Back in March the New York Times published an interactive piece entitled, “Can You Create a Diverse College Class Without Affirmative Action?” The article, by Aatish Bhatia and Emily Badger, examined a study conducted by Stanford researchers Sean Reardon and Demetra Kalogrides to determine whether alternatives to race-based admissions preferences could increase diversity at elite colleges and universities.
Reardon and Kalogrides ran simulations for four scenarios to calculate how each would impact the racial composition of a class, the percentage of low-income students, and the average SAT score. Utilizing real-life academic and demographic information from the high school class of 2013, they developed models for an institution with a five percent admit rate, where 500 of 10,000 applicants would be admitted.
Scenario one involved giving students from low-income backgrounds a boost equivalent to 150 points on the SAT. The second scenario added a second, 150-point boost for students coming from high-poverty high schools. Scenario three was based on finding “outliers,” students who outperform their peers from similar backgrounds. Dartmouth used the outlier theory as one rationale for its return to requiring test scores. The final scenario involved expanding the applicant pool by identifying qualified students who aren’t applying to elite colleges. The authors of the study report report that scenario four produces the biggest shift toward low-income students, with Hispanic students benefiting the most.
So how did each scenario impact the composition of the projected entering class?
Scenario % Black/Hispanic % Low Income Avg. SAT
One 13 10 1390
Two 18 11 1380
Three 25 17 1340
Four 32 23 1320
The models above demonstrate the complex relationship that different levers have on the composition of a desired class. Making progress in one area carries with it a cost in another. Hoping to increase the percentage of diverse and low-income students will probably lower the average SAT score of a class. Admission offices have to decide which institutional priority takes precedence.
ECA is always interested in underlying assumptions, and there are several in this study worth examining..
The researchers state that they “use SAT scores as a simplified measure of academic merit.” I wonder if that holds up to scrutiny.
Clearly one appeal of SAT scores as a metric is that they are simple, a way to categorize a student in four digits. I have questioned before whether we measure what we value, or value what is easily measured.
But are SAT scores a measure of academic merit? I have never found SAT scores to be random. With rare exceptions my good students scored well and weaker students didn’t. But does a strong test score indicate “academic merit” or the ability to test well?
The “academic merit” arguments hearken back to a century ago when testing was seen as an objective way of measuring human potential. Today we know how flawed and even ignorant that was, with subtle ties to the eugenics movement. Even David Coleman would have a hard time making the argument today that SAT scores measure “academic merit.” An individual’s score is influenced by the environment he or she grows up in, both at home and at school. In addition, the ability of wealthy families to pay for test prep and multiple test administrations means that identical scores don’t mean the same thing. SAT scores may be simple, but not a measure of academic merit.
The Stanford study makes an important distinction between race and income. It states that, “Income is a relatively weak proxy for race in admissions. A preference for lower incomes produces just that: students with lower incomes, not necessarily a much larger share of Black or Hispanic students.” It points out that one criticism of race-based admission preferences was that the Black and Hispanic students admitted under those policies were “overwhelmingly middle- and upper-income students.” Of course the elephant in the room is that admitting more low-income students carries with it the obligation to provide more financial-aid resources.
The boost given by the researchers to low-income students (150 points on the SAT in scenario one and 300 points for students from high-poverty schools in scenario two) would seem to come into conflict with the “diamond in the rough” theory advanced by some of the Ivy League institutions that have reinstated their test score requirement. The argument is that the SAT identifies students who would otherwise be overlooked as applicants for the nation’s elite colleges.
But how many diamonds in the rough are out there? According to the 2022 annual report for the SAT “Suite of Assessments,” one percent of African-American test takers scored above 1400 on the SAT, or about 2000 students nationally. For test takers self-identifying as Hispanic, it was two percent, or approximately 8000 nationally. Those numbers aren’t broken down by income, but it’s a very small pool even if everyone in it is from a low-income background, and that’s likely not the case. So does the SAT serve to find diamonds in the rough, or do students from low-income backgrounds and schools require a significant boost because their test scores are an impediment to admission?
That leads to the biggest flaw in the Reardon/Kalogrides study. They begin with a desired result, a certain composition of the class, and then attempt to reverse engineer the admissions process to produce the desired outcome. Is that any different from what highly-selective/rejective colleges did with regard to race and other factors in “shaping” a class? Probably not. Critics of race-based affirmative action were convinced that colleges were putting a thumb on the scale to achieve a certain level of diversity.
If I were an admissions dean, charged with helping my institution achieve a variety of strategic goals through the admissions process, I’d be tempted to do the same thing. But admitting a class rather than admitting students as individuals serves institutions much better than it does students. Every applicant deserves equal consideration, and that doesn’t happen in an admissions process that is reverse engineered to produce a previously desired outcome. Chief Justice Roberts’ opinion in the Harvard/UNC cases finds that race can be taken into account as part of the evaluation of an individual’s qualifications, whereas setting goals to achieve certain levels of diversity within the entire class is impermissible.
There are certainly challenges to that approach. How does a college or university handle a situation where its diversity numbers drop precipitously from year-to-year?
Perhaps the most interesting idea in the Stanford study is that the ultimate solution is expanding outreach and identifying students we are now missing. That’s a huge challenge at any time, but especially this year when the FAFSA rollout fiasco has challenged colleges’ ability to enroll their classes for this fall. We may agree that higher education has an obligation to serve the public interest and not just institutional interest, but according to Maslow’s hierarchy of needs it’s hard to worry about self-actualization when you are worried about where your next meal is coming from. When you are on a flight and the emergency oxygen masks drop down, you are supposed to first put on your own mask.
There is, however, a group of colleges that do have the opportunity to make a difference. That, of course, is the group of “elite” colleges and universities that admit a minuscule percentage of applicants, the kinds of places that receive the vast majority of attention from media outlets like the New York Times. We need that group of colleges to step up and lead with a sustained commitment to do pro bono work long before students enter high school. It also requires from all of us a more sustained effort to identify the qualities, both academic and non-academic, that lead to success. It won’t be easy, but the things in life that are most valuable aren’t easy.