It was an unsettling experience at Harvard that spurred Hoxby to study the students she now is obsessed with helping. In the summer of 2004, then-president Lawrence Summers and his brain trust were frustrated that the school was still largely a place for the affluent. Despite the fact that low-income students had long had virtually a free ride, only 7 percent of the class was coming from the bottom quartile of income, while nearly a third came from families earning more than $150,000 a year. So the school announced to much fanfare that it would officially be free for those with less than $40,000 in annual family income (now up to $65,000). No loans, just grants to cover the entire cost. The administration figured the program would instantly flush out superstar high-school seniors from unexpected places—hardscrabble Midwestern farming communities, crime-pocked cities too small for a recruiter to visit, maybe even a small Nebraska town where a girl with straight A’s seemed destined to languish in her local community college.
But when April rolled around, there was nothing to celebrate. The number of incoming freshmen with family incomes below $40,000 was virtually flat, fewer than 90 in a class of 1,500, a tiny bump of only 15 or so students. Other elite institutions that had quickly matched Harvard’s program reported even more depressing statistics.
So Hoxby, who was on the faculty at the time, began to analyze what had gone wrong. A former Rhodes scholar with a PhD from MIT, she had almost single-handedly created the field of educational economics. Her previous work had measured whether charter schools raise student achievement, whether class size really mattered and how school vouchers worked.
The problem captured her immediately. She had analyzed the data enough to know that many qualified low-income students were not applying to selective schools. While Harvard could afford to step up its expensive outreach—in recent years it and other top schools have increased the proportion of low-income students to as much as 20 percent—Hoxby estimated that there were huge swaths of kids who were being overlooked.
“Caroline,” says Harvard’s Fitzsimmons, “has a great heart as well as a great intellect. And like every economist, she hates waste, especially a waste of human capital.”
First she had to figure out how many qualified students were actually out there—and where. The College Board and its counterpart, the ACT, which administers another admissions test, knew who had high scores, but not who was poor. Test-takers are asked about family income, but only about 38 percent respond, and, as Hoxby says, “lots of kids have no idea what their parents make.” Colleges glance at application ZIP codes, but that’s a blunt instrument, especially in vast rural areas. Ironically, “need-blind” admissions, used by about 60 top schools, had contributed to the dearth of information. The policy, instituted to make sure the process didn’t favor wealthy students, precludes schools from asking applicants about their household income.
So Hoxby, 47, and co-author Christopher Avery, a public policy professor at Harvard’s John F. Kennedy School of Government, tackled a monumental data challenge. They decided to look at every senior in the U.S. in a single year (2008). They devised a complicated set of cross-references, using block-by-block census tract data. They matched each student with an in-depth description of his or her neighborhood, by race, gender and age, and calculated the value of every student’s house. Parents’ employment, education and IRS income data from zip codes were also part of the mix. They even tracked the students’ behavior in applying to college.
The results were shocking. They found approximately 35,000 low-income kids with scores and grades in the top 10 percentile—and discovered that more than 80 percent of them didn’t apply to a single selective institution. In fact, a huge proportion applied to only one college, generally a non-selective school that required only a high-school diploma or a GED, and where a typical student had below-average scores and grades.
Mostly from rural backgrounds, crumbling industrial outposts or vast exurbs, these students had been falling through the cracks for generations. Elite institutions traditionally concentrated on a small number of cities and high schools in densely populated, high-poverty areas, places that had reliably produced talented low-income students in the past. Smaller markets, such as Nashville, Topeka and Abilene, rarely got a look. Kids in rural settings were even less likely to catch the eye of college admissions staff, especially with college counselors an endangered species—the ratio of counselors to students nationally is 333 to one.
“When you’re in admissions, you go to the schools that you know, to areas likely to have a number of kids like that,” says Hoxby. “You might have a school in New York, for instance, that has a really great English teacher whose judgment you trust. You work your contacts, just like in everything else.”