In 2003 affirmative-action policies in higher education dodged a bullet when a sharply divided U.S. Supreme Court reaffirmed the use of race in admissions at the University of Michigan Law School in Grutter v. Bollinger. But a second bullet is now headed straight toward the heart of race-based preferences. The projectile is not likely to kill affirmative action outright but rather significantly change its nature. In the future, such programs are likely to place much greater emphasis on socioeconomic status and reduce the role of race. Fresh research suggests that this new balance would make eminent sense.
The 5-4 Grutter decision supporting affirmative action included a caveat that colleges and universities could use race only if they could not create sufficient racial diversity using race-neutral means. Justice Anthony Kennedy—now the swing justice on a more conservative court—dissented in Grutter, saying the Court should “force educational institutions to seriously explore race-neutral alternatives,” rather than being satisfied by Michigan Law School’s “profession of its own good faith.”
Opponents of affirmative action see in Kennedy’s dissent a large bull’s-eye and have shrewdly brought a case challenging affirmative action at the University of Texas at Austin. For a number of years following the Fifth Circuit’s 1996 decision in Hopwood v. State of Texas, barring the use of race in admissions, Austin employed three race-blind alternatives. First, it considered grades and test scores in the context of factors like “socioeconomic status, whether the applicant is from a single-parent home, language spoken at home, family responsibilities, socioeconomic status of the school attended, and average SAT or ACT score of the school attended in relation to the student’s test scores.” Second, Texas employed the “top 10 percent” plan, which allowed students in the top 10 percent of their high-school classes to automatically enroll at Austin. Third, Texas adopted new scholarship programs for low-income and first-generation college students.
When Grutter gave the green light to consider race, however, Austin began to do so, as a supplement to its race-blind initiatives. The plaintiffs in the latest challenge, Fisher v. Texas, say the renewed use of race is unconstitutional. They also point out that in 2004, the use of race-neutral alternatives produced a class that was 4.5-percent black and 16.9-percent Hispanic—exceeding the 4-percent black and 14-percent Hispanic representation achieved before the initial ban on race. The case is now before the U.S. Court of Appeals for the Fifth Circuit.
If the Fisher case ends up in front of the Supreme Court, it seems likely that Justice Kennedy will join conservatives in striking down the use of race at Austin—but he might also leave the door open a crack for a diminished use of race at universities where race-neutral plans don’t work to produce sufficient racial diversity. That compromise is, in essence, the same one Kennedy employed in the court’s 2007 decision nixing racial-integration plans at the elementary- and secondary-school level in Louisville and Seattle. In his concurring opinion, he said that using the race of individual students might be possible as a “last resort” where race-neutral alternatives had failed.
A Supreme Court ruling effectively encouraging universities to use a lot of race-neutral considerations and a little bit of race (only if absolutely necessary) would essentially reverse the weights of current affirmative-action programs. Today many researchers find that selective universities give big preferences based on race and small or no preferences based on socioeconomic status. Thomas J. Espenshade, a professor of sociology at Princeton University, and his co-author Alexandria Walton Radford, found that selective private universities provide a preference that is the equivalent of 310 SAT points for African-Americans, 130 points for low-income students, and 70 points for working-class students. Those findings back up earlier research. William G. Bowen and colleagues found that at selective institutions, among students within a given SAT range, being an underrepresented minority boosts one’s chance of admission by 28 percentage points, but poor students receive “essentially no break in the admissions process; they fare neither better nor worse than other applicants.” Anthony P. Carnevale and Stephen J. Rose, in a 2004 Century Foundation study, found that at the most selective 146 institutions, race-based affirmative action tripled the representation of black Hispanic students, while low-income and working-class applicants received no boost at all.
