons between persons raised in one-child families (i.e., “only children”) and persons raised with siblings. This exclusion is based on the reasonable grounds that only children “cannot be compared with any siblings and are thus uninformative for the analysis of birth-order effects” (3, p. 14228). Nevertheless, it is possible to compare personality trait levels between persons from one-child families and persons raised with siblings, keeping in mind the possibility that only-child/non-only-child status may involve causes different from those associated with number of siblings.

Research comparing the personalities of only children and children with siblings has sometimes focused on traits relevant to selfishness (or lack thereof), given the stereotype of greater selfishness in only children than in children with siblings (e.g., 7). Some studies have found essentially no difference in narcissism between adults who grew up as only children and adults who grew up as children with siblings (e.g., 8, 9), and another study found no difference in Big Five Agreeableness (10). Other studies have found only children to be more selfish than children with siblings, with effect sizes ranging from small (11, d ≈ 0.10 for low Honesty-Humility) to medium (12, d ≈ 0.30 for valuing “power” over “benevolence”).

By definition, comparisons of only children with non-only children are comparisons of persons in a sibship of 1 with persons in a sibship of 2 or more. However, one can also compare persons of varying sibship sizes, from 1 through 2, 3, and beyond. In (4), sibship size showed weak negative correlations with all 10 Project Talent Personality Inventory scales, indicating that persons from larger sibships had modestly lower self-report scores across a variety of socially desirable trait scales.

Present Study

In the present research, we reexamine the possibility that adult personality trait levels may differ between birth order categories and across sibship sizes. We use longer and more differentiated personality measures, within larger and more geographically diverse samples than those of previous investigations.

The participants of our samples completed a self-report personality inventory to obtain feedback about their levels of the HEXACO personality dimensions (13, 14): Honesty-Humility (H), Emotionality (E), Extraversion (X), Agreeableness [versus Anger] (A), Conscientiousness (C), and Openness to Experience (O). The HEXACO Extraversion, Conscientiousness, and Openness dimensions are similar to the Big Five dimensions of the same names, but the relations of other three HEXACO dimensions with Big Five Agreeableness and Neuroticism are more complex; note that HEXACO Agreeableness [versus Anger] is not the same as Big Five Agreeableness.

Participants of the first sample (N = 710,797, hereafter Sample 1) indicated their birth order category (first, last, middle, only), and participants of the second sample (N = 74,920, hereafter Sample 2) indicated their sibship size (1, 2, 3, 4, 5, 6+) in addition to their birth order category (SI Appendix, Table S1 for the descriptive statistics and alpha reliabilities for the study variables). Both questions asked participants about the household in which they grew up and thus assessed birth order and sibship size in the social rather than the biological sense. Following previous research, we refer to children of a household as siblings, without distinguishing full or half-siblings, without distinguishing biological, adopted, step-, and foster siblings, and without distinguishing siblings from other child relatives such as cousins.

Results

Birth Order and Personality.

Table 1 shows the standardized mean scores on the six HEXACO scales for the birth order categories (oldests, youngests, middles, and onlys) in Sample 1 and Sample 2, after controlling for respondent sex and age (the latter as a linear variable). (Results for the 25 facet scales are shown in SI Appendix, Tables S2–S5; note that reliabilities of facet-level scales are lower than those of factor-level scales.) Also given in Table 1 is a column showing the largest difference for any pair of birth order categories within each sample. The differences between these z-scores correspond closely to d statistics, so we refer to those differences in terms of d. The patterns of differences were strikingly similar across two samples, so we describe them together.

Table 1.

