The University of Pennsylvania has promised $100 million to increase faculty diversity. Yale has just pledged $50 million, Johns Hopkins $25 million. These sums are large, but the goal has been an object of urgent attention for decades. University administrators and faculty have long been dedicated to increasing the numbers of blacks, Latinos, and women, among others, in their teaching ranks.
But despite their intentions and efforts, the desired degree of diversity has not materialized. The reason is there is a pipeline problem. For example, in 2014, black students earned just 1.8 percent of doctoral degrees in the physical sciences. Even if every physics department in the nation were to recruit black Ph.D.s, there wouldn’t a big enough pool to effect much statistical change.
Julie Posselt, an assistant professor of higher education at the University of Michigan, wishes to increase faculty diversity by expanding that pipeline and thinks that this pipeline problem has much to do with how Ph.D. students are selected. In her new book, Inside Graduate Admissions: Merit, Diversity, and Faculty Gatekeeping, she writes that faculty “in highly ranked doctoral programs rely on admissions criteria that undermine their diversity and equity aims.” Posselt set out to find out why by “observing and interviewing graduate admissions committees in astrophysics, biology, classics, economics, linguistics, philosophy, physics, political science, and sociology”—ten departments at three unnamed high-powered research universities.
Posselt exposes the many ways in which professors, their occupational attachment to deliberation notwithstanding, don’t deliberate very well. To start with, their time is limited. Facing hundreds of applications, they seek simple ways to discard as many of them as possible. Posselt found that several committees use a GRE (Graduate Record Examination) cutoff to eliminate applications without otherwise considering them. This culling contradicts the advice of the Educational Testing Service that creates and administers the GRE.
As we’ll see later, Posselt may be too quick to dismiss the predictive power of GRE’s. But there is little question that the faculty members Posselt observed and interviewed, whether because of time constraints, or because of unrealistic assumptions about what small differences in GRE scores can tell them, overemphasize those scores.
But at least GRE’s have some connection to students’ actual intellectual abilities. Often, even when they have a short list to work with and can conduct a “holistic” review, faculty members rely on ill-defined standards like “grit,” “brilliance,” and even “coolness” to distinguish candidates from each other. Of course, lacking a mathematical or scientific solution to the question of who the best candidate is, faculty are bound to fall back on standards about which reasonable people disagree. That is, they have to exercise their judgments.
There is nothing intrinsically wrong with that, but Posselt catches professors grasping at any handy straw. In one such situation, two classics candidates receive special consideration in part because they hail from rural New England. One committee member “envisioned it as a ‘pastoral’ place for early socialization in classics.” In another case, a political science candidate “with mediocre grades was admitted at least in part because a committee member thought it was ‘cool’ that she wrote for an online magazine.” Her gain was the loss of “another borderline applicant who had written a book.” Candidates are made or unmade by a “single ambiguous line in a letter of recommendation” or even a “weekend hobby.”
In this admissions Wild West, decisions will sometimes favor members of groups that diversity advocates wish to recruit, as grit or spark or an interesting story can be drawn on to promote a candidate who otherwise might not pass muster with most committee members. But Posselt suggests it frequently works against them, as faculty members, like other human beings, tend to favor people who resemble them, and, other things equal, committees might overlook good candidates who do no not resemble them. Even in the absence of that rationale, I see no reason to discount the studies that suggest even our progressive faculties exhibit bias against certain groups under some circumstances.
Posselt includes among these groups that face graduate admissions bias not only the usual suspects but also candidates from religious colleges (“I would like to beat that college out of her,” says a committee chair) and, perhaps most strikingly candidates from East and South Asia, who faced a higher bar because they were thought to be better at taking exams. Just as Posselt says, it seems hypocritical that while “many expressed worries that test preparation rendered Chinese students’ scores suspect, not once did a respondent express similar concern about formal test preparation courses convoluting the meaning of a wealthy American student’s score.”
