The Columbia Deal

On July 23rd, Columbia University signed an agreement with the federal government that covers a broad range of issues including hiring and admissions practices, disciplinary procedures, curricular oversight, and international applications. I’ll discuss these in detail in a moment (with a focus on admissions) but first want to make a couple of broader points.

I have felt for some time now that there are two competing views on higher education in the White House, and the fate of American universities hinges on which one eventually prevails.

The view associated most closely with JD Vance would like to see universities “taken over or destroyed” in order that conservatism in America can thrive.¹ The endgame from this perspective is the installation of administration loyalists at the helm, the shuttering of departments considered to be ideologically captured, political litmus tests for admissions and hiring, and the phasing out of tenure. In my opinion, such heavy-handed interference in internal governance would lead to the end of America’s global dominance in higher education and research.

The second view is more pragmatic and based on the understanding that universities deliver a service that is in strong global demand, help prevent our trade deficits from being even larger, fuel innovation in the private sector, and contribute significantly to productivity growth. Those who hold this view are not interested in killing off the golden goose, as long as institutions are prepared to make significant changes to their operations.

A great deal will depend on whether Columbia and the resolution monitor agree on the interpretation of such terms as merit, but at first glance it seems to me that the deal reflects the more pragmatic view. This keeps alive the possibility that American universities will maintain their perch atop the global higher education hierarchy. My guess is that most other elite institutions (including Harvard, despite having filed suit against the federal government) would accept something resembling this deal if it were offered.

With that out of the way, consider the details.

From Columbia’s perspective, the following two clauses are crucial. First, the agreement “is not an admission in whole or in part by either party, and Columbia expressly denies liability regarding the United States’ allegations or findings.” And second, no provision of the agreement “shall be construed as giving the United States authority to dictate faculty hiring, University hiring, admission decisions, or the content of academic speech.” I don’t think that Columbia could have signed the agreement (or remained viable as a leading research university) had these red lines been crossed.

In exchange, the university has committed to a number of very significant steps:

• It will review its regional studies programs “to ensure the educational offerings are comprehensive and balanced.”

• It will “maintain merit-based admissions policies” and will not “unlawfully preference applicants based on race, color, or national origin in admissions.”

• It will “provide that all hiring and promotion practices for faculty and administrative roles are grounded solely in individual qualifications and academic and professional merit, and shall not use race, color, sex, or national origin as a factor—implicit or explicit—in hiring decisions.”

• It will provide “single-sex housing for women who request such housing and all-female sports, locker rooms, and showering facilities.”

• It will “add an additional administrator… who will serve as a liaison to students concerning antisemitism issues.”

• It will establish processes to provide that students are committed to “civil discourse, free inquiry, open debate, and the fundamental values of equality and respect” and “develop training materials to socialize all students to campus norms and values more broadly.”

• It will provide federal authorities “with all disciplinary actions involving student visa-holders resulting in expulsions or suspensions, and arrest records” and ensure that disciplinary actions are “determined without regard to immigration status.”

• It will “take steps to decrease financial dependence on international student enrollment.”

This is just a partial list, each element of which warrants a separate post. Here I want to focus on the admissions directive, because so much hinges on how merit is understood and how compliance with the directive is to be assessed.

The resolution monitor, chosen by mutual consent, is Bart M. Schwarz. In order to assess compliance with the resolution, Columbia will provide him with “admissions data… showing both rejected and admitted students broken down by race, color, grade point average, and performance on standardized tests, in a form permitting appropriate statistical analyses.” In addition, it will provide “non-individualized, statistical information regarding enrolled students… including the composition of the class broken down by race, color, national origin, grade point average, and performance on standardized tests.”

How might this data be used to determine whether the university is abiding by its commitment to maintain admissions policies that do not preference applicants based on race, color, or national origin?

I can think of two possibilities.

The monitor could look at the likelihood of acceptance conditional on an index of past performance (such as a weighted average of grade point average and a standardized test score). If there are substantial differences in admissions probability by race, color, or national origin at any given score, this might be interpreted as evidence of discrimination.

In addition, the monitor could look at data on enrolled students, to see whether there are differences across groups in the distribution of performance. If one group has a significantly higher level of performance than another, the monitor may conclude that it has been discriminated against, on the logic that members of this group must have faced a more demanding standard for admission.²

Both of these approaches are based on a very narrow conception of merit, and would rule out admissions policies that seek to identify diamonds in the rough (students with great potential but limited prior opportunities) and students who could make substantial contributions to society in later life but may achieve less than stellar grades while enrolled.

Consider some examples to illustrate.

When Dartmouth decided to return to a policy requiring all undergraduate applicants to submit standardized test scores, one of the reasons given was that a test-optional policy made it harder for the university to identify high achieving students from less advantaged backgrounds. Many such students were choosing not to submit scores, and were hurting their admissions prospects as a result. This was because the admissions office considered both absolute and relative scores—of two applicants with the same score, the one from a high school with a lower distribution of scores was considered to have greater latent ability.

Such an admissions policy, while defensible on grounds of merit, would fail a test requiring students of all groups to have the same probability of admission conditional on past performance. Those in groups that tend to be in schools with higher overall score distributions would appear to be discriminated against, even though the policy does not condition acceptance on group membership in any way.

As a second example, consider Roland Fryer’s argument that while his math GRE scores were not quite as high as those of his peers in graduate school, “what I endured to achieve those scores—a father in prison and a mother I had yet to meet—were important context.” Such background information is valuable in the interpretation of scores, but if it is taken into account by an admissions office, the resulting data would again fail the monitor’s test.

Finally, consider the manner in which the Reverend John Brooks selected a cohort of twenty African American men for admission to the College of Holy Cross in 1968, based on “ambition, leadership potential, and strong character rather than the right family pedigree or the right test scores.” Few if any of the members of this group would have been selected on standard criteria, and most of them struggled from time to time once enrolled. But many went on to achieve great professional success, including Clarence Thomas, Edward P. Jones, Theodore Wells, and Stanley Grayson. The data generated by this selection process would have suggested discrimination under naive tests of merit, but the problem in this case is with the tests rather than with the process.

Amartya Sen has argued that “the idea of meritocracy may have many virtues, but clarity is not one of them.” If Columbia wants to persuade the monitor that the university is abiding by the resolution, while generating data that fails naive tests of merit-based processes, they will have to develop alternative metrics that are transparent and defensible in plain language. It will not be an easy task. And the decisions made by the resolution monitor over the next couple of years may turn out to be more consequential than any made by a university president or board of trustees.

Columbia’s iconic Low Library

1. I have quoted extensively from Vance in earlier posts, and repeat here his words from a podcast conversation in September 2021: “There is no way for a conservative to accomplish our vision of society unless we're willing to strike at the heart of the beast. That's the universities… Unless we're willing to de-institutionalize the left in those institutions—or destroy the institutions absent that—we are going to continue to make the most powerful academic actors in our society actively aligned against us. The only way to work is to actually take some of these institutions over.”

2. This is basically the outcome test for discrimination proposed by Gary Becker. For example, if a group is discriminated against in the process of loan origination, their repayment rates will be higher, since they have to be unusually credit-worthy to get the loan in the first place.