Study of Elite College Admissions Data Suggests Being Very Rich Is Its Own Qualification
Elite colleges have long been filled with the children of the richest families: At Ivy League schools, one in six students has parents in the top 1 percent.
A large new study, released Monday, shows that it has not been because these children had more impressive grades on average or took harder classes. They tended to have higher SAT scores and finely honed résumés, and applied at a higher rate — but they were overrepresented even after accounting for those things. For applicants with the same SAT or ACT score, children from families in the top 1 percent were 34 percent more likely to be admitted than the average applicant, and those from the top 0.1 percent were more than twice as likely to get in.
The study — by Opportunity Insights, a group of economists based at Harvard who study inequality — quantifies for the first time the extent to which being very rich is its own qualification in selective college admissions.
The analysis is based on federal records of college attendance and parental income taxes for nearly all college students from 1999 to 2015, and standardized test scores from 2001 to 2015. It focuses on the eight Ivy League universities, as well as Stanford, Duke, M.I.T. and the University of Chicago. It adds an extraordinary new data set: the detailed, anonymized internal admissions assessments of at least three of the 12 colleges, covering half a million applicants. (The researchers did not name the colleges that shared data or specify how many did because they promised them anonymity.)
The new data shows that among students with the same test scores, the colleges gave preference to the children of alumni and to recruited athletes, and gave children from private schools higher nonacademic ratings. The result is the clearest picture yet of how America’s elite colleges perpetuate the intergenerational transfer of wealth and opportunity.
“What I conclude from this study is the Ivy League doesn’t have low-income students because it doesn’t want low-income students,” said Susan Dynarski, an economist at the Harvard Graduate School of Education, who has reviewed the data and was not involved in the study.
In effect, the study shows, these policies amounted to affirmative action for the children of the 1 percent, whose parents earn more than $611,000 a year. It comes as colleges are being forced to rethink their admissions processes after the Supreme Court ruling that race-based affirmative action is unconstitutional.
“Are these highly selective private colleges in America taking kids from very high-income, influential families and basically channeling them to remain at the top in the next generation?” said Raj Chetty, an economist at Harvard who directs Opportunity Insights, and an author of the paper with John N. Friedman of Brown and David J. Deming of Harvard. “Flipping that question on its head, could we potentially diversify who’s in a position of leadership in our society by changing who is admitted?”
Representatives from several of the colleges said that income diversity was an urgent priority, and that they had taken significant steps since 2015, when the data in the study ends, to admit lower-income and first-generation students. These include making tuition free for families earning under a certain amount; giving only grants, not loans, in financial aid; and actively recruiting students from disadvantaged high schools.
“We believe that talent exists in every sector of the American income distribution,” said Christopher L. Eisgruber, the president of Princeton. “I am proud of what we have done to increase socioeconomic diversity at Princeton, but I also believe that we need to do more — and we will do more.”
Affirmative action for the rich
In a concurring opinion in the affirmative action case, Justice Neil Gorsuch addressed the practice of favoring the children of alumni and donors, which is also the subject of a new case. “While race-neutral on their face, too, these preferences undoubtedly benefit white and wealthy applicants the most,” he wrote.
The new paper did not include admissions rates by race because previous researchhad done so, the researchers said. They found that racial differences were not driving the results. When looking only at applicants of one race, for example, those from the highest-income families still had an advantage. Yet the top 1 percent is overwhelmingly white. Some analysts have proposed diversifying by class as a way to achieve more racial diversity without affirmative action.
The new data showed that other selective private colleges, like Northwestern, N.Y.U. and Notre Dame, had a similarly disproportionate share of children from rich families. Public flagship universities were much more equitable. At places like the University of Texas at Austin and the University of Virginia, applicants with high-income parents were no more likely to be admitted than lower-income applicants with comparable scores.
Less than 1 percent of American college students attend the 12 elite colleges. But the group plays an outsize role in American society: 12 percent of Fortune 500 chief executives and a quarter of U.S. senators attended. So did 13 percent of the top 0.1 percent of earners. The focus on these colleges is warranted, the researchers say, because they provide paths to power and influence — and diversifying who attends has the potential to change who makes decisions in America.
The researchers did a novel analysis to measure whether attending one of these colleges causes success later in life. They compared students who were wait-listed and got in, with those who didn’t and attended another college instead. Consistent with previous research, they found that attending an Ivy instead of one of the top nine public flagships did not meaningfully increase graduates’ income, on average. However, it did increase a student’s predicted chance of earning in the top 1 percent to 19 percent, from 12 percent.
