It’s not rare to see college students team up with people of the same ethnicity to work on projects. However, recent research from Harvard University indicates that working with people like oneself may weaken the quality of science research.
Richard Freeman and Wei Huang of Harvard University announced the discovery in the paper “Collaborating With People Like Me: Ethnic Co-Authorship Within the U.S.,” released in February as a National Bureau of Economic Research working paper.
Researchers analyzing citations and journal publications determined that working with people from more diversified ethnic backgrounds enhances the quality of scientific publication, while “homophily is associated with publication in a lower-impact-factor journal” and fewer citations, regardless of the authors’ previous publishing performance.
The results of the two years of research were released in 2013. “We tested the results continuously,” Huang said, “and the result turned out to be stable in different sampling and data.”
The study examined the ethnic identity of the authors of more than 1.5 million scientific papers written solely within the U.S. from 1985 to 2008. Freeman and Huang created a list of co-authors who had U.S. addresses based on the Thomson-Reuters Web of Science database for those years.
From the mid-1970s to the 2000s, the proportion of U.S.-based authors with European surnames fell while the proportion of those with names originating in non-European regions rose. This indicates that the globalization of science increased diversity among scientists and engineers in the U.S. Today, more and more names that indicate origins in Asia, Africa and Latin America appear on the papers published in the U.S.
Freeman concluded the ideas shared between co-authors from different ethnic background helped produce better publications by providing perspectives. According to him, writing with someone from a different background is almost like working with somebody from outside of your comfort zone.
“So there is an economic principle here, we say it’s more costly to produce a paper with somebody you are far away from, with different backgrounds and harder communication levels,” Freeman said. “Therefore, they are going to do that only if you can write a better paper.”
When in was published, the paper drew a good deal of attention and media coverage. However, it also initiated discussion and debate. Some people argued that citation doesn’t necessarily indicate the quality of science.
Freeman said citation is affected by two things: the quality of the paper and the number of people who appreciate the work and cite it.
He pointed out that a diverse team of co-authors also helps the paper circulate more widely. “Having a great paper gets widely circulated because it gets a lot of people from different backgrounds, different countries, different universities involved. That’s positive, not negative, ” Freeman said.
Huang agreed with this view, saying that citation is the most widely used indicator and the most effective one compared to others. “Citation maybe is not the perfect measure of the paper quality,” Huang said, “but citation and quality are highly related.”
Some critics also argue that cognitive diversity may not really reflect the true intellectual diversity.
To address this point, Freeman said in a phone interview that ethnic diversity is an indicator of diversity of thought. “If you were assigned within a company, you want to have teams of people with different views, different perspectives, and the ethnicity is an indicator,” Freeman said.
“You are more likely to have diversity of … ideas showing up in these papers in the work as opposed having everybody of exactly the same backgrounds,” Freeman said.
“We want a better-functioning economy,” Freeman said, “and this is trying to understand how science works.”
Freeman and Huang plan to continue their research in related areas, including collaborations between males and females, older and younger people and the stability within a diversified research team.
Xin Sheng is an editorial intern at Diversity Executive. She can be reached at firstname.lastname@example.org.