New research by Carnevale and Jeff Strohl, however, suggests a shift to a program of affirmative action based primarily on socioeconomic status, but including race, would have strong empirical grounding. The roots of Carnevale’s new research reach back to the 1990s, when he was a vice president at the Educational Testing Service and began work on a project quantifying racial and economic disadvantages students face. Carnevale hoped to find a way to identify “strivers,” those talented, disadvantaged students who beat the odds and score far better on the SAT than expected, given the obstacles that stand in their way. The notion of identifying and admitting strivers proved unpopular among university leaders, who didn’t want to have more outside pressure to admit low-income students, and the College Board killed the project. Most universities preferred, as Walter Benn Michaels has acidly observed, to focus on “what color skin the rich kids should have.”
Carnevale and Strohl have now updated and revised the strivers idea for a forthcoming Century Foundation volume that I edited entitled Rewarding Strivers: Helping Low-Income Students Succeed in College. The authors find that the most economically and racially disadvantaged applicant is expected to score an astounding 784 points lower on the combined math and verbal SAT than the most advantaged student. While the average student is expected to score 1054, the most advantaged student is expected to score 1328 and the least advantaged student 544. That 784-point gap accounts for a remarkable 65.3 percent of the 1200-point possible score range on the math and verbal SAT (400-1600 points). In other words, it’s as if, in a 100-yard dash, advantaged kids start off 65 yards ahead of disadvantaged kids when the gun goes off.
It is significant that, of those 784 points, only 56 are based on race per se (being black as opposed to white). The vast majority of the obstacles are socioeconomic in nature. For example, having a father who is a laborer as opposed to a physician costs an applicant 48 points on average. Attending a school where 90 percent of classmates are low-income predicts an SAT 38 points lower than one where no peers receive subsidized lunch. And having a parent who is a high-school dropout, as opposed to highly educated, costs 43 points.
Carnevale and Strohl believe that universities should use both race and class in admissions. The 56 points are meaningful, they say, and there appear to be interaction effects between race and class, though the data set employed was too small to establish the extent of interaction. At the same time, the authors acknowledge that using a more robust measure of wealth (net worth) than they had available to them might in fact eliminate the predictive value of race per se altogether. (Their wealth variable was limited in scope.)
The relatively small role of race in the findings should not be read to suggest that racism and discrimination are things of the past but rather that economic criteria can often capture continuing discrimination. Racial discrimination in the employment sector is reflected in lower earnings for black families. Likewise, pervasive discrimination in the housing market is reflected in the fact that black and Latino students are much more likely to live in neighborhoods of concentrated poverty than are white students of similar income.
Carnevale and Strohl do not advocate mechanically adding SAT points to the raw scores of disadvantaged applicants. They draw a distinction between “deservedness” and “readiness,” pointing out that the most disadvantaged student who scores 700 on the combined SAT will not perform at the level of the most advantaged student who scores 784 points higher, and the lower-scoring student would most likely fail at a selective college. But the authors do suggest that universities consider how far a student has come, as well as what her raw scores are.
The evidence implies that far more strivers could be admitted to selective institutions than are admitted now. Their earlier research found that under a pure merit-based system, combined with socioeconomic affirmative action, the bottom half of the socioeconomic distribution could raise its representation at selective institutions from 10 percent to 38 percent, while still allowing universities to have the same graduation rates as they do now.
Of course, admitting strivers would cost much more money, both to enroll the students and to help them persist to graduation. Another chapter in Rewarding Strivers—by the former New York Times education editor Edward B. Fiske—finds that the University of North Carolina at Chapel Hill’s program of financial aid and academic support has enhanced graduation rates by almost 10 percent among low-income students.
A Supreme Court ruling curtailing the use of race and effectively pushing universities in the direction of class-based affirmative action could create an interesting political space to increase financial aid. Conservatives have embraced President Obama’s statement that his own daughters, as privileged African-Americans, don’t deserve affirmative action and that poor white students do. It is possible, but by no means certain, that the two sides could come together in the future around more financial aid, reduced reliance on race, and a new, more vibrant form of class-based affirmative action.