For outcomes other than earnings, the effect was even larger — it nearly doubled the estimated chance of attending a top graduate school, and tripled the estimated chance of working at firms that are considered prestigious, like national news organizations and research hospitals.
“Sure, it’s a tiny slice of schools,” said Professor Dynarski, who has studied college admissions and worked with the University of Michigan on increasing the attendance of low-income students, and has occasionally contributed to The New York Times. “But having representation is important, and this shows how much of a difference the Ivies make: The political elite, the economic elite, the intellectual elite are coming out of these schools.”
The missing middle class
The advantage to rich applicants varied by college, the study found: At Dartmouth, students from the top 0.1 percent were five times as likely to attend as the average applicant with the same test score, while at M.I.T. they were no more likely to attend. (The fact that children from higher-income families tend to have higher standardized test scores and are likelier to receive private coaching suggests that the study may actually underestimate their admissions advantage.)
An applicant with a high test score from a family earning less than $68,000 a year was also likelier than the average applicant to get in, though there were fewer applicants like this.
Children from middle- and upper-middle-class families — including those at public high schools in high-income neighborhoods — applied in large numbers. But they were, on an individual basis, less likely to be admitted than the richest or, to a lesser extent, poorest students with the same test scores. In that sense, the data confirms the feeling among many merely affluent parents that getting their children into elite colleges is increasingly difficult.
“We had these very skewed distributions of a whole lot of Pell kids and a whole lot of no-need kids, and the middle went missing,” said an Ivy League dean of admissions, who has seen the new data and spoke anonymously in order to talk openly about the process. “You’re not going to win a P.R. battle by saying you have X number of families making over $200,000 that qualify for financial aid.”
The researchers could see, for nearly all college students in the United States from 1999 to 2015, where they applied and attended, their SAT or ACT scores and whether they received a Pell grant for low-income students. They could also see their parents’ income tax records, which enabled them to analyze attendance by earnings in more detail than any previous research. They conducted the analysis using anonymized data.
For the several elite colleges that also shared internal admissions data, they could see other aspects of students’ applications between 2001 and 2015, including how admissions offices rated them. They focused their analysis on the most recent years, 2011 to 2015.
Though they had this data for a minority of the dozen top colleges, the researchers said they thought it was representative of the other colleges in the group (with the exception of M.I.T.). The other colleges admitted more students from high-income families, showed preferences for legacies and recruited athletes, and described similar admissions practices in conversations with the researchers, they said.
“Nobody has this kind of data; it’s completely unheard-of,” said Michael Bastedo, a professor at the University of Michigan’s School of Education, who has done prominent research on college admissions. “I think it’s really important to good faith efforts for reforming the system to start by being able to look honestly and candidly at the data.”
How the richest students benefit
Before this study, it was clear that colleges enrolled more rich students, but it was not known whether it was just because more applied. The new study showed that’s part of it: One-third of the difference in attendance rates was because middle-class students were somewhat less likely to apply or matriculate. But the bigger factor was that these colleges were more likely to accept the richest applicants.
The largest advantage for the 1 percent was the preference for legacies. The study showed — for the first time at this scale — that legacies were more qualified overall than the average applicant. But even when comparing applicants who were similar in every other way, legacies still had an advantage.
When high-income applicants applied to the college their parents attended, they were accepted at much higher rates than other applicants with similar qualifications — but at the other top-dozen colleges, they were no more likely to get in.
“This is not a sideshow, not just a symbolic issue,” Professor Bastedo said of the finding.
One in eight admitted students from the top 1 percent was a recruited athlete. For the bottom 60 percent, that figure was one in 20. That’s largely because children from rich families are more likely to play sports, especially more exclusive sportsplayed at certain colleges, like rowing and fencing. The study estimated that athletes were admitted at four times the rate of nonathletes with the same qualifications.
“There’s a common misperception that it’s about basketball and football and low-income kids making their way into selective colleges,” Professor Bastedo said. “But the enrollment leaders know athletes tend to be wealthier, so it’s a win-win.”
There was a third factor driving the preference for the richest applicants. The colleges in the study generally give applicants numerical scores for academic achievement and for more subjective nonacademic virtues, like extracurricular activities, volunteering and personality traits. Students from the top 1 percent with the same test scores did not have higher academic ratings. But they had significantly higher nonacademic ratings.
