A study of nearly 19,000 faculty in computer science, business, and history departments in the United States and Canada finds that top universities like the ones that help give Massachusetts its international renown form a kind of insular academic club, hiring most of their faculty from others in a small, elite network.
Just a quarter of the 461 academic departments studied produce the majority, somewhere between 71 to 86 percent, of all tenure-track faculty, the researchers found. The pattern was not simply explained by differences in merit, and the researchers argue that it may have "profound implications for the free exchange of ideas" in their study published this week in Science Advances.
The pattern could shape what type of research is done, what collaborations occur, and which ideas catch on easily.
The work by a trio that included a postdoctoral researcher at the Harvard School of Public Health provides a detailed quantitative look at the role prestige and hierarchy play in academia. Half of the faculty in each discipline studied are trained at a small number of institutions. In computer science, 18 institutions produce half the faculty; in history, eight schools account for half the professors; and in business, 16 institutions produce half.
In the rankings, which are based on where doctoral students end up landing faculty jobs, Harvard and the Massachusetts Institute of Technology were both ranked in the top 10 in business and computer science; Harvard was ranked first in history, Brandeis University was eighth.
"People have beliefs about how they think the system works," said Daniel Larremore, a postdoc at Harvard who specializes in studying networks and devotes most of his attention these days to studying the spread of the malaria parasite. "Our study provides hard data to substantiate and then unpack the nuances of this market."
Among the findings:
■ Between 68 and 88 percent of the faculty at the most elite institutions also earned their graduate degrees at those elite institutions, classified as the top 15 percent.
■ Most people will slip down the slope, moving to a faculty job at an institution that is less highly ranked than the one where they received their graduate degree. This was true overall, but the hierarchy was even steeper for women in the fields of computer science and business.
■ Because of this slippage, a professor will on average train two to four times fewer new faculty than their own mentors did.
The researchers found that it is implausible that the disparity is explained solely by differences in the merit of people graduating from the institutions; if the hiring pattern reflected differences in the quality, it would mean that faculty who earned their doctoral degrees from the top 10 departments were two to six times more productive than faculty in those top 21 to 30 departments.
Larremore noted that this was an incomplete picture of the productivity of any given program, since people who earn PhDs often go on to other jobs outside academia. He also noted that people defy the odds and rise from lower-ranked universities to the top. Understanding how and why those people rise could be a fascinating subject for future study.
He noted that his own experience provides an interesting perspective. He earned his doctoral degree in applied math from the University of Colorado at Boulder, not one of the most prestigious institutions in his field. Now, he has a fellowship at Harvard. He has yet to go on the job market to see whether his own anecdotal experience compares to the data set.
"I guess what I'd say is if I had seen this study before going into graduate school or before looking for fellowships, I don't think it would have deterred me. Knowing how steep the mountain is wouldn't have deterred me from trying to climb it," Larremore said. "But for many people, knowing how steep it is could help make more informed decisions."