Working together

Think manager, think male

In 1973 the psychologist Virginia Schein gave a list of personality traits to American managers and asked them to describe successful managers, women in general, and men in general. The traits that went with “manager” matched the traits they gave to men far better than the ones they gave to women. She called the result “think manager, think male.” Did it only describe the United States? A 1996 study by Schein and colleagues asked 361 men and 228 women studying management in Japan and China (the results were pooled for the two countries) and found the same thing: both men and women saw the traits of a successful middle manager as ones more common in men. Among male students the pattern held in all five countries studied, including Britain and Germany; among female students it held everywhere except the United States, where women saw men and women as equally likely to have the traits. A meta-analysis of the later literature found that leader stereotypes were still masculine on all three of its measures, but that the masculinity of the stereotype had been declining over time.

The stereotype is not the same as the numbers, but the numbers are lopsided too. Katharine Graham of the Washington Post became, in 1972, the first woman to be chief executive of a company on the Fortune 500 list. By 2023 there were 53 women among the 500, or 10.6 percent (52, or 10.4 percent, a year later). In China, a business-leader survey by Grant Thornton put women’s share of senior management at 35.3 percent in 2025, against 34.0 percent in the United States, though the survey does not say how it defines senior management. At the very top of politics the picture is different: the 24-member Politburo of the Chinese Communist Party elected in 2022 has no woman in it. The Wall Street Journal gave the barrier its name in 1986, in a story headed “The Glass Ceiling.”

Does a woman have to be better to be seen as equal? A well-known study of orchestras tried to measure it. By 1970 women were fewer than one in twenty players in the five top American orchestras. In a study first circulated in 1997 and published in 2000, the economists Claudia Goldin and Cecilia Rouse looked at auditions, most of them held behind a screen so that the committee could not see the player. A screen raised a woman’s chance of getting past the first round, and the authors estimated that the change could account for a good share of the rise in women among new hires. They also stated plainly that some of their estimates had large standard errors and that one effect, in a later round, went the other way. It is a suggestive study, not a closed case.

Do women lead differently?

The popular account says women are more empathic, more patient, and more ethical as leaders, and so make better human-resources managers. The studies are more cautious. A meta-analysis of 45 studies by Alice Eagly and colleagues found that women leaders were somewhat more “transformational” (inspiring and developing others), and more likely to reward good work, while men were more likely to manage by exception or to avoid deciding at all; the differences were small. An earlier review found that men and women leaders were equally effective overall: men did better in roles defined in more masculine terms, and women in roles defined in less masculine terms. Whether women are less corrupt is similar. A famous 2001 paper found that countries with more women in government had less corruption, but a later study of 76 democratic-leaning countries found the link was not universal and was strongest where officials can be voted out. That is a conditional pattern in correlational data, not a trait of women.

Who competes, and who asks

In 2007 the economists Muriel Niederle and Lise Vesterlund had volunteers solve arithmetic problems, first for a fixed payment per correct answer and then in a tournament, where only the winner of each four-person group was paid. They then let people choose which to do again. Women and men performed equally well, but about twice as many men chose the tournament, and differences in ability, in beliefs about winning, and in dislike of risk explained only part of the gap. Was it nature, or something about the society? In 2009 the economists Uri Gneezy, Kenneth Leonard, and John List ran the same game among the Maasai of Tanzania, who trace descent through men, and the Khasi of India, who trace it through women. Among the Maasai, about half the men and a quarter of the women chose to compete, close to the pattern in American students; among the Khasi the pattern reversed, with more women than men choosing the tournament.

China offers a natural experiment of its own. In 2009 the economist Y. Jane Zhang ran the same game with 272 high-school students in Ninglang county, Yunnan: Han Chinese, who were exposed to the communist reforms of marriage, education, and work, and the Yi and the Mosuo, who were largely exempt from them. About 48 percent of Han girls and 63 percent of Han boys chose the tournament, a gap of 15 points that was not statistically significant in so small a sample. Among the Yi the figures were 38 and 60, and among the Mosuo, who trace descent through women, 48 and 75. When Zhang allowed for confidence, skill, and attitude to risk, she found no significant gap for the Han and gaps of about 24 and 30 points for the Yi and Mosuo. Her reading is that reforms promoting women’s work may have raised girls’ taste for competition, a suggestion she describes as a hypothesis: the study is observational and small, and the matrilineal Mosuo show the largest gap of all, a reminder that “matrilineal” does not simply mean “egalitarian.”

What about risk? The idea that women avoid risk is one of the best known in economics, and one of the most qualified. In 15 sets of experiments with one simple investment game women invested less. But a review of 54 replications of another standard test of risk attitudes, covering about 7,000 people, found differences in fewer than one study in ten, and tiny ones once pooled; another reanalysis found that the male and female distributions overlap by more than 80 percent. Two field experiments in Yunnan, among the Mosuo and the neighboring Yi, found women somewhat more risk-averse in both groups, and the gap smaller among the Mosuo.