At one of the colleges that shared admissions data, students from the top 0.1 percent were 1.5 times as likely to have high nonacademic ratings as those from the middle class. The researchers said that, accounting for differences in the way each school assesses nonacademic credentials, they found similar patterns at the other colleges that shared data.
The biggest contributor was that admissions committees gave higher scores to students from private, nonreligious high schools. They were twice as likely to be admitted as similar students — those with the same SAT scores, race, gender and parental income — from public schools in high-income neighborhoods. A major factor was recommendations from guidance counselors and teachers at private high schools.
“Parents rattle off that a kid got in because he was first chair in the orchestra, ran track,” said John Morganelli Jr., a former director of admissions at Cornell and founder of Ivy League Admissions, where he advises high school students on applying to college. “They never say what really happens: Did the guidance counselor advocate on that kid’s behalf?”
Recommendation letters from private school counselors are notoriously flowery, he said, and the counselors call admissions officers about certain students.
“This is how the feeder schools get created,” he said. “Nobody’s calling on behalf of a middle- or lower-income student. Most of the public school counselors don’t even know these calls exist.”
The end of need-blind admissions?
Overall, the study suggests, if elite colleges had done away with the preferences for legacies, athletes and private school students, the children of the top 1 percent would have made up 10 percent of a class, down from 16 percent in the years of the study.
Legacy students, athletes and private school students do no better after college, in terms of earnings or reaching a top graduate school or firm, it found. In fact, they generally do somewhat worse.
The dean of admissions who spoke anonymously said change was easier said than done: “I would say there’s much more commitment to this than may be obvious. It’s just the solution is really complicated, and if we could have done it, we would have.”
For example, it’s not feasible to choose athletes from across the income spectrum if many college sports are played almost entirely by children from high-earning families. Legacies are perhaps the most complicated, the admissions dean said, because they tend to be highly qualified and their admission is important for maintaining strong ties with alumni.
Ending that preference, the person said, “is not an easy decision to make, given the alumni response, especially if you’re not in immediate concurrence with the rest of the Ivies.” (Though children of very large donors also get special consideration by admissions offices, they were not included in the analysis because there are relatively few of them.)
People involved in admissions say that achieving more economic diversity would be difficult without doing something else: ending need-blind admissions, the practice that prevents admissions officers from seeing families’ financial information so their ability to pay is not a factor. Some colleges are already doing what they call “need-affirmative admissions,” for the purpose of selecting more students from the low end of the income spectrum, though they often don’t publicly acknowledge it for fear of blowback.
There is a tool, Landscape from the College Board, to help determine if an applicant grew up in a neighborhood with significant privilege or adversity. But these colleges have no knowledge of parents’ income if students don’t apply for financial aid.
Ivy League colleges and their peers have recently made significant efforts to recruit more low-income students and subsidize tuition. Several now make attendance entirely free for families below a certain income — $100,000 at Stanford and Princeton, $85,000 at Harvard, and $60,000 at Brown.
At Princeton, one-fifth of students are now from low-income families, and one-fourth receive a full ride. It has recently reinstated a transfer program to recruit low-income and community college students. At Harvard, one-fourth of this fall’s freshman class is from families with incomes less than $85,000, who will pay nothing. The majority of freshmen will receive some amount of aid.
Dartmouth just raised $500 million to expand financial aid: “While we respect the work of Harvard’s Opportunity Insights, we believe our commitment to these investments and our admissions policies since 2015 tells an important story about the socioeconomic diversity among Dartmouth students,” said Jana Barnello, a spokeswoman.
Public flagships do admissions differently, in a way that ends up benefiting rich students less. The University of California schools forbid giving preference to legacies or donors, and some, like U.C.L.A., do not consider letters of recommendation. The application asks for family income, and colleges get detailed information about California high schools. Application readers are trained to consider students’ circumstances, like whether they worked to support their families in high school, as “evidence of maturity, determination and insight.”
The University of California system also partners with schools in the state, from pre-K through community college, to support students who face barriers. There’s a robust program for transfer students from California community colleges; at U.C.L.A., half are from low-income backgrounds.
M.I.T., which stands out among elite private schools as displaying almost no preference for rich students, has never given a preference to legacy applicants, said its dean of admissions, Stuart Schmill. It does recruit athletes, but they do not receive any preference or go through a separate admissions process (as much as it may frustrate coaches, he said).
“I think the most important thing here is talent is distributed equally but opportunity is not, and our admissions process is designed to account for the different opportunities students have based on their income,” he said. “It’s really incumbent upon our process to tease out the difference between talent and privilege.”