The Chronicle Review

What Sotomayor Gets Wrong About Affirmative Action

June 17, 2014

Affirmative-action programs at colleges are being squeezed by America’s most democratic mechanism (the voter referendum) and its least (the judiciary.) The two forces merged in late April when the U.S. Supreme Court ruled, 6-2, that Michigan voters have a right to amend their state Constitution to ban racial preferences in admissions at public universities. In so doing, the court affirmed laws in eight states that have 29 percent of America’s high-school population and more than 40 percent of its Hispanic residents.

In the case, Schuette v. Coalition to Defend Affirmative Action, the court’s only Hispanic member, Justice Sonia Sotomayor, wrote a widely acclaimed dissent, in which she challenged Chief Justice John Roberts’s colorblind approach to college admissions as "out of touch with reality."

A new report that I edited, just released by the Century Foundation and the Lumina Foundation, suggests, however, that the concerns of both justices can be met: Alternatives to race-conscious affirmative-action, if properly structured, would produce more diversity than just concentrating on race.

According to a chapter by Anthony P. Carnevale and his colleagues at Georgetown University in the new report, The Future of Affirmative Action: New Paths to Higher Education Diversity After Fisher v. University of Texas, using socioeconomic preferences and/or plans that admit a top percentage of students from every high school, if structured properly, could produce even higher levels of black and Hispanic representation at the most selective colleges than racial preferences now achieve. That approach would work because it reflects economic disadvantages that are often shaped by racial discrimination.

Sotomayor’s dissent in Schuette is a strong reminder of the importance of race. "Race matters to a young man’s view of society when he spends his teenage years watching others tense up as he passes, no matter the neighborhood where he grew up," she wrote. "Race matters because of the slights, the snickers, the silent judgments that reinforce that most crippling of thoughts: ‘I do not belong here.’ "

Past and present discrimination certainly help explain why black and Hispanic-Americans on average have lower incomes and smaller savings than whites do, and why even middle-class blacks live in neighborhoods with higher poverty rates than low-income whites. It does not necessarily follow, however, that racial preferences are the best response. The plurality opinion in Schuette, joined by the chief justice, observed that the preference system "has the potential to become itself the source of the very resentments and hostilities based on race that this Nation seeks to put behind it."

Fortunately, there are policies that recognize the effects of racial discrimination and growing socioeconomic divisions while also avoiding the toxic use of race in deciding who gets ahead. That third path is essentially what the Supreme Court sought to foster in its 2013 decision in Fisher v. University of Texas at Austin, which affirmed the goal of racial diversity but said that colleges must seek alternative ways of producing it before resorting to affirmative action.

In Schuette, Sotomayor wrote that preferences provide the only realistic path to racial inclusion in higher education, correctly noting that race-neutral alternatives have failed to produce adequate diversity at three high-profile institutions—the University of Michigan at Ann Arbor, the University of California at Berkeley, and the University of California at Los Angeles.

But Michigan, Berkeley, and UCLA are outliers. In a 2012 study of 10 leading universities where race was banned, my colleague Halley Potter and I found that seven other selective public institutions were able to preserve or even exceed black and Hispanic levels of representation by using race-neutral strategies such as socioeconomic affirmative action.

Why did Michigan, Berkeley, and UCLA fail? Look more closely at Michigan. While the university has taken modest steps to increase socioeconomic (and thereby racial) diversity—by factoring in disadvantage and providing scholarships for students transferring from community colleges—it could be doing much more. For example, it still gives preferences in admission to the children of alumni, something that other colleges facing bans on racial preferences—such as the University of Georgia and Texas A&M University—have stopped doing. Research finds that legacy preferences tend to benefit wealthy and white students. (Berkeley and UCLA have done a better job of promoting socioeconomic diversity and have dropped legacy admissions.) Michigan also still provides substantial "merit" aid to wealthy students, thereby diverting funds from need-based aid. And a recent report found the university to be among the "worst offenders" among institutions showering enormous pay and bonuses on administrators. In all, only 15 percent of Michigan students are eligible for federal Pell Grants, which are need-based, compared with more than 25 percent at public flagship universities nationally.

Moreover, questions have been raised about the data cited by Sotomayor. In 2010 the Education Department changed its methodology for categorizing students by race and ethnicity, requiring colleges to report separately students who are members of two or more races. "So a drop in the number of black students reported at a university from 2009 to 2010," a Chronicle article noted, "doesn’t necessarily mean that there were actually fewer black students."

