Why do men tend to fare better in competitive environments than women do? It is a question researchers have long pondered. More than a decade ago, two economists—Claudia Goldin of Harvard and Cecilia Rouse of Princeton—found sex-based discrimination in the hiring of orchestra musicians. Beginning in the 1970s, many orchestras in the United States, hoping to remedy a dearth of women in their ranks, began conducting blind auditions. Musicians performed behind a screen so the selection committee could not see them. Goldin and Rouse compared several orchestras before and after the change—and found that women were selected 30 percent more often when auditions were blind.
For Paola Sapienza, a professor of finance at the Kellogg School of Management, the finding was fascinating—yet it did not explain the reason for the bias to begin with. Was the gender bias a result of “belief-based discrimination,” in which the judges assumed that women are less competent; of “taste-based discrimination,” in which the judges simply preferred men for reasons other than competence, such as a belief that the audience prefers male musicians; or of “statistical discrimination,” in which gender confers some other valuable information (e.g., men are preferable because they presumably do not take as much time off)? “The data in the paper were able to establish a bias,” says Sapienza, “but they could not distinguish between the three of them.”
Distinguishing Between Types of Discrimination
So Sapienza, along with Ernesto Reuben, an assistant professor at Columbia University; Pedro Rey-Biel, an associate professor at the Universitat Autònoma de Barcelona; and Luigi Zingales, a professor at the University of Chicago, set out to look more closely at gender bias in competitive settings. They wanted to know why organizations might fail to select high-performing women for jobs at which the women would excel. They devised an experiment that eliminated the possibility of discrimination based on taste or statistics, to test whether belief-based discrimination could account for gender bias on its own.
“There is big evidence that ten years out women tend to earn 60 percent less than men.”
They divided MBA students into teams and had each team select a leader to represent them in a competition, which involved doing a series of calculations. The subjects in the study had all previously performed the same task, two years earlier, so they could use their past performance as a means of predicting their chances of winning this time around. Before the selection took place, the students were asked to recall how well they had done in the same task two years ago. They were also asked to predict how well they would do on the same task again.
Members of the team whose leader won the competition all received a cash prize, so everyone on the team had a clear incentive to choose the best leader; the leader received no more money than anyone else. Teams had five minutes in which each participant could make the case for how well they thought they would do in the competition, and then to choose a leader based on that information. “Barring any explicit discrimination against women—which would be unlikely in an experiment with university students—groups should aim to select their most talented individual irrespective of gender,” Sapienza and her colleagues wrote.
But that is not what happened. Instead, women were selected as group leaders 33.3 percent less frequently than they should have been based solely on how well they did in the earlier competition. “Women are selected much less often as leaders than is suggested by their individual past performance,” the researchers wrote.
“There is big evidence, even among women MBAs, that ten years out women tend to earn 60 percent less than men,” Sapienza explains, “even when they start in the same field with very similar salaries. The question is, Why do women lag behind over time? To what extent is it the result of evaluation that’s done outside, and how much does it interact with the self-promotion of men and women?”
To understand what was driving their results, Sapienza and her colleagues tested three possible theories as to why women may be less frequently selected as leaders. The first was a difference in the way men and women judge their own abilities. The second was a difference in how men and women describe their own abilities. And the third was a difference in how men and women deal with what the researchers call “agency problems,” or how they “respond to conflicts of interest between their own interest and the group’s.”
The answer turned out to be theory number two. By far the most important reason why women were being underrepresented as team leaders was how they portrayed their abilities relative to men’s. Both men and women in the study were inclined to overstate how well they had done in the earlier competition and how well they would do in the future, but men were far more willing to do so.
What troubled Sapienza and her colleagues, though, was not that women tended to overstate their abilities less than men. It was that the people evaluating candidates for leadership did not discount men’s tendency to boast more. “The fact that men tend to overstate, that’s not a surprise,” Sapienza says. “But if everyone knows that, why don’t they just say, ‘If men say 5, it must be 3.’ There is some discounting, but it’s very minimal.”
These biases are a cause for some concern—in the real world, companies and organizations that can identify and choose their best members have an advantage over those that cannot. On a personal level, to overcome this, women are typically coached to boast more, Sapienza says. But that does not help companies and organizations, which would then face a field of candidates who greatly inflate their reported abilities. The solution, Sapienza suggests, is to handicap the biggest braggarts, leveling the playing field for everyone.
Related reading on Kellogg Insight