Data Mining Points Path to Helping Female and Minority Scientists

Jennifer Waddell

It’s not what you know. It’s whom you know.

That’s not a typical adage of academe, but it may be a lot closer to the truth than many faculty members might like to admit. And accepting it might be a crucial insight for overcoming some chronic obstacles to career advancement facing women and minorities in the sciences.

Advocates of the concept include Griffin M. Weber, an assistant professor of medicine at Harvard University who has created an academic version of “six degrees of separation” to map out how closely one author of a scientific study is connected to other successful journal authors.

Dr. Weber’s fundamental finding is that such a measure of personal “reach” is one of the best predictors of whether a researcher stays at a university and gets promoted. And unlike more-traditional measures such as grant-application success or publication rates, it’s a relatively easy way to direct needed help.

“It isn’t necessarily more important” than grants or journal-article counts, Dr. Weber said. “But it is an important component of predicting a person’s career that people really haven’t looked at before.”

That insight offers help to someone like Joan Y. Reede, who has spent 25 years leading minority-faculty development at Harvard Medical School. Dr. Reede said she had long regarded the lack of quality personal connections as a key obstacle facing women and minorities in science, but felt she needed more empirical evidence to gain political acceptance of that reality as a basis for action.

Dr. Reede’s own data tell her that a classroom instructor in the top quartile, as measured by academic reach, is three times as likely as one in the bottom quartile to be promoted to assistant professor. And an assistant professor in the top quartile is twice as likely to be promoted to associate professor, she said.

Without more evidence to validate and explore that finding, however, efforts to help young researchers often founder for lack of direction and institutional support, she said.

“It is really a huge change in our ability to understand the experiences of individuals, to intervene, to look at outcomes, to reassess and revise,” Dr. Reede said of Dr. Weber’s work, which covers more than 5,000 Harvard Medical School faculty members. “We haven’t had this ability in the past.”

A Virtual Map

Beyond Harvard, such analyses appear to be rare. One 2011 study, at the University of Washington’s School of Medicine and involving American Indian and Alaska Native scientists, also affirmed the benefit of encouraging collaborative relationships. But it was based on only 29 trainees and mentors.

For the most part, said one author of that study, Dedra Buchwald, then a professor of epidemiology at Washington and now a professor of medicine at Washington State University, the process of helping minority faculty members in career advancement entails a much more ambitious program of training meetings every two months on the campus. And convening those meetings can be costly because such faculty members, rare at most institutions, need to travel to Seattle from several universities, she said.

Dr. Buchwald said she does wonder about the practical value of a strategy for advancing female and minority faculty members that’s based largely on arranging and encouraging collaboration on research projects. That may work within a single large institution such as Harvard, she said, because that probably would mean that faculty members seeking promotion become better known to the people voting on their candidacies. That benefit may not be as clear-cut when project collaborators are at other institutions, she said.

Meaningful collaboration also requires more than arranged partnerships. A study published last month by researchers at the Massachusetts Institute of Technology found that women are more likely than men to abandon engineering career tracks, and it attributed that problem to negative dynamics in research teams that tend to discourage women from participating.

Both universities and their funding agencies are under pressure to figure it out. Those that have been seeking solutions include the National Institutes of Health, the largest single provider to universities of basic-research money. The NIH has found that its black grant applicants have a 35 percent lower chance of success than its white applicants.

Dr. Weber’s project involves creating a virtual map of which researchers have been co-authors with other researchers, and then using that information to identify what potential future authorship collaborations would give a researcher the most value in terms of expanding his or her network.

Testing it on himself, Dr. Weber identified another Harvard researcher whose network would give him the greatest benefit in terms of filling holes in his own network of personal reach. It may not make sense to propose a research project with that faculty member just to build his network, Dr. Weber said, but in some cases that type of information might contribute to a decision on whom to approach for a future collaboration.

Dr. Weber also acknowledged that his findings do not prove a direct cause and effect. “We do not yet know whether having a larger co-author network influences promotion votes, or having qualities that get a person promoted — for instance, grants or highly cited articles — attracts new collaborators,” he said.

Future solutions based on the metrics of network reach probably will need to emphasize balance, Dr. Weber said. The idea will be to assemble teams in which members gain from the expansion of their networks, he said, without stretching people so far that they can’t work together productively.

Paul Basken covers university research and its intersection with government policy. He can be found on Twitter @pbasken, or reached by email at paul.basken@chronicle.com.

Correction (6/29/2016, 9:30 a.m.): This article originally cited Dedra Buchwald’s professional affiliation at the time of the research. It has been updated to add her current affiliation.