River Dance

National Review, Nov 9, 1998 by Ramesh Ponnuru

The Shape of the River: Long-Term Consequences of Considering Race in College and University Admissions, by William G. Bowen and Derek Bok (Princeton, 472 pp., $24.95)

THE debate over racial preferences takes place on almost every conceivable level: economic, moral, constitutional. In the course of this debate, numerous empirical claims have been made. For instance, critics suggest that preferences harm even their beneficiaries by putting them in academic environments a notch or two above what they can handle. Until now it has been difficult to sort out these claims, not least because institutions that practice preferential treatment have been reluctant to disclose the data. With The Shape of the River, the former presidents of Princeton and Harvard, William G. Bowen and Derek Bok, attempt to make the definitive statistical case for preferences. (The title, in case you're wondering, is a pretentious elaboration of the notion of an educational "pipeline" leading from primary school to high school to college and beyond; its point appears to be to let B&B sprinkle in otherwise irrelevant quotes from Mark Twain.) The number-crunching is certainly impressive: there are about 150 graphs and charts in here, many of them-okay, I didn't examine every one-illuminating. But the focus is limited (partly because most colleges aren't selective at all and thus don't use preferences) to 28 elite colleges and universities, including Princeton but not Harvard. B&B provide some perspective on the degree of preference such selective schools give to racial minorities. People who think these institutions admit certain students "just because they're black" will learn that most black applicants to the five schools for which B&B had the most data had higher SAT scores than the average white test-taker, and that the average score for black matriculants to these schools was higher than the average for all matriculants in 1951. Useful also is B&B's point that substantial gaps would still separate the average SAT scores of blacks and whites on a particular campus even under a race-neutral standard: there are just too few really high-scoring blacks. Still, it remains true that among those five schools, a black applicant with scores between 1200 and 1249 has three times as great a chance of acceptance as a white in the same score range. The meat of the book's argument is that black students with the same SAT scores do better the more selective their college, even if their scores are far below the norm. "Doing better" can be measured by graduation rates, post-graduate degree attainment, earnings, and self-reported satisfaction. Black graduates of these 28 schools were actually more likely to get degrees in law or medicine than were their white counterparts (of course, preferences apply here too). The worm in the bud is that black students have lower grades-their grades on average, even if you include those students who made it without preferences, put them in the 23rd percentile of students-and grad- uation rates than their white classmates. B&B argue that this does not confirm the mismatch thesis of preference critics because the likelihood of black students' graduation rises with the selectivity of the school (again, holding SATs constant). But this is true for white students as well. Hence Abigail Thernstrom, co-author with her husband, Stephan, of last year's America in Black and White, seizes on the racial disparity in dropout rates-and this rises with selectivity. At the 28 colleges the black dropout rate (25 per cent) is 78 per cent higher than the white (14 per cent); at the most selective of these it is three times higher (15 vs. 5 per cent). B&B refer in passing to the possibility that under a race-neutral standard students who were refused admission to the most selective schools would simply drop down one tier to a slightly less selective school, but the authors then glide past the point (or seem to: their prose is murky) to reach their conclusion that the number of black students at the 28 colleges would fall by 50 per cent. Mrs. Thernstrom figures that taking the point seriously reduces the decline to something more like 14 per cent. No doubt it will be a while before the statistical dust settles. But those of us who are not social scientists need not fret about the outcome: not much turns on it. The policy dispute has concerned state-sponsored preferences-California's Proposition 209 affects only state universities-not the practices of elite private schools. And the key issue that opponents of preferences have raised is fairness, not lifetime earnings. But B&B insist on misstating the opponents' case: "There is a widespread misconception that scores and grades represent the only truly valid considerations in deciding whom to admit to a selective institution." I am aware of no critic of racial preferences who denies that athletic ability, artistic talent, community service, and other attributes and achievements should be considered in admissions decisions. What critics object to is the use of one criterion, race, which is not a proxy for those qualities and has no relationship to any reasonable definition of merit. But B&B insist on deconstructing "merit" (their sneer quotes), which takes them in some odd directions. At one point they note that "grades and test scores are a reflection not only of effort but of intelligence, which in turn derives from a number of factors," as if this were some sort of indictment. As these passages suggest, B&B are simply hopeless on moral reasoning. That's why they can't explain why they oppose quotas. Or consider the handicapped-parking analogy they borrow from Thomas J. Kane: "Eliminating the reserved space would have only a minuscule effect on parking options for non-disabled drivers. But the sight of the open space will frustrate many passing motorists who are looking for a space." Leave aside the revealing comparison of blackness to a disability. The notion that ending preferences would only marginally increase white students' odds of acceptance, to which B&B return several times, assumes what is at issue: that it is appropriate to classify students by race. Critics are concerned about unfairness to individuals, not to a racially defined group. B&B are even less persuasive in arguing for the educational value of racial diversity. So what if most students at the 28 schools made friends across racial lines? Most people everywhere do, including people who don't go to college at all. Finally, B&B fail to explain why elite colleges matter. Their graduates are too small a segment of the population to significantly narrow the black-white earnings gap, as the authors seem to suggest, or to create a strong black middle class. What B&B are really talking about, then, is the formation of a multiracial elite (with the proper attitudes); but as liberals, they're uncomfortable saying so. The Shape of the River may have something to say about some subsidiary arguments in the racial-preferences debate. But it dodges the main issues and settles nothing. Finally, it ignores everything upstream: the far more consequential question of how to improve K-12 education, particularly in the cities. And to that task, racial preferences are, at best, an affirmative distraction.