The characteristics shown in the table below have a special bearing on the People-Things dimension.
PERSONALITY DIFFERENCES RELATING TO THE PEOPLE-THINGS DIMENSION
Warm, outgoing, attentive to others
Costa:
Kajonius:
Del Giudice: +0.89
Inventory: 16PF
Sensitive, aesthetic, sentimental
Costa:
Kajonius:
Del Giudice: +2.29
Inventory: 16PF
Cooperative, accommodating, deferential
Costa:
Kajonius:
Del Giudice: +0.54
Inventory: 16PF
Shows warmth toward others
Costa: +0.33
Kajonius: +0.07
Del Giudice:
Inventory: FFM
Altruistic concern for others
Costa: +0.43
Kajonius: +0.51
Del Giudice:
Inventory: FFM
Sympathizes with others
Costa: +0.31
Kajonius: +0.57
Del Giudice:
Inventory: FFM
Enjoys the company of others
Costa: +0.21
Kajonius: +0.05
Del Giudice:
Inventory: FFM
Straightforwardness, not demanding
Costa: +0.43
Kajonius: +0.40
Del Giudice:
Inventory: FFM
Source: Costa, Terracciano, and McCrae (2001); Del Giudice, Booth, and Irwing (2012); Kajonius and Mac Giolla (2017); Kajonius and Mac Giolla (2017). Positive scores indicate women score higher.
A person who is warm, sympathetic, accommodating, altruistic, and sociable amounts to the stereotype of a human being, male or female, who is more attuned to people than things. Women are more likely to have that profile than are men.
People who are somewhat to the other side of each trait in the table are reserved, utilitarian, unsentimental, dispassionate, and solitary—which amounts to the stereotype of a human being, male or female, who is more attracted to things, broadly defined, than to people. Men are more likely to have that profile than are women.
With the 16PF inventory, just two factors, sensitivity (d = +2.29) and warmth (d = +0.89), tell most of the story. With the FFM inventory, the individual effect sizes from both studies are modest, with the largest being just +0.57 for “sympathizes with others.” But a scan of the table also makes an obvious point: All five of the FFM traits add up. I do not have the raw data for computing the aggregated difference (D) on the traits included in the table, but some other statistics will give you a sense of the overall sex difference in personality that they reflect.
For the 16PF inventory, the Del Giudice study calculated the value of D for all 15 factors. It was 2.71, a huge difference that would leave only 10 percent overlap between two normal distributions.10 Even when the extremely large difference (+2.29) on just one of the factors, sensitivity, is excluded, the value of D is 1.71, corresponding to 24 percent overlap between two normal distributions.11 If instead we use the mean of those 15 separate effect sizes (again excluding sensitivity), the overall difference would be estimated at just +0.44—a dramatic illustration of the difference between averaging effect sizes and aggregating them.
With regard to the FFM inventory, we have reason to be confident that aggregating the effect sizes for the five traits most closely related to the People-Things dimension would produce a D much larger than their mean d of +0.32. To give you an idea, a large-sample (n = 8,308) administration of the FFM in 2006 had an average sex difference in d of +0.30.12 The value of Mahalanobis D for that dataset was 0.98.13
THE EVIDENCE FROM INFANCY
Measuring personality sex differences in infancy is tough, and the instruments for doing so are not nearly as precise as instruments for older children. Different studies come up with different estimates of some relationships, and almost all of the studies need replication. The most dramatic example of a finding from infancy, which led to considerable publicity, was a 2002 study presenting evidence that newborn girls no more than two days old after birth showed stronger interest in a human face while the newborn boys showed stronger interest in a mechanical mobile.14 It is a single, unreplicated study with a sample of 102, not proof to take to the bank, but its finding was in line with many other studies that have found personality sex differences in infants.
On average, infant girls cry longer than boys in response to recordings of another baby crying, believed to be a primitive empathic reaction.15
On average, infant girls hold eye contact with an adult human longer than boys do.16
On average, infant girls show more expressions of joy than boys at the appearance of the mother.17
On average, infant girls are more responsive to maternal vocalizations than infant boys.18
On average, infant girls are more distressed by maternal “still face” than infant boys.19
On average, infant girls show visual preferences for objects with human attributes while boys show more visual preferences for balls and vehicles.20
On average, infant girls are more likely to initiate and respond to joint attention.21
In Erin McClure’s meta-analysis of 20 studies of facial expression processing in infants, the six studies for which effect sizes were reported or could be calculated had a weighted effect size of +0.92 favoring girls.[22]
Sex Differences in Personality Worldwide
So far, I have presented nothing indicating that these personality differences are hardwired. Maybe that’s just the way little girls and little boys are brought up in the United States and other Western cultures. That’s where cross-national comparisons come in. The legal and social status of women varies widely around the world. Some Islamic cultures still keep women at a level of legal subservience little better than Western women experienced until the twentieth century. Some sub-Saharan African cultures still take the superiority and dominance of men for granted and organize daily life accordingly. At the other extreme are countries in Western Europe and especially Scandinavia that have erected elaborate structures to require gender parity in all economic and social matters.