But the biggest problem in Sotomayor’s analysis is that the three campuses she cites are national universities that compete for talented black and Hispanic students with other prestigious institutions that continue to use racial preferences. A black student who gets into Berkeley without a preference also probably gets into an even more highly ranked private institution, like Stanford University, with a preference. How is Berkeley to compete?

Now consider: What would happen if top colleges all played by the same rules?

In their new study, Carnevale, Stephen J. Rose, and Jeff Strohl conducted a simulation of what would happen if race were eliminated from admissions decisions across the board, and colleges instead employed socioeconomic preferences and/or plans that admit the highest-scoring students from every high school. Because of the unfortunate association between race and class in American society, the researchers found that such plans would improve the combined black and Hispanic representation at the top 193 colleges. Moreover, in most cases, the plans could also raise, rather than reduce, average SAT scores at institutions.

As the figures above show, at the most selective 193 institutions, under a system of race-conscious affirmative action, legacy preferences, athletics preferences and the like, 11 percent of students are either black (4 percent) or Hispanic (7 percent). The average combined verbal and math SAT scores for all students at those colleges is 1230.

Shifting to a system of admissions based purely on merit as defined by SAT scores would see an increase in average scores to 1362, but racial diversity would plummet (4 percent Hispanic, 1 percent black).

Under a class-based affirmative-action program, which considers a variety of socioeconomic disadvantages, the combined representation would grow from today’s 11 percent to 13 percent (10 percent Hispanic, 3 percent black.) Average SAT scores would remain high, at 1322.

Under a system in which the top 10 percent of test-takers in every high school were admitted, the combined minority proportion would rise from 11 percent today to 17 percent (11 percent Hispanic, 6 percent black). The combined SAT scores would average 1254, slightly higher than today.

Finally, if a top 10-percent plan were merged with economic affirmative action, the combined minority representation would rise further, to 23 percent (14 percent Hispanic, 9 percent black.) But combined SAT scores would drop from 1230 to 1160. Although this option provides the biggest payoff in terms of diversity, the modest 70-point drop in average SAT scores might have a small impact on graduation rates.

Under these various scenarios, racial diversity would suffer at colleges below the top 193—but it already suffers under racial-preference programs, which increase diversity at the top but decrease it the next rung down.

Shifting from racial considerations would substantially increase socioeconomic diversity. While those in the bottom socioeconomic half currently enjoy access to just 14 percent of seats at selective colleges, that would rise to 46 percent under socioeconomic affirmative action, 31 percent under a top-10-percent plan, and 53 percent under a program combining the two.

Achieving racial diversity by such alternative means is a matter of fairness and equity: While race matters in allocating opportunity, class is an even more significant barrier to success. Although the achievement gap by race used to be twice as large as the achievement gap by income, today the reverse is true.

In the news-media coverage around Justice Sotomayor’s dissent in Schuette, race and class were often conflated. It was noted that she had grownup in public housing, which is part of what makes her story so inspiring. But today most affirmative-action beneficiaries do not fit the Sotomayor mold: At selective colleges, 86 percent of blacks are middle or upper class, according to one study. At the top 20 law schools, another study finds, 89 percent of black students come from the top half of the socioeconomic distribution, and 66 percent from the top quarter.

Considering socioeconomic disadvantage and geography in admissions implicitly recognizes that race still matters in American society. As the Georgetown University law professor Sheryll Cashin demonstrates in her new book Place, Not Race (Beacon Press), economic disadvantage and neighborhoods are highly racialized. At the same time, Cashin observes that her own black sons are educationally privileged and do not need a leg up in admissions. Moreover, by addressing what she calls a superficial level of "diversity by phenotype," racial preferences can also allow colleges to avoid creating student bodies that are socioeconomically diverse as well.

Attacks on affirmative action are likely to mount. The Project on Fair Representation is busy recruiting plaintiffs to bring lawsuits against Harvard University, the University of North Carolina at Chapel Hill, and the University of Wisconsin at Madison. Meanwhile anti-preference initiatives are possible in Missouri, Ohio, Tennessee, and Utah.

It is incumbent upon those who care about social mobility and racial inclusion to come up with alternatives that implicitly recognize that race matters in American society—and that, today, class matters even more.

Richard D. Kahlenberg, a senior fellow at the Century Foundation, is editor of The Future of Affirmative Action: New Paths to Higher Education Diversity After Fisher v. University of Texas, just released as an e-book by the Century and Lumina Foundations.