Mean standardized scores on HEXACO factor scales by birth order category (participant age and sex controlled)

	Only	Oldest	Youngest	Middle	max |d| (Highest vs lowest)
Sample 1
Honesty-humility	−0.119	−0.039	0.048	0.086	0.205 (Middle vs only)
Emotionality	0.021	0.004	0.011	−0.043	0.064 (Only vs middle)
Extraversion	−0.070	0.021	−0.016	0.032	0.102 (Middle vs only)
Agreeableness	−0.133	−0.034	0.035	0.108	0.241 (Middle vs only)
Conscientiousness	−0.033	0.011	−0.015	0.026	0.059 (Middle vs only)
Openness to experience	0.082	0.035	−0.051	−0.046	0.133 (Only vs Youngest)
95% CI = ±	0.007	0.004	0.004	0.006
Sample 2
Honesty-humility	−0.142	−0.046	0.042	0.101	0.243 (Middle vs only)
Emotionality	0.008	0.001	0.018	−0.037	0.056 (Youngest vs Middle)
Extraversion	−0.093	0.022	−0.019	0.037	0.130 (Middle vs only)
Agreeableness	−0.158	−0.024	0.014	0.112	0.270 (Middle vs only)
Conscientiousness	−0.062	0.012	−0.020	0.044	0.105 (Middle vs only)
Openness to experience	0.105	0.042	−0.052	−0.056	0.161 (Only vs middle)
95% CI = ±	0.022	0.012	0.013	0.016

For Sample 1, Ns = 90,922 (Only), 273,421 (Oldest), 218,078 (Youngest), and 128,376 (Middle). For Sample 2, Ns = 7,945 (Only), 28,846 (Oldest), 23,786 (Youngest), and 14,343 (Middle).

Honesty-Humility and Agreeableness showed the largest differences across birth order categories. For both of these factor scales, the rank ordering of birth order categories from the highest to lowest average scores was middle, youngest, oldest, only. In Samples 1 and 2, respectively, the differences between middles and onlys were d = 0.21 and 0.24 for Honesty-Humility and d = 0.24 and 0.27 for Agreeableness.

Differences across birth order categories were smaller for other dimensions, but some were noteworthy. For Openness to Experience, onlys and oldests averaged higher than middles and youngests (ds > 0.10 for both samples); that is, scores were slightly lower for persons having older siblings. For Extraversion, middles and oldests averaged higher than youngests and onlys, with ds ≥ 0.10 for middles versus onlys; that is, scores were slightly higher for persons having younger siblings.

Sibship Size and Personality.

The largest personality differences between birth order categories usually involved middles and onlys. This result suggests that the birth order differences could reflect sibship size differences: onlys by definition are from families with one child, middles by definition are from families with three or more children, and both oldests and youngests are from families with two or more children but in many cases exactly two children. We added a sibship size variable in Sample 2 (Materials and Methods) so that we could examine this issue, after having noticed the birth order differences of Sample 1.

Fig. 1 shows the standardized mean scores on the six HEXACO scales in Sample 2 by sibship size after controlling for respondent sex and age. The largest differences were observed for Honesty-Humility and Agreeableness, both of which showed increasing means as a function of sibship size. Mean scores were d = 0.30 higher (Honesty-Humility) or d = 0.36 higher (Agreeableness) for sibship size of 6+ (N = 3,226) than for sibship size of 1 (N = 7,945). The Extraversion and Conscientiousness factors also showed increases across sibship sizes, with d > 0.15. For the Openness to Experience factor, sibship size of 1 averaged d ≈ 0.10 higher than the other sibship sizes, whose means were nearly equal.

Fig. 1.

Mean z-scores on HEXACO factor scales by sibship size. Error bars show 95% CI. Ns were 7,945 (sibship size 1), 30,216 (sibship size 2), 20,598 (sibship size 3), 8,785 (sibship size 4), 3,605 (sibship size 5), and 3,226 (sibship size 6+).

We note that religiousness is related to number of offspring in contemporary Western countries (e.g., 15, 16) and also to self-report personality trait levels, particularly for some traits relevant to prosocial versus antisocial tendencies (17). We therefore also show in Fig. 1 the sibship size differences in the HEXACO variables after adding controls for upbringing religiousness and current religiousness. As seen in that figure, the addition of these religiousness controls reduced the differences in Honesty-Humility and in Agreeableness between sibship sizes of 1 and 6+ by about one-quarter, to about d = 0.25. For Extraversion and for Conscientiousness, differences between sibship sizes of 1 and 6+ were reduced below d = 0.10. (The religiousness variables themselves showed increasing means as a function of sibship size: the differences in upbringing religiousness and in current religiousness between sibship sizes of 1 and 6+ were d = 0.86 and d = 0.73, respectively.)