Asking for more

If competing and asking are different from ability, they might help explain pay. A meta-analysis of 123 effect sizes from 10,888 negotiators by Mazei and colleagues found that men did get better economic outcomes on average, but that the gap depended strongly on the situation: it was smaller when negotiators had experience, when they knew the range of possible deals, and when they were negotiating for someone else, and it reversed in the situations that were least awkward for women. In a natural field experiment by Andreas Leibbrandt and John List, nearly 2,500 job seekers responded to job advertisements. When the ad did not say wages were negotiable, men were more likely to ask for more, and women to signal that they would work for less. When the ad said wages were negotiable, the difference vanished. The men also preferred jobs with vague wage rules, which made the gaps larger there.

Pay itself is a different measure. In the United States in 2014 full-time women earned about 79 percent of men’s annual earnings and 83 percent of weekly earnings, according to an overview by Francine Blau and Lawrence Kahn. Allowing for education and experience, the hourly figure was about 82 percent in 2010; adding industry, occupation, and union status, about 92 percent, though “unexplained” is not the same as discrimination, and occupation and industry are themselves partly choices that bias and expectations may shape. In 2023 the economist Claudia Goldin won the Nobel prize in economics for advancing our understanding of women’s labor market outcomes. We found no single reliable headline figure for China’s pay gap, though a 2024 study reported that the gap has not narrowed in the 21st century and is largest in the private sector and smallest in the public sector.

Trust, and who cooperates

In 1995 the economists Joyce Berg, John Dickhaut, and Kevin McCabe devised a game of trust that has been run thousands of times. A volunteer receives $10 and may send any part of it to an unseen partner. The experimenter triples whatever is sent, and the partner decides how much to send back. In their first group, 32 pairs of University of Minnesota students, 30 of the 32 senders sent something, and the average sent was $5.16 out of $10. The average sent back was $4.66, so trusting cost the senders about 50 cents on average. The $5.16 is an overall average. The paper reports no breakdown by sex, so it cannot be a difference between men and women, as is sometimes retold. Try the game below.

Does sex matter? In one study of American participants, men trusted more, and women proved more trustworthy by sending back more. And are women more cooperative, men more competitive? A meta-analysis of 272 comparisons in social dilemmas over 50 years found an overall difference of −0.05, essentially zero. The context mattered: groups of men cooperated more than groups of women, while women cooperated more in mixed-sex pairs and in larger groups. The idea that women cooperate and men compete does not describe how people behave in these games.

Do mixed teams do better?

A popular argument for mixed teams is that groups with women are smarter. It comes from a 2010 study in Science by Anita Woolley and colleagues. They put 699 people into groups of two to five, gave them tasks such as visual puzzles, brainstorming, moral judgments, and negotiating over limited resources, and found that how well a group did on one kind of task predicted how well it did on the others, a “collective intelligence” factor. It did not depend much on the average or highest intelligence of the members. It went with the members’ average social sensitivity, with how evenly they took turns in talking, and with the proportion of women in the group. Social sensitivity was measured with the “reading the mind in the eyes” test discussed in Lesson 3.

Did it hold up? A 2017 attempt with 312 people in three studies found that the members’ individual intelligence explained about 80 percent of the differences between groups, and did not find support for the effect of women or turn-taking. A much larger analysis in 2021, pooling 22 studies and 5,279 people in 1,356 groups, did find that groups with a higher proportion of women scored higher, and that this was explained by their members’ better average social perceptiveness. It also found that how a group worked together predicted its score better than the individual skills of its members. The two results are difficult to reconcile, and they leave a small, real, but not yet fully understood effect.

Real teams have been studied too. In a field experiment in the Netherlands, Hoogendoorn, Oosterbeek, and van Praag assigned 550 first-year business students at random to 45 teams that ran a small company for a year. Teams with an equal mix of women and men sold more and earned more profit than teams dominated by men; the estimates were roughly halved by more robust statistical methods, there were too few female-dominated teams to say the pattern was symmetric, and none of the explanations the authors tested was supported. A 2022 study of 6.6 million medical-science papers found that mixed-gender teams produced more novel and more cited work, and the more balanced the team the better the result, though that is a correlation. Against this, Alice Eagly noted in 2016 that meta-analyses of women on boards and of gender diversity in task groups find average effects that are “null or extremely small,” and that effects look positive in some contexts and negative in others. A reasonable conclusion is that diversity is no guaranteed boost, but that the best teams make use of what each member can do.

Use everyone’s strengths

The old Chinese statesmen did not need a meta-analysis to see that a leader’s job is to draw on other people. The Han Feizi (third century BCE) says, in the section on “the way of the ruler” in chapter 48: 下君尽己之能，中君尽人之力，上君尽人之智, “the lowest ruler uses up his own ability, the middling ruler uses up the strength of others, the best ruler uses up the wisdom of others.” An American general of the Second World War said something like it. George Patton, who led the US Seventh Army in Sicily and the Third Army in France, wrote in his reflections on war that people should be told what to do and not how to do it, and “they will surprise you with their ingenuity.” Both ideas assume that the people you lead differ from one another and from you, and that the differences are a resource.

That is also the best use of what this course has found. Knowing the average of a group tells you where to look, not what you will see. The overlap between men and women on almost every trait is so large that the person in front of you is better known by asking than by guessing. And when a message matters, say it clearly and check it. Chinese has a proverb for that: 口说无凭，立字为据 (kǒu shuō wú píng, lì zì wéi jù), “spoken words prove nothing; write it down as evidence.” Minds also fill in what they expect to hear, as Course 10 shows, which is another reason to check what each of you understood.