The $6 Solution

A White House summit on college opportunity in January featured more than one Cinderella story. A young man named Troy Simon told the invited group of educators and advocates how he lived for a year in an abandoned building in New Orleans and did not learn to read until he was 14, but made his way to Bard College, where he was a sophomore studying American literature. He introduced Michelle Obama, whose journey from her modest Chicago neighborhood to Princeton University serves as the emotional core of the administration’s campaign to broaden college access.

But stories like Mr. Simon’s and Ms. Obama’s are rare. Many accomplished high-school students don’t even consider places like Bard and Princeton if they’re not born into the country’s elite. Students from the lowest income quartile make up less than 4 percent of the enrollment of the country’s most selective colleges. They may make it to college, but they’re likely to be, to use the lingo of the moment, “undermatched"—attending less-selective colleges than their grades and test scores suggest they could.

At the White House, David Coleman, president of the College Board, described its efforts to expand options for top students from less-privileged backgrounds. “We must be committed to propelling them into the opportunities they have earned,” he said.

The summit was just the latest and largest stage for a conversation about undermatching, which in the past 18 months has leapt from the pages of academic journals into the mainstream conversation. None of the research has received more buzz than a pair of studies by Caroline M. Hoxby, a professor of economics at Stanford University.

Ms. Hoxby’s studies, featured at the White House summit, deliberately avoid using the term “undermatching"—she says students’ college-going decisions are more complicated than that—but, even so, her research has become synonymous with the term. The two papers consider low-income students whose grades and test scores put them in the top 4 percent of achievement and colleges ranked in the top 230 or so by Barron’s. As a result, much of the discussion now revolves around that world.

One paper, written with Christopher Avery, a professor at Harvard University’s John F. Kennedy School of Government, finds there are tens of thousands of such students and describes their college-application behavior. The other, written with Sarah E. Turner, a professor of economics at the University of Virginia, tests whether mailing such students information and application-fee waivers changes where they apply and ultimately enroll. The College Board is now expanding that experiment, which found that the low-cost intervention made a difference.

The White House is enthusiastic. Helping students make better college choices, says James Kvaal, deputy director of the Domestic Policy Council, is “an inexpensive and effective way to help our country produce more college graduates.” But critics say policy makers expect it to achieve more than it possibly could.

They contend that the hype around undermatching, particularly the focus on high-achieving, low-income students, is a distraction. Colleges and policy makers should be grappling with broader, systemic changes needed to make the higher-education system more equitable. It “diverts attention from other things that would have better impacts,” says Michael Bastedo, an associate professor and director of the Center for the Study of Higher and Postsecondary Education at the University of Michigan at Ann Arbor.

In the end, the focus on undermatching is friendly to the status quo, says Anthony P. Carnevale, director of Georgetown University’s Center on Education and the Workforce. “It is uncritical of the selection system. It basically says keep doing things the way you’re doing them and we can find a few more exceptions to the rule and go on and admit all the rich kids.”

So why has the research attracted such a spotlight? Well, for many reasons. Its path to prominence reflects policy makers’ hunger for measurable problems and cost-effective solutions. It highlights how comfortable elite colleges are with the existing hierarchy of institutions. Perhaps most of all, it shows how badly Americans want to believe that their society—and the higher-education system in particular—rewards talent, not privilege.

While the term “undermatching” is fairly new, researchers have been studying college choice for decades. The purpose of policy discussions used to be access, simply getting more students into college, says Jessica Howell, executive director of policy research at the College Board. Where they went wasn’t terribly important.

But the focus has shifted in recent years, from access to completion. And once graduation is a goal, distinctions among colleges do matter, a lot.

Using data from local public schools, the University of Chicago Consortium on Chicago School Research published a pivotal study in 2008 demonstrating that students there did not attend the kinds of colleges their level of preparation suggested they could. That study and more by the Chicago consortium and other groups considered a broader population than just students who might go to the top colleges.

Another group of researchers, William G. Bowen, Matthew M. Chingos, and Michael S. McPherson, soon found a similar pattern in North Carolina, which they documented in the book Crossing the Finish Line in 2009, coining the term “undermatching,” as far as they can tell. One of the book’s main findings was that whatever students’ grades and test scores, they had a better shot at graduation at a more selective college.

At the time, that was a surprising insight, says Mr. McPherson, a higher-education economist and former college president who now leads the Spencer Foundation.

According to conventional wisdom, if a student went to a college where he would be at the bottom academically he might flounder. So discovering that students were actually better served at a more-selective college—at least when it came to graduating—had important implications. If they undermatched, they risked the possibility of starting college and leaving without a degree.

Crossing the Finish Line also established that students with lower socioeconomic status undermatched at higher rates than did their more-privileged peers.

Beyond the new focus on completion, some elite colleges’ decisions to offer more-generous financial aid also prompted interest in undermatching, says Ms. Howell. The policies raised the question of whether there was an untapped pool of low-income students who would apply if they thought they could afford such colleges—and be admitted.

Ms. Hoxby and Mr. Avery set out to answer that question for Harvard in particular, whose new policy, announced in 2004, said, among other things, that parents earning less than $40,000 would not be expected to contribute to its cost for their children. With the program in place, the share of students in Harvard’s entering class who were low-income increased, the researchers found—the percentage of students with family incomes of $60,000 or less grew from just under 15 percent for those who entered in 2004 to more than 16 percent for those who entered in 2005—without affecting its admissions standards.

But many apparently qualified low-income students still hadn’t applied. How many? Another paper by Ms. Hoxby and Mr. Avery estimated that 25,000 to 35,000 high-achieving, low-income students are among each year’s high-school graduates in the United States. The vast majority, the paper found, do not apply to any selective college.

In March 2013, Ms. Hoxby announced the successful results of the experiment testing the information packets and fee waivers, a randomized controlled trial. Now there was both evidence of the scale of the problem and an inexpensive solution—each set of materials cost only about $6. Right away, the College Board announced that it would adapt and expand the experiment. The organization is always looking for solutions that are evidence-based and scalable, says Steve Colón, vice president of its Access to Opportunity program: This was both.

All of this led to a flurry of public attention. The experiment was written up in The New York Times. Last June, Ms. Hoxby presented a policy proposal for expanding the program at an event in Washington put on by the Brookings Institution. In September, the State of Delaware announced that it would use similar strategies to reach out to students. Then came the White House summit.

The gathering was part of the administration’s effort to reform higher education in ways that don’t require Congressional action. And policy makers are naturally attracted to cheap solutions.

This one is so cheap, says Mr. McPherson, “it’s like magic.”

But how effective will changing the colleges students consider really be at producing more graduates? Ms. Hoxby and Ms. Turner say they are not making any claims. They declined to be interviewed for this article but agreed to answer some questions jointly via email. “We hypothesize that improving information about college choices might improve college outcomes significantly,” they wrote. But “to think of collegiate attainment as one problem in need of one solution oversimplifies.”

Undermatching research in general has met with at least one serious scholarly critique, in a review essay in which Mr. Bastedo, of the University of Michigan, and Allyson Flaster, a Ph.D. candidate there, questioned some of its core assumptions.

Measuring undermatching, they wrote, means separating colleges into tiers of a hierarchy and predicting which students would be likely to get in where. They challenge researchers’ ability to do either adequately. It isn’t clear that colleges are sorted on the “margins that matter,” they argue. The differences among colleges are significant when it comes to the extreme ends of the spectrum, they write, but the middle is more muddled. And even when students’ grades and test scores are known, predicting their chances of admission is fraught, given selective colleges’ holistic evaluations. And in the end, they argue, getting more low-income students to attend match colleges won’t be enough to make the higher-education system less stratified.

Ms. Hoxby, who has a reputation for responding sharply to criticism, told Inside Higher Ed to “simply ignore this low-quality study.”

She and Ms. Turner declined to comment on that reaction in their emailed responses to The Chronicle, but provided a more detailed response to the criticism. “We have repeatedly emphasized that distinctions among colleges need to be measured accurately and their effects on outcomes need to be tested using rigorous, scientific methods,” they wrote. “First, we are using a randomized controlled trial to test whether students have different outcomes as a result of attending different colleges. Second, in our recent paper, we cite only the convincingly causal evidence on the effects of attending different colleges.” The researchers didn’t make assumptions, they argue, but ran an experiment.

If the policy discussion is fixated on elite higher education, that’s not accidental. Scholars and policy wonks are often themselves the products of elite colleges, says Jeffrey Smith, a professor of economics and public policy at Michigan who has studied undermatching. That’s the world they know, and its future membership matters to them.

Finding opportunities for high-achieving, low-income students also resonates emotionally. To be flip, says Awilda Rodriguez, a research fellow at the American Enterprise Institute who has studied undermatching, “it’s the homeless-to-Harvard story.” Nobody said selective college admissions is fair, but the idea that qualified low-income students simply don’t know they could go to elite colleges feels viscerally unfair.

And focusing on such students is appealing because the message experts want to send them is simple. Researchers—and the College Board—