Birth Order and Personality within Sibship Sizes.

Fig. 2 shows separately by sibship size the standardized mean scores on the HEXACO factor- and facet-level scales in Sample 2 for oldests, youngests, and middles, after controlling for respondent sex and age. For sibship size 2, youngests and oldests are reported, and for sibship size 3, middles are also added; we also show results for those three birth orders in the combined 4, 5, and 6+ sibship sizes, for which sample sizes were smaller.

Fig. 2.

Mean z-scores on HEXACO factor scales by birth order category within each sibship size (Participant age and sex controlled). Error bars show 95% CI. For sibship size 2, Ns were 16,525 (oldest) and 136,91 (youngest). For sibship size 3, Ns were 7,665 (oldest), 6,262 (youngest), and 6,671 (middle). For sibship size 4+, Ns were 4,433 (oldest), 3,620 (youngest), and 7,563 (middle).

Within sibship sizes, oldests averaged nearly d = 0.10 lower in Honesty-Humility than did both middles and youngests, whose means were about the same. Oldests also averaged lower in Agreeableness than did both middles and youngests, but with smaller differences. Among the other factors, the largest differences involved Openness to Experience, where oldests averaged d ≈ 0.10 higher than both middles and youngests.

Discussion

Two dimensions relevant to cooperative tendencies—Honesty-Humility and Agreeableness—showed similar differences between birth order categories and across sibship sizes. The differences are of moderate size, as the middle-versus-only differences exceeded d = 0.20, and the sibship size 6+ versus sibship size 1 differences reached or exceeded d = 0.30. The sibship size differences were reduced by about 25% when religiousness variables (upbringing and current) were controlled. Within sibship sizes, oldests averaged slightly lower than middles and youngests on both dimensions, with differences of nearly d = 0.10 for Honesty-Humility and less for Agreeableness. The finding that differences between middles and both youngests and oldests in Honesty-Humility and Agreeableness became smaller when sibship size was controlled suggests that these birth order differences mostly reflect differences in sibship size.

The largest effect sizes reported here fall between the conventional thresholds for small and medium-sized effects (d = 0.20 and 0.50, respectively; 18). Consider the difference in HEXACO Agreeableness between sibship sizes of 1 and of 6+, which was d = 0.36. Given that the distribution of HEXACO Agreeableness is approximately normal, this difference implies that if one person is selected at random from each of the two sibship size groups, there is a 60% chance that the higher-scoring person will come from the sibship size 6+ group. A difference of this size also implies that the threshold that divides the top 5% of the sibship size 1 group (i.e., its 95th percentile) would be exceeded by about 10% of the sibship size 6+ group; conversely, about 10% of the sibship size 1 group would be below the 5th percentile of the sibship size 6+ group.

Comparisons with Previous Findings.

These results showing birth order and sibship size differences in cooperativeness-related traits are in sharp contrast to those obtained from the two previous large-sample studies (3, 4). We first compare the sibship size results of the present study with those obtained by Rohrer et al. (3). We focus on their results for Big Five Agreeableness, which usually correlates positively with HEXACO Agreeableness and Honesty-Humility (19). The values reported in Table S2 of Rohrer et al. allow us to calculate z-scores for Big Five Agreeableness among persons from sibship sizes of 2 (z = −0.005; N = 8964), 3 (z = 0.024; N = 5497) and 4 (z = 0.022; N = 2569). These values show no systematic increase across sibship sizes. In contrast, our results for HEXACO Agreeableness and Honesty-Humility do show clear monotonic increases across sibship sizes. For both dimensions, sibship size 4 averaged d ≈ 0.12 higher than sibship size 2.

