Faculty hiring and retention: The influence of prestige

The rounds of interviews conducted by American universities that lead to new academic hires are less important than the reputation of the university at which candidates took their PhDs, suggests a new study published in the journal Nature in October.

The study shows that the prestige of a small group of American universities leads to a self-perpetuating hierarchy: a small minority of universities supply a large majority of faculty across a wide range of fields, and graduates of these schools who do not get jobs in the higher prestige universities move to faculty jobs in lower prestige universities.

“Quantifying hierarchy and dynamics in US faculty hiring and retention” also shows that with the exception of those educated in Canada or Britain, foreign-trained professors leave the profession at a much higher rate than do American-trained professors.

Additionally, it found that while the percentage of women professors in the STEM (science, technology, engineering and mathematics) fields has gone up over the past decade, that increase is due less to recent increases in hiring of women professors than it is to the retirement of mostly male professors.

The study, conducted by a team led by Professor Daniel B Larremore and graduate student K Hunter Wapman, both from the department of computer science and the BioFrontiers Institute at the University of Colorado Boulder, involved almost 300,000 faculty members in 10,612 departments in 368 PhD-granting American universities.

“We analysed the academic employment and doctoral education of tenure-track faculty at all PhD-granting US universities over the decade 2011-2020, with the goal of understanding the factors that shape hiring and attrition in the US professoriate,” said Larremore and Wapman.

“Our analysis shows universal inequalities in which a small minority of universities supply a large majority of faculty across fields – inequalities which are created during the hiring but exacerbated by attrition, and reflect steep hierarchies of prestige.

“Our results also indicate that gains in women’s representation over this decade result from demographic turnover and earlier changes made to hiring, and are unlikely to lead to long-term gender parity in most fields,” says Larremore and Wapman.

According to Antonio Duran, professor of higher and post-secondary education in the division of educational leadership and innovation at Arizona State University, Larremore’s study of the composition of the US professoriate “certainly shapes the experiences that our college students have on campus, the outcomes they report and the ideas that they wrestle with”.

“Therefore, it is critically important that individuals understand patterns in faculty hiring, especially when taking into account social identities and the systemic barriers associated with these identities.”

How prestige structures faculty

Building on the findings reported in Larremore’s 2015 article “Systemic inequality and hierarchy in faculty hiring networks”, this study shows how ‘prestige’ structures American university faculty.

In faculty hiring, perceived prestige can be seen in the fact that 80% of faculty members were graduates of just over 20% of the nation’s universities. Even more striking is the fact that 13.8% (one in eight) went to a handful of elite universities: University of California (UC) Berkeley, Harvard University, University of Michigan, University of Wisconsin-Madison and Stanford University.

The next grouping, made up of 13 universities, trains another 20% of graduate students who go on to be professors.

To show how iniquitous this is, Wapman plotted these figures using the Gini coefficient. A measure of inequality developed by economists, the Gini coefficient is plotted on a scale ranging from zero to one: the higher the number, the greater the inequality. For example, the Gini coefficient for Brazil’s income inequality is 0.48, while the figure for the United States is 0.42. The Gini coefficient for Canada is 0.33; Ukraine’s is 0.26.

Using the Gini coefficient, Wapman’s charts showed that inequality in “faculty production” was 0.75, which is 0.12 points higher than the figure economists find for South Africa (0.63), the country the Gini coefficient rates as the most inequitable.

At 0.67, inequality is lowest in education, and medicine and health, likely because many universities across the country have education, medical and nursing faculties. Engineering, by contrast, has an overall Gini coefficient of 0.73.

“These large Gini coefficients, alongside our network-based hiring analyses, show that prestige plays a large role in shaping hiring, with most professors showing ‘downward mobility’ by moving from higher prestige PhDs to lower prestige faculty jobs; very few faculty show upward mobility.”

Social closure

While Larremore’s data didn’t exactly find the “old boy’s network”, it allowed his team to construct models that showed that faculty hiring networks linked the Massachusetts Institute of Technology, Carnegie Mellon University, UC Berkeley, Cornell, Harvard, Stanford, Yale and Washington University far more than might have been expected given chance alone.

According to Larremore, treating flows of faculty between US universities as a network leads to a natural, recursive definition of prestige: a department is prestigious if its graduates are hired by other prestigious departments. Further, since these universities train the bulk of the professors who teach at the nation’s other universities, the methods, ethos and practices of these universities become, by default, the dominant ones in America’s universities.