However, Posselt doesn’t distinguish enough between standards involving judgment that seem irrelevant, like “coolness,” and standards that seem quite relevant, like prior research experience and prior coursework in the discipline. With respect to the latter, Posselt asks suspiciously, “What is the ‘right preparation’? What would it have looked like for these…applicants to be ‘up to the bar’”? As if it is arbitrary for a linguistics admissions committee to favor applicants from “great, active linguistics departments where they probably had taken two years’ worth of graduate coursework already.”
Similarly, Posselt views with skepticism professors who are very confident that a student who has perfect G.R.E. scores and perfect grades is uncommonly intelligent. Because Posselt is interested in the way in which the committees she studied were “constructing symbolic boundaries that guided their assessments of fit and belonging,” she suspects every standard of being a questionable means of distinguishing “us” from “them.”
But selecting on the basis of prior research and course experience is not like selecting on the basis of, as Dr. Seuss would have it, butter side up or butter side down. Some standards probably predict student success, others are probably meaningless.
But this is where reality starts to encroach on Posselt’s version of fairness. Admittedly, as Posselt says, the standard of graduating from a rigorous, highly rated program in the subject matter “presumes a level of socialization” into an academic research community that “privileged applicants” are more likely to have. Using it may reinforce existing inequalities.
But so may requiring a college degree or any other such obviously sensible measure. Unless our sole purpose is to achieve racial and gender balance, we have to be able to distinguish between reasonable and unreasonable standards for admission, rather than assuming that all standards are the equivalent of the signs rival gangs use to distinguish themselves from each other. To be fair, Posselt does not take that view. But she leans toward it.
Let one more example suffice to illustrate her perspective. Posselt reports on a political science admissions committee meeting. The chair, a political philosopher, argues that GRE scores should be discounted, as some studies suggest that such scores predict only first year grades. Three professors with statistical expertise respond that these studies are flawed because of the samples they use. The chair loses the argument. Posselt, strangely, reads this result as “methodological majoritarianism” because the chair is a political theorist and his opponents are quantitative researchers. That makes the outcome seem arbitrary. The statistics gang won. But the chair wasn’t making a “political philosophy” argument. He was making an argument that depended on statistical knowledge, and he therefore lost on the very grounds upon which he attempted to stand to those with superior statistical expertise.
Notably, Posselt cites the kind of studies in question without responding to the argument that prevailed in the meeting. That may be in part because Posselt has written a fighting book. She thinks that graduate admissions committees, although they value diversity as one standard among others, need to value it still more. Yet she provides ample evidence to the contrary, suggesting that university administrations provide strong incentive to admit “diversity” applicants. And that even the quantitatively oriented political science department just admitted a student who qualified for a diversity fellowship, even though he had an abysmal quantitative GRE score. Or that a push from even one committee member concerned about diversity may be enough to get a candidate in, especially if someone agrees to mentor the student.
Additionally, Posselt’s intention to expand the pipeline of minority applicants to faculty jobs through graduate admissions is almost doomed to failure without looking beyond that single step. It is likely that graduate admissions committees are dealing with the same kind of pipeline problem that hiring committees face. Yes, women earn only about 20 percent of the Ph.D.s awarded in physics but they also earn only about 20 percent of the undergraduate degrees awarded in physics. All the systemic bias Posselt thinks she sees results, at least in physics, in women being admitted in rough proportion to their actual numbers in the pool of possible applicants. To her credit, Posselt, who pays very little attention to this pipeline issue, does highlight one program that addresses it, a cooperative effort between Vanderbilt University and Fisk University that prepares candidates from “underrepresented” groups to enroll and excel in Vanderbilt’s STEM Ph.D. programs.
Posselt is right that professors who purport to stand for rational discourse could stand to do a much better job of reflecting on the grounds of the judgments they make about applicants to their graduate programs. But her book, though it is full of interesting details about the cases she studied, will convince only the already converted that graduate admissions committees need to be pushed harder on diversity.