It is less straightforward to compare the sibship size results of the present study with those of Damian and Roberts (4). However, one of their scales, Social Sensitivity, was found in a later study (6) to be strongly positively correlated with Big Five Agreeableness (and moderately correlated with the other Big Five factors). Damian and Roberts reported a correlation of -.07 between sibship size and Social Sensitivity. In their study, sibship size had a mean of 4.04 and an SD of 2.15. The difference between sibship sizes of 1 and 6 therefore represents a 2.32 SD difference (i.e., (6 – 1)/2.15), which gives an estimate of d = −0.16 (i.e., −0.07 × 2.32) between those two sibship sizes for Social Sensitivity. That is, the results of Damian and Roberts are in the opposite direction to those of the present study, as they found lower levels of this cooperativeness-related trait (Social Sensitivity) with larger sibship sizes. However, we think that this difference across sibship sizes is not due to any cooperativeness-relevant construct but rather due to some socially desirable response tendency that is common to the Project Talent Personality Inventory scales: As noted in the Introduction, nine of those 10 scales showed intercorrelations between 0.22 and 0.61, averaging 0.44. Those same nine scales correlated −0.04 to −0.07 with sibship size.

Why did our research find higher levels of cooperativeness-related traits in larger sibships when previous large-sample studies did not? We think that our use of the HEXACO structural model in the present research is not the likely explanation. Even though HEXACO Honesty-Humility is not well represented in many Big Five personality measures (such as those used in Rohrer et al.), the present findings involved not only Honesty-Humility but also HEXACO Agreeableness and the interstitial Altruism facet (SI Appendix, Table S3), the content of which is represented in Big Five measures and also seems to be represented in the Social Sensitivity scale of the Project Talent Personality Inventory.

We suspect instead that the null (or even opposite) findings of the previous large-sample studies are due in part to the measures used in those studies. Specifically, the Big Five Agreeableness scales used by Rohrer et al. were very brief or very narrow in their content. Also, as we have already noted, the Project Talent Personality Inventory scales used by Damian and Roberts were substantially intercorrelated, suggesting a lack of discriminant validity. We think that the use of the HEXACO-100 scales in the present research, with their higher reliabilities, broader content coverage, and smaller interscale correlations, is likely to have given more accurate indications of birth order and sibship size differences in personality. It is also possible that the participants of the present study were highly motivated to provide accurate self-reports in order to receive a meaningful personality profile.

We note that most of the participants in the present research were from English-speaking countries and that the remaining participants also responded in English. We do not know the extent to which our results would generalize to samples of persons from non-English-speaking countries who do not read English.

Origins of Differences in Honesty-Humility and Agreeableness Between Sibship Sizes and Birth Orders.

To the extent that having more siblings is associated with higher Honesty-Humility and higher Agreeableness, one can consider why this happens. A common-sense possibility is that when one has more siblings, one must more frequently cooperate rather than act on selfish preferences. This ongoing situation might then promote the development of cooperative tendencies generally. Given that both Honesty-Humility and Agreeableness are related positively to sibship size, any such development of cooperative tendencies apparently involves both a tendency not to exploit others (Honesty-Humility) as well as a tendency not to overreact to perceived exploitation by others (Agreeableness). Ashton & Lee (13) discuss these two aspects of cooperative or reciprocally altruistic tendency.

The associations between sibship size and cooperative traits became somewhat smaller when we added controls for religiousness. This result suggests that religiousness contributes modestly to these associations, such that religious families tend to have more children (15) and such that the children of those families tend to be slightly more prosocial as adults (17). Most of the link between sibship size and cooperative traits remained when religiousness was controlled, with d ≈ 0.25 between sibship sizes of 1 and 6+ for each of Honesty-Humility and Agreeableness. However, other variables (not measured in the present research) may also contribute to the observed associations between sibship size and cooperative traits.