Cultures around the world have other deep differences that affect both women and men—for example, the intensely family-oriented cultures of much of Asia compared to the individualism of the Western tradition. And yet despite this extremely wide range of environments in which children are raised, sex differences in personality are remarkably similar around the world.
The same article that reported the results for American adults on the Costa-McCrae inventory also reported them for 25 other countries.23 In 2005, McCrae and Antonio Terracciano used observer reports from 50 cultures, 22 of which had not been included in previous studies. The next table shows effect sizes for the same five traits from the Costa-McCrae inventory shown in the previous table, adding the results from the international samples.
High-end descriptors: Shows warmth toward others
Questionnaire data
U.S.: +0.33
25-nation sample: +0.23
Observational data
50-nation sample: +0.29
High-end descriptors: Appreciates art and beauty
Questionnaire data
U.S.: +0.34
25-nation sample: +0.35
Observational data
50-nation sample: +0.31
High-end descriptors: Has altruistic concern for others
Questionnaire data
U.S.: +0.43
25-nation sample: +0.25
Observational data
50-nation sample: +0.33
High-end descriptors: Sympathizes with others
Questionnaire data
U.S.: +0.31
25-nation sample: +0.28
Observational data
50-nation sample: +0.39
High-end descriptors: Enjoys the company of others
Questionnaire d
ata
U.S.: +0.21
25-nation sample: +0.14
Observational data
50-nation sample: +0.26
Source: Costa, Terracciano, and McCrae (2001); McCrae and Terracciano (2005). All samples are adults. Positive scores indicate women score higher.
The results show universally higher female means and similar effect sizes on the individual traits. Even taken country by country, the number of anomalies was remarkably small. The Costa study reported effect sizes for extraversion, agreeableness, and openness for 26 populations in 25 countries—78 effect sizes in all. The signs for 77 out of the 78 were positive (women scored higher).24 The McCrae study of 50 cultures reported country-by-country effect sizes for 49 populations in 46 countries. Of the 147 effect sizes reported, 139 were positive. The largest of the negative effect sizes (i.e., higher for males) was trivially small (d = –0.05).25
This consistency is all the more remarkable considering that the 50 nations included ones from East Asia, South Asia, the Mideast, Africa, Europe, South America, and North America, and nations that ranged from the most impoverished and traditional (e.g., Uganda, Burkina Faso) to the wealthiest and most sex-egalitarian (e.g., Sweden, Denmark). The great cultural and economic disparities across these countries make it difficult to see how all of them could produce uniform socialization of girls to be more warm, altruistic, sympathetic, sociable, and artistically sensitive than men.
Sex Differences in Personality and a Society’s Gender Egality
I use gender egality in preference to gender equality to signify not just progress toward diminishing sex differences but also institutional, legal, and social changes intended to put men and women on an equal footing. The question at hand is whether sex differences in personality are smaller in countries that have made the most progress.
The theories of socialization and of social roles that I summarized in chapter 1 necessarily expect that the answer is yes. If sex differences in personality are artificial, diminishing the causes of artificial differences must eventually lead to smaller differences.26 The only question is how long it will take. This brings us to a counterintuitive finding that seems to cut across a variety of sex differences: Many sex differences in cognitive repertoires are wider rather than smaller in countries with greater gender egality. Personality traits offers the first example.
The Evidence for Wider Personality Differences in Advanced Countries
The Costa study. The Costa study discovered this startling result as they examined the scores for individual nations in their pioneering study. The wrong nations had the largest sex differences: “Sex differences are most marked among European and American cultures and most attenuated among African and Asian cultures,” they wrote.[27]
To convey this finding more systematically, I employ the UN’s annual Gender Inequality Index (GII). It is based on maternal mortality rate, adolescent birth rate, women’s share of seats in parliament, percentage of women with at least some secondary education, and women’s labor force participation.[28] A high score on the GII indicates high inequality.
The results correspond to widespread impressions that Western Europe has the best record for sex equality. Among the 70 nations with data on personality and a GII score, the five nations with the best (meaning lowest) scores on the GII were Switzerland, Denmark, the Netherlands, Sweden, and Iceland. The five nations with the worst (highest) GII scores were Burkina Faso, Congo, Egypt, Pakistan, and Tanzania.
As noted, both socialization and social role theories of sex differences predict that effect sizes should diminish as gender egality increases. Translated into a prediction about the Costa data, this means that a correlation between the absolute size of the sex difference and the GII should be positive (greater inequality is associated with greater personality differences).29 In the Costa study, those correlations were not only negative—the “wrong” sign—but substantially so: –.61 for emotional stability, –.57 for extraversion, –.49 for openness to emotion, and –.42 for agreeableness.30 On average, personality differences were wider in countries with greater gender egality.
