Richard Sander
New York Daily News, January 30, 2022
Monday, the Supreme Court announced that in the 2022-23 term it will hear challenges to the use of race in college admissions at Harvard and the University of North Carolina. The court has thus decided to wade into another highly controversial question — like abortion and gun control — about how exactly the Constitution should protect individual rights. Unlike these other divisive social problems, however, the issue of race-based affirmative action is not, I believe, mainly a question of conflicting values. Rather, most disagreement stems from confusion about the underlying facts.
For example, public opinion polls consistently show that most Americans support “affirmative action” in higher education to create opportunities for racial minorities — but large majorities also oppose the use of “racial preferences” by colleges. The likely explanation of this difference is that most people support better outreach and better services to identify promising disadvantaged applicants, and efforts to make sure that admissions processes are fair and accessible. But they oppose “reverse discrimination” against white and Asian-American applicants, or the use of “double standards” in admissions.
So which perception is accurate? Both are, to a degree. At many community colleges, second-tier state universities and regional private colleges, admissions are either not competitive or simply require that students meet some specified qualification (for example, having at least a 3.0 high school GPA). So affirmative action at these local schools is largely a matter of outreach and remediation. At elite colleges and universities, in contrast, racial preferences are usually large and often huge. At Harvard, a student whose overall academic credentials put her in the middle of the pack of admitted students has only a 25% chance of being admitted if she is Asian-American, but a 74% chance of admission if she is Hispanic and a 94% chance of admission if she is Black, according to an analysis by economist Peter Arcidiacono, the plaintiff’s expert. At the University of North Carolina, an out-of-state applicant whose credentials are at the level of a typical UNC student has a 25% chance of admission if he is Asian-American, but a 99% chance of admission if he is Black. The admissions double standards not only exist; they are so large that they create a big academic gulf, equivalent to hundreds of SAT points, between students of different skin colors.
The academic gulf in admissions standards then leads to other problems. One of the most serious of these is “academic mismatch,” which occurs when a student finds herself at a university where most of the other students have substantially higher qualifications. A mismatched student who is interested in a science or technical (“STEM”) field has a dramatically lower chance of graduating with a STEM degree than she would experience at a less-elite university. A mismatched law student is more than twice as likely to fail the bar exam as the same student would be at a less elite law school. Other examples abound, and have been documented in over a dozen peer-reviewed studies.
The “mismatch effect,” contrary to some claims, is definitely not a consequence of a student’s race. Highly qualified Black students admitted without preferences (or with only a small preference) do fine, and white students admitted with a large preference (for example, some athletes or legacy preference beneficiaries) run into the same academic problems as students admitted with a large affirmative action preference.
Large racial preferences also cause a serious “social mismatch” problem. A careful 2013 study at Duke University, which uses large racial preferences similar to those at Harvard and UNC, found that entering freshmen initially formed friendships that reflected the diversity of the freshman class, but that friendships became more stratified over time along lines of academic performance. The “A” students tend to hang out with other “A” students, the “B” students with other “B” students, and so on. Since large racial preferences create large racial differences in college performance, the high-performing students tend to be predominantly white and Asian-American; the low-performing students tend to be Black or, to a lesser degree, Hispanic. This academic stratification by race thus precipitates a social stratification. The Duke study showed that these social effects are driven by student performance, not by race: High-performing Black students continued to have very racially diverse friendship networks, but struggling Black students were increasingly segregated in predominantly minority networks. Other national data, and a controlled experiment at the Air Force Academy in Colorado produced similar results. Large preferences produce counterproductive social effects.
From this perspective, the demands of many Black students at elite college campuses to have segregated dormitories, separate graduations, “safe” spaces on campus, and ethnically focused majors are less puzzling. College administrators don’t tell students that they are using large racial preferences, but Black students will notice, as soon as grades come out during freshman year, that a disproportionate number of their Black friends are in academic difficulty. In the absence of any other information, it is not surprising that they conclude that the campus is “hostile” to Black success and is putting them in harm’s way. In a sense, that is exactly what college administrators are doing.
Both socially and academically, then, racial preferences in admissions tend to do the opposite of what their proponents claim: They reduce the diversity of college interactions, and they make it more difficult for minorities to be academically successful.
Another empirical reality is that few genuinely disadvantaged students receive admissions preferences. Higher education leaders speak endlessly about diversity, but what most of them mean is only racial diversity. Across all elite colleges, only 3% of undergraduates come from families whose incomes put them in the bottom quartile of the American income distribution; over 70% come from families in the top quartile. In other words, these colleges do a lousy job of fostering social mobility; Black students targeted by these programs are often upper-middle-class.
There is a better way. When California voters passed a 1996 measure that banned the use of racial preferences in state institutions, the University of California (“UC”), where I work, launched a large-scale effort to improve local pipelines from high school to college. The system invested tens of millions of dollars in counseling programs, formed partnerships between college campuses and inner-city schools, and made sure that many more low-income students understood exactly what they needed to do in high school to qualify for UC admission. The university also instituted mild socioeconomic preferences — large enough to increase the admissible pool of students, but small enough to avoid mismatch problems.
The results were stunning and, because California has such large minority populations, the benefits were disproportionately reaped by minorities. Soon, a third of UC students (compared to a sixth at most flagship state universities) were low- or moderate-income Pell grant recipients. Without racial preferences, minority enrollments rose only gradually, but those enrolling had far higher success rates. The number of Blacks earning bachelor degrees rose 70% after race-neutrality, and the number of Blacks earning STEM degrees nearly tripled. For Hispanics, the number of bachelor’s degrees quadrupled and bachelor’s degrees in STEM fields quintupled.
If the University of California and Black and Hispanic Californians were so much better off without racial preferences, why do other colleges and universities continue to use them? One reason is the “first-mover” problem. In private conversation, more than one elite university president has acknowledged to me that his or her school has a serious “science mismatch” problem. But if we acknowledge the problem, they say, and our competitors don’t, then all of our Black recruits will turn us down and go to the other schools.
The other, equally paradoxical problem is that it has become very dangerous for anyone in an academic environment to speak openly about the realities of university preference systems and their consequences. Last spring, a well-meaning Georgetown law professor was caught on tape discussing with a colleague the problem of Black students disproportionately receiving low grades. She was immediately pressured into resigning, and her colleague was suspended, apparently for the crime of listening! A distinguished medical school professor at the University of Pittsburgh published, in a peer-reviewed journal, an article discussing affirmative action in medical schools and the resulting “mismatch” problem. A storm erupted that led the journal to retract the article and Pitt to strip the professor of his administrative duties. Nothing in his article was inaccurate; his offense lay simply in discussing uncomfortable truths.
Over the 20 years I have worked on this issue, I have observed a dramatic retrenchment in once-customary policies of sharing anonymized data with researchers on student outcomes. Administrators literally fear transparency.
Higher education, in other words, has become trapped within the bizarre realities of its system of racial preferences. Aside from the most extreme ideologues, few people who are familiar with the facts (and not subject to administrative pressures) think this system makes any sense. Many, many academics secretly hope that the Supreme Court will abandon its past practice of deferring to higher education leaders on the matter and reassert the right of college students of all races to be evaluated and treated without regard to the color of their skin.
The court has a choice. So does the country.
Affirmative Action Debate 