As noted above, sibship size differences in Honesty-Humility and Agreeableness were of moderate size, with ds above 0.30 when comparing a sibship size of 6+ versus a sibship size of 1. By contrast, birth order differences in these dimensions within a given sibship size were quite small, with even the largest differences—typically between oldests on the one hand and middles or youngests on the other hand—only reaching d ≈ 0.10. If we assume provisionally that these differences represent causal influences of family structure on these personality characteristics, then they suggest a modest common (i.e., shared, between-family) environment influence due to sibship size plus a weaker unique (i.e., nonshared, within-family) environment influence due to birth order independent of sibship size.

The suggestion of a possible sibship size effect might appear to conflict with the recurring finding of little or no common environment influence on personality trait levels. However, the latter finding is easily reconciled with the present results. Recall that the difference in Agreeableness between sibship size groups of 1 and 6+ was d = 0.36, which would correspond roughly to r = 0.18 if those groups were of equal size. However, when the correlation between Agreeableness and sibship size is computed across all our participants—two-thirds of whom had sibship sizes of 2 or 3—the value is only r = 0.09. Even if this correlation were entirely due to a causal effect of sibship size on Agreeableness, it would represent a common environment effect of only about 1%.

Sibship size differences in Honesty-Humility and Agreeableness could be due in part to heritable differences in those dimensions between persons having more or fewer children. But this genetic contribution is unlikely to be large, because midparent/offspring regressions for personality traits tend to be modest [about 0.25 based on the mean single parent/offspring correlations of 0.17 in (20)]. Even if we assume generously that parents of 6+ children average d = 0.50 higher in these dimensions than do parents of 1 child, this would not account for more than about a d = 0.125 difference between the two offspring groups. Nevertheless, a partially genetic origin to these sibship size differences in personality trait levels is plausible.

Sibship size differences in cooperativeness traits might also result from a tendency for parents to have more children when their current children tend to be cooperative. There is some evidence that parents are more likely to have another child when their current child tends to be prosocial (21). However, this effect is rather weak, and we think it accounts for at most a small part of the link between sibship size and cooperativeness traits.

Origins of Only/Non-Only Differences.

Our results showed that onlys, as compared with persons raised with siblings, were about d = 0.15 lower in Honesty-Humility and Agreeableness and about d = 0.10 higher in Openness to Experience. The origins of these rather small differences may vary across personality dimensions. Recall that mean scores on Honesty-Humility and Agreeableness showed a monotonic increase with sibship size. Although the differences between sibship sizes of 1 and 2 were somewhat larger than the differences between successive sibship sizes thereafter (Fig. 1), the lower means for onlys (i.e., sibship size 1) seem mainly to reflect a continuation of a trend for lower means with smaller sibship sizes. In contrast, mean scores on Openness were higher for a sibship size of 1 than for a sibship size of 2, but there was no consistent decrease across sibship sizes between 2 and 6+. Thus, for Openness, the results suggest some special influences for onlys, with these influences being shared to some extent by oldests (cf. Panel 6 of Figs. 1 and 2).

Openness to Experience and Birth Order.

As with previous results, Openness to Experience also showed associations with birth order, with oldests and onlys averaging higher than youngests and middles. When sibship size was controlled, the largest differences observed for the Openness to Experience facet scales involved Inquisitiveness (SI Appendix, Table S4), which assesses intellectual curiosity. This result is comparable with findings for a self-report Intellect scale in Rohrer et al. (3). These differences might reflect a tendency for the presence of older siblings to hinder slightly the intellectual development of younger siblings.

The finding of higher average levels of Openness to Experience in oldests than in youngests appears consistent with differences in the numbers of oldests and youngests in our sample. The ratio between them was above 55:45 rather than the expected 50:50 (Table 1). Furthermore, the proportion of oldests was significantly higher than the expected proportion within each sibship size (i.e., 55%, 37%, 30%, and 27% for sibship sizes of 2, 3, 4, and 5). We suspect that this ratio reflects a tendency for oldests to be more inquisitive than nonoldests, because persons who voluntarily complete an online personality inventory tend to be more intellectually curious (22).

