What If Colleges Embraced Affirmative Action for Class Instead of Race?

Even if the Supreme Court rules race-based affirmative action to be unconstitutional, schools could still be able to maintain diversity.

Susan Walsh/AP Photo
Michigan's 2006 ban on affirmative action means de-facto segregation, a coalition fighting the ban argues in a brief filed to the Supreme Court. The challenge to the law, which the Court is now considering, starts with the assumption that the number of minority students admitted to Michigan's top colleges would shrink under the ban.
A new approach at the University of Colorado (Boulder) suggests that some colleges could maintain racial diversity while de-emphasizing race in the admissions process. Since 2011, the state flagship has given extra consideration to disadvantaged applicants and to those whose grades and test scores are above average for their economic circumstances.
Even if the state were to ban race-based affirmative action, the class-based framework should increase low-income admissions, maintain minority representation, and even better, predict success in college. It's a model that could help other universities become more diverse.
"What we hope is that this result will ease fears—at least a little bit—that an end to race-based affirmative action will be absolutely devastating for racial diversity on college campuses," says Matthew Gaertner, a research scientist in the center for college and career success at Pearson. Gaertner helped the admissions office devise its new framework while a graduate student at Boulder.
CU-Boulder is a moderately selective institution that primarily serves Colorado and the West. Only about 20 percent of the university's over 29,000 students are African-American, Asian-American, Hispanic, or Native American. The university has been working to increase underrepresented-minority enrollment for the past six years.
Like other selective colleges, CU-Boulder has a set of criteria it uses to weigh marginal applicants—those whose grades and test scores aren't high enough to guarantee admission, but who would still add value to the freshman class. Factors include extracurricular activities, strength of the senior schedule, and race.
The university had to rethink its approach in 2008, when a state ballot initiative threatened to ban affirmative action. "I was very concerned about what could potentially happen to our campus," says Kevin MacLennan, CU-Boulder's director of admissions. He visited flagship institutions in some of the eight states that have banned affirmative action, and he found that they had experienced a significant drop in ethnic and racial diversity.
Even in Texas, where the majority of public-school students are Latino and where the top 10 percent of every high school class gain automatic admission to state universities, the state flagship has found it necessary to consider race in order to increase diversity. Earlier this year, the Supreme Court sent a challenge to the affirmative-action plan at the University of Texas (Austin) back to a lower court.
Social and economic forces are making elite schools both more white and more affluent, a trend documented by Anthony Carnevale and Jeff Strohl of Georgetown University's Center on Education and the Workforce. Inequality is a big part of the problem. Low-income African-Americans and Latinos tend to grow up in more-concentrated areas of poverty than their white peers. Other factors may also be at work: African-American and Latino students are less likely to enroll in selective colleges even when they have solid grades and test scores.
Gaertner set out to create a class-based affirmative-action framework for CU-Boulder that would take into account resources available to a child at home and in high school. He came up with two measures. The "disadvantage index" measures the likelihood that the applicant will enroll in college at all, given his socioeconomic status. The "overachievement index" measures whether an applicant's grades and test scores exceed the scores usually achieved by students of his socioeconomic status.
"When students apply and they demonstrate severe socioeconomic disadvantage, or extraordinary overachievement relative to that disadvantage, they're given a substantial boost in the admissions process," Gaertner says. Overachievers from all backgrounds get extra consideration, and so do severely disadvantaged students, even those who aren't achieving at high levels.
Each index is a statistical model that weighs many factors, including the teacher-student ratio at the applicant's high school, the number of dependents the applicant's parents are supporting, and whether the student's first language was English. Gaertner created a computer program that runs the math and produces a number that admissions officers use to classify students as moderately disadvantaged, severely disadvantaged, or not disadvantaged; and as exhibiting high overachievement, extraordinary overachievement, or none at all.
Two experiments prove that the indicies are working. In 2009, Gaertner had admissions officers review 478 applications, first under CU-Boulder's race-based policy and then under the new class-based policy, with all racial identifiers removed. Officers ended up admitting 9 percent more underrepresented minority students under the race-blind policy than and 20 percent more students of very low socioeconomic status.
In 2010, CU-Boulder ran another experiment, this time on 2,000 applications deemed borderline for admission. Half were evaluated using the new class-plus-race approach, and half using the old approach that used race alone. The hybrid approach resulted in a 13 percent increase in acceptance rates for the poorest students, a 17 percent increase for underrepresented minority students, and a 32 percent increase in the lowest-income, minority students. The results of both experiments were recently published in Harvard Law and Policy Review.
Colorado's ballot initiative didn't pass, so today CU-Boulder uses both race and the new indicies as factors in admissions decisions. In general, disadvantaged students admitted through this process don't perform as well at CU-Boulder as the typical student, Gaertner says. "Their grades and test scores don't indicate they're automatic admits, so you'd sort of expect them to have a little bit lower outcomes than typical undergraduate," he says.
But students identified as overachievers excel. "Their college outcomes—that's grades, credit hours earned, and graduation at four years and six-year graduation rate—is actually higher than typical undergraduates," Gaertner says of overachievers.
Constructing the new admissions factors has had an additional benefit. "The more we got into this, it actually expanded our view of diversity," MacLennan says. The admissions team now pays more attention to rural students and students coming from under-resourced high schools.
While the indicies are a good fit for CU-Boulder, it's not clear whether they'd work just as well at very selective schools, Gaertner says. And it's important to remember that improving on-campus diversity involves a spectrum of efforts.
"The indicies are wonderful, but they can't just stand by themselves—or even part of the admissions process," MacLennan says. Outreach and recruitment comes first. Then, once students are admitted, the university needs to help every admitted student make it to graduation.