A further factor that contributes to “social closure” in the American professoriate is the phenomenon of self-hiring, ie, universities hiring their own graduates. In comparison to Russia, South Africa and Portugal, where self-hires account for 36%, 67% and 73%, respectively, writes Larremore, in the US self-hires account for 11% of US-trained faculty.

This figure is statistically far greater than expected, reflecting a marked deviation from what is otherwise prestige-driven hiring.

The rates of self-hiring were not, however, the same across all fields. It was lowest in the humanities (4.5%) and social sciences (6%). By contrast, in medicine and health, it was 16.7%.

“These differences indicate that self-hiring practices vary by domain, though we don’t know why. We hope to investigate this in future work,” says Wapman.

One of the surprising findings of this study is that self-hires have a higher attrition rate than do non-self-hires. The rate of attrition for self-hires is 1.2 times the rate for other faculty. Nursing experienced the lowest rate (0.9-fold), while the highest rate was found in criminal justice and criminology (1.9-fold), and industrial engineering (1.8-fold).

Larremore’s study, Duran stressed in his e-mail, “unearths the unsaid, but always present – that those seeking careers in the professoriate are always contending with many factors out of their control,” including the perceived prestige of the university where they were trained and gendered hiring structures.

Hiring of women faculty is flatlining

At first glance, Larremore’s numbers appear to support the belief that the situation vis-à-vis women’s employment as professors in the academy has steadily improved. While male professors made up 64% of professors on the tenure track, in the decade beginning in 2011, the share of professorships held by women increased by about a third, 12.5% to 17.1%.

In all eight broad areas of study and in 80 of 107 individual fields, women’s representation increased. The lone field where women’s representation dropped was nursing, traditionally the field where women were vastly over-represented in the professoriate. (The eight domains are applied sciences, education, engineering, humanities, mathematics and computing, medicine and health, natural sciences, and the social sciences.)

In analysing these numbers, Larremore’s team asked whether the decade-long increase of women in the professoriate was a function of simple demographics. In other words, did their numbers record who was entering or leaving (retiring from) the profession?

To their great surprise, they found that, when comparing 2011 and 2020, the proportion of women among newly hired faculty in 100 of the 107 fields did not change significantly. In seven fields, including gender studies and nursing, it actually decreased.

“We then compared inflows of new hires with outflows of departing faculty over our decade of observation,” says Larremore. “We found that in all eight domains, and 103 of 107 fields, newly hired faculty were substantially more likely to be women than their departing counterparts.

“However, this pattern in all-cause attrition is driven by dramatic demographic turnover, with retirement-age faculty skewing heavily towards men. The implication of this is that the overall increases in women’s representation over this period are primarily due to changes in faculty hiring that predate the decade we studied.

“If you compare it to the 1950s, yes, women’s representation has improved greatly. But if you look at new hires in medicine and health in 2012 it was about 52% and now it’s about 54%. In other words, the hiring of women by universities has essentially flatlined over the past decade. It’s unlikely to improve, especially in STEM fields, without further changes to women’s representation among new faculty and procedures.”

For his part, Duran drew attention to how Larremore and Wapman’s work revealed that while “women faculty are increasing, patterns still point to their continued under-representation in the academy”.

Loss of foreign-trained faculty

My interview with Larremore and Wapman began with me asking about the reason the rate (38%) for foreign-trained professors leaving their professions was so much higher than the rate (31%) for professors educated in the United States, Canada or Britain.

One reason, Larremore hypothesised, was positive: professors from developing countries returning to their counties of origin with both knowledge and valuable teaching experience.

Another was personal. “Anecdotally, we’ve heard of others who have studied foreign-born and -trained faculty and they have said that if you’re from another country and you come to the US to be a professor, there may be some pressures or desires to go back to live in a place where everybody speaks your native language,” he says.

Difficulties in gaining purchase on the steep prestige hierarchy that is the American professoriate is one more reason foreign-trained professors resign from their positions at a much higher rate than do their American, Canadian or British-born and -trained peers.

“Eighty-nine percent of US professors,” Larremore said at the end of our interview, “got trained within the US; only 11% come from outside the US. So, the system isn’t closed. But it’s definitely not porous.

“I know from talking to colleagues that the Italian higher education system is very different. There, people are interested in living near where their families have lived for generations, where they speak the dialect. And so moving far away is sort of less interesting for them. Accordingly, Italy is less of a big market and much more geographically organised, whereas the US system is geographically flat, but very prestige oriented.”