Studies based on large data sets have an obvious limitation: they can only demonstrate that diversity is correlated with better performance; they cannot demonstrate that diversity causes better performance. However, research on racial diversity in small groups allows some causal conclusions to be drawn. Again, the findings are clear: diversity benefits groups that value innovation and new ideas.
In 2006, I collaborated with Margaret Neale of Stanford University and Gregory Northcraft of the University of Illinois at Urbana-Champaign to investigate the impact of racial diversity on small decision-making groups in an experiment in which information sharing was required for success.
Our subjects were University of Illinois undergraduate students taking business classes. We formed three-person groups, some with all white members, others with two whites and one nonwhite member, and had them participate in a murder mystery exercise. We made certain that all group members had access to the same information, but we also provided each member with important clues that only he or she was aware of. To figure out who committed the murder, the group members would have to share all of the information they had during the discussion. Groups with racial diversity outperformed groups with no racial diversity significantly. Being with people who are similar to us leads us to believe that we all have the same information and the same point of view. This perspective, which prevented the all-white groups from processing information effectively, is what stifles creativity and innovation.
Similar findings have been made by other researchers. In 2004, Anthony Lising Antonio, a Stanford Graduate School of Education professor, collaborated with five colleagues from the University of California, Los Angeles, and other institutions to investigate the impact of racial and opinion composition in small group discussions. The study included over 350 students from three universities. For 15 minutes, group members were asked to discuss a current social issue (either child labour practices or the death penalty). The researchers drafted dissenting opinions and had both black and white participants deliver them to their respective groups. When a black person presented a dissenting viewpoint to a group of whites, the viewpoint was perceived as more novel, leading to broader thinking and consideration of alternatives than when the same dissenting viewpoint was presented by a white person.
The lesson is that when we hear opposing opinions from someone who is not like us, it causes us to think more deeply than when it comes from someone who looks like us. A longitudinal study published last year tracked the moral development of students on 17 campuses who took a diversity class in their freshman year. The analysis led the researchers to a strong conclusion: students who were trained from the start to negotiate diversity demonstrated much more sophisticated moral reasoning by the time they graduated. This was especially true for students with lower academic ability.
This effect transcends race and gender. Last year, for example, management professors Denise Lewin Loyd of the University of Illinois, Cynthia Wang of Oklahoma State University, and Robert B. Lount, Jr. of Ohio State University and I asked 186 people if they identified as a Democrat or a Republican, then had them read a murder mystery and decide who they thought committed the crime. The subjects were then asked to prepare for a meeting with another group member by writing an essay expressing their point of view. More importantly, in every case, we informed the participants that their partner disagreed with them but that they would need to reach an agreement with the other person. Everyone was told to prepare to persuade their meeting partner to come around to their side; however, half of the subjects were told to prepare to persuade a member of the opposing political party, and the other half were told to persuade a member of their own party.
As a result, Democrats who were told a fellow Democrat disagreed with them prepared for the debate less well than Democrats who were told a Republican disagreed with them. Republicans followed a similar pattern. When we disagree with someone from a different social group, we are motivated to work harder. In ways that homogeneity does not, diversity shocks us into action. As a result, diversity appears to result in higher-quality scientific research.
In 2014, two Harvard University researchers used Thomson Reuters' Web of Science, a comprehensive database of published research, to examine the ethnic identity of the authors of 1.5 million scientific papers written between 1985 and 2008. They discovered that papers written by people from different ethnic groups receive more citations and have higher impact factors than those from the same ethnic group. Furthermore, they discovered that stronger papers were associated with a greater number of author addresses; geographical diversity, as well as a greater number of references, reflects greater intellectual diversity.
This essay was first published in Scientific American in 2014. Katherine Phillips revised and updated it in 2017 to include new research.