The McCrae study. The McCrae study applied the same measures plus one for conscientiousness to a larger sample of nations, using an observational measure of personality traits rather than self-reports. It found the same thing as the Costa study. The correlations between the Gender Inequality Index with the effect sizes for sex differences were once again all in the “wrong” direction and all substantial: –.61 for openness to emotion, –.57 for emotional stability, –.56 for extraversion, –.47 for conscientiousness, and –.43 for agreeableness.[31]
The Schmitt study. In 2008, an international team of behavioral scientists consisting of American David Schmitt, Austrian Martin Voracek, and two Estonians, Anu Realo and Jüri Allik, drew on one of the largest cross-cultural studies of personality ever conducted, part of the International Sexuality Description Project, with three aims in mind.
First, the team wanted to see if the findings of the Costa and McCrae studies generalized to another instrument for measuring personality. The Schmitt study used the Big Five Inventory, consisting of 44 self-report items, rather than the FFM inventory.
Second, the team wanted to increase the range of nations in the database. In all, they obtained personality measures from 55 nations, including 14 that were not part of either the Costa or McCrae studies.
Third, the Schmitt study undertook an elaborate set of tests to determine whether artifacts explained the widening personality differences in advanced countries.
The short version of the answers presented at length in the Schmitt study is that (1) the Big Five Inventory showed essentially the same cross-national patterns that the FFM inventory had produced; (2) the addition of new nations allowed an extension of the conclusions that the Costa and McCrae studies had reached; and (3) the arguments for an artifactual explanation of the widening gap in advanced nations were not borne out by the analyses.32
The Giolla study. In 2018, Erik Mac Giolla and Petri J. Kajonius published the results for a database with a more extensive (120-item) version of the FFM for 22 countries with uniformly larger sample sizes per country (at least 1,000) than the samples used by the other studies. Uniquely, this study also calculated Mahalanobis D—the method for aggregating individual effect sizes that I described in chapter 1. The index of gender egality used for the study was the Global Gender Gap Index (GGGI) published by the World Economic Forum. The index is scored from 0 to 1, with 1 meaning gender equality (or better conditions for women) on all of the 14 indicators.33
In this case, the “right” correlation with the absolute size of the sex difference is negative (a higher score on the GGGI is associated with smaller sex differences). Instead, all of the correlations between personality differences and the GGGI were positive, which means they were all in the “wrong” direction and all were substantial: +.33 for emotional stability, +.33 for openness, +.48 for conscientiousness, +.49 for agreeableness, and +.53 for extraversion. The correlation between the GGGI and the aggregate statistic D was 0.69. The size of D was much larger than the average value of the effect sizes. Mean D was 0.89 compared to a mean for Cohen’s d of 0.24—further evidence of how much difference aggregating conceptually related indicators makes.34
The Falk study. Also in 2018, economists Armin Falk and Johannes Hermle published their analysis of the Global Preferences Survey conducted in 2012. The indicators were not of personality traits per se, but of six preferences that in turn are consistent with personality traits. Four of these preferences were in the social domain: altruism, trust, positive reciprocity (a preference for rewarding positive behaviors), and negative reciprocity (a preference for punishing negative behaviors). Two were nonsocial and had more direct implications for economic behavior: risk-taking and time discounting (preference for a future larger reward than an immediate smaller reward). The sex differences on the four social preferences were all on the side of People-oriented personality traits: On
average, women preferred altruism, trust, and positive reciprocity more than men and were more averse to negative reciprocity than men. In the two nonsocial preferences, men preferred risk-taking and waiting for a larger reward more than women. The analysis employed representative samples from 76 countries.
All of the sex differences on these traits became larger as countries became more economically developed and more egalitarian in their social policies. The correlations of preferences with the authors’ Gender Equality Index were all in the wrong direction: +.51 for altruism, +.41 for trust, +.13 for positive reciprocity, +.40 for negative reciprocity, +.34 for risk-taking and +.43 for patience.35 Greater equality was associated with larger sex differences. The authors did not report an aggregated effect size. However, they did create an index incorporating all six preferences. The correlation between the size of the sex difference on the combined index and the Gender Equality Index was +.56. The correlation was even larger (+.67) for a measure of national wealth, per capita GDP. Or as the authors put it, “These findings imply that both economic development and gender equality exhibited an independent and significant association with gender differences in preferences.”36
Five different studies, based on different measures of personality and national gender egality, analyzing data from dozens of countries, all found the same pattern: overall consistency in male-female differences in personality, but larger differences in the most advanced countries.[37]
Explaining Wider Personality Differences in Advanced Countries
Why haven’t the sex differences in personality gotten smaller in countries that have aggressively adopted gender-egalitarian policies? Why instead, and contrary to all expectations, have they tended to widen?
Costa and his coauthors hypothesized that in traditional societies with strong sex roles, people see behavioral sex differences as socially mandatory, not the result of personal dispositions, whereas people in advanced societies are more likely to see them as evidence of personal dispositions.38 Another possibility is that people tend to compare themselves to others of their own sex in traditional cultures, whereas in advanced cultures people compare themselves to the whole population. For example, a woman in a traditional culture may rank herself on kindness relative to other women. She may be of the opinion that women tend to be kinder than men, but that doesn’t enter into her self-report. In an advanced culture, perhaps a woman compares her kindness to others of both sexes, and a sex difference emerges.39
Human Diversity Page 5
