Gender-Balanced Teams Do Better Work

Larger scientific teams have become much more common over the last 50 years. And these larger teams outperform smaller ones, producing research that is more highly cited by other scholars. That was the key finding of a 2007 paper by the Kellogg School’s Brian Uzzi, a professor of management and organizations, and Benjamin Jones, a professor of strategy.

Now the researchers are looking not just at the size of a team but its composition, too. In new research, Uzzi and Jones found that the gender mix on teams matters. Mixed-gender teams produce more novel and impactful scientific research than teams made up of only men or only women, according to the paper, which was coauthored by Yang Yang of the University of Notre Dame, Yuan Tian of New York University, and Teresa Woodruff of Michigan State University.

In fact, “the more gender-balanced the team is, the better the team does,” Uzzi says. Simply put, “men and women are both part of the recipe for success in science. We’re better together.”

Gender-Diverse Teams Perform Better

Uzzi, Jones, and their colleagues analyzed 6.6 million biomedical science papers published from 2000 to 2019, using an algorithm to infer authors’ genders from their names. (This method, while imperfect and unable to capture complexity in gender expression or identity, was extremely efficient, and produced figures consistent with official data about the gender makeup of medical-school faculty.)

Next, the researchers calculated the impact and novelty of each paper at the time of its publication. Impact was measured by the number of citations the paper received; a paper was considered highly cited if it was in the top five percent of citations in a given year.

The researchers measured novelty by examining the citations each paper contained. “If I see in the reference section of a paper that they cite Leonardo da Vinci and Einstein—and Einstein and Leonardo da Vinci have been cited together in a lot of other papers—we’d consider that a conventional mixture,” Uzzi explains. “But if Einstein is paired for the first time with Maya Angelou in the references of a scientific paper, that would be a novel combination,” and one that suggests new knowledge is being advanced. The researchers used this citation information to compute a novelty “score” for each paper in the data set.

Mixed-gender teams significantly outperformed same-gender teams on both novelty and impact, the researchers found.

A mixed-gender team of six or more researchers was 9.1 percent more likely to produce a novel paper and 14.6 more likely to produce a highly cited paper than a same-gender team of the same size. What’s more, the novelty and impact benefits were strongest when teams were gender-balanced—that is, a team of three men and three women was more likely to produce novel and highly cited research than one with four men and two women.

They then tested whether these findings would persist even when accounting for a variety of factors unrelated to gender that could influence the outcome of a team’s scientific research.

For example, the researchers considered the possibility that teams with men and women might have more areas of expertise represented, which could explain why those teams’ research appeared more novel. There was some truth to that story, they found: mixed-gender teams “do tend to have more diversity of expertise,” Uzzi says. “But you still get a separate bump in novelty just from having a mixed-gender team.”

The benefits of mixed-gender teams also remained when they looked at the size of the authors’ professional networks and the geographic diversity of the team, as well as the areas of research represented, such as cardiology or neurology.

In addition, the researchers conducted a preliminary investigation of team gender makeup in other areas of science. They looked at over 20 million more papers in the 18 disciplines in science besides medicine and found that the same pattern extended to papers published in all scientific fields—not just medicine—over the last 20 years.

Gender Diversity Is Better for Creativity

While the research didn’t directly address the question of why mixed-gender teams outperform same-gender teams, Uzzi has a general hypothesis. “We think that gender affects the process by which scientists generate ideas and then select the best ideas to follow,” he says. In other words, perhaps the exchange of ideas is more lively, creative, and constructive in a gender-diverse group.

Whatever the exact reason for the benefit, Uzzi says he’s trying to use it to the fullest in his own research. In his lab, for example, he’s careful to make sure research teams are gender diverse. “I do feel like it leads to better outcomes,” he says, “and I think it helps to make the process more generative and fun.”

Mixed-Gender Teams Are More Common—But Still Underrepresented

Yet, not everyone is doing this. The researchers found that while the percentage of mixed-gender teams is growing—between 2000 and 2019, the percentage of four-person teams that were mixed-gender grew from 60 to 70 percent—there are still fewer than would be expected if teams were assembled without regard for gender.

To quantify this, the researchers developed a theoretical model that created teams randomly, while holding constant factors such as the medical subfield represented, the team’s geographic diversity, and authors’ number of citations. This model revealed that mixed-gender teams are underrepresented by as much as 17 percent.

“The benefits of gender diversity are kind of hidden,” Uzzi says, “and because they’re hidden, they’ve been underutilized. People are not taking advantage of this potential approach to doing better science.”