Conclusion

In an editorial about the findings of Damian and Roberts (4) and Rohrer et al. (3), Damian and Roberts (23, p. 14120) argued—reasonably, in our view—that those studies

“should be the standard against which any new studies on birth order and personality are considered. The largest, most methodologically sophisticated studies in existence show little or no functional relation between birth order and personality. Newer data will have to provide evidence for much larger effects in equally large samples to counter the weight of the evidence.”

We believe that the present data—which have the additional advantage of better differentiated personality measures—provide such evidence. The weight of that evidence now indicates that personality trait levels do differ as a function of birth order and sibship size.

Materials and Methods

The data for this report were obtained from the hexaco.org website, where interested persons have completed the self-report HEXACO Personality Inventory—Revised (HEXACO-PI-R) in its 100-item version (14) to receive feedback about their personality profiles. The HEXACO-PI-R includes six broad factor scales assessing the major dimensions of personality: Honesty-Humility, Emotionality, Extraversion, Agreeableness (versus Anger), Conscientiousness, and Openness to Experience. Each factor scale includes four narrower “facet” trait scales; a 25th facet scale represents a blend of three HEXACO dimensions. Each facet scale in the HEXACO-100 consists of four items, of which one to three are reverse-scored; each item has a 1-to-5 response scale. See ref. 14 for evidence of the convergent and discriminant validity of the HEXACO-100 variables. After completing the inventory and before receiving their feedback, participants also answer a series of demographic questions, such as those for age, sex, country, and others.

Two of these demographic questions are the focus of the present report. Beginning in August 2015, a question about birth order was administered:

Which of the following describes your position in the household in which you grew up?

• The only child

• The oldest child (of two or more children)

• The youngest child (of two or more children)

• A “middle” child (neither oldest nor youngest of three or more children)

Beginning in June 2023, a question about sibship size was also administered:

• When you were a child, how many children (including yourself) were raised together in your household?

$$123456 or more$$

Sample 2 includes persons who participated after the sibship size question was introduced; Sample 1 includes all other participants of this report. Participants of Sample 2 had the option of completing the HEXACO-100 in any of 43 languages, but to maximize comparability with Sample 1, we report results only for English-language participants. In both samples, over 60% of participants were from the United States of America, the United Kingdom, Canada, or Australia.

The questions above do not allow information about the specific birth ranks of middle children from families with four or more children, about the relative ages of children in the household, about any family dissolution or reconstitution, and about whether any children were twins or higher-order births. To the extent that these complications tend to make some respondents less prototypical exemplars of the various birth orders or sibship sizes, they should tend to weaken any differences between groups, causing our results to underestimate those differences.

Participants of both samples also answered two questions about their religiousness: “How religious was your upbringing?” and “How religious do you consider yourself to be?” Both questions were answered on a 7-point scale (1 = not religious at all; 7 = extremely religious). These two religiousness items (“religiousness of upbringing” and “current religiousness”) intercorrelated 0.49 and 0.51 in samples 1 and 2, respectively. We note that single-item religiousness scales similar to ours have shown test–retest reliabilities above 0.90 and correlations of about 0.90 with full-length religiousness scales (see 24).

For Sample 1, participants’ responses were screened such that data were retained for analysis from all participants who completed all HEXACO-100 items, who passed all three “attention check” items, who passed two indices assessing variation and consistency in personality item responses (14), who reported their sex and an age between 18 and 94 years, and who responded to both religiousness items. For Sample 2, the above requirements were again applied, but participants in this sample were excluded if their response to the birth order question was inconsistent with their response to the sibship size question [specifically, (a) a birth order of “only” and a sibship size other than 1, (b) a birth order of either “oldest” or “youngest” and a sibship size of 1, or (c) a birth order of “middle” and a sibship size of 1 or 2].

Ethics approval for this study was obtained from the University of Calgary Conjoint Faculties Research Ethics Board (REB14-0792).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

The conditions of participant consent do not allow us to post the datasets to a publicly available repository. The datasets will be sent to researchers upon request. The SPSS code files are available in https://osf.io/vs42a/ and can be used with the SPSS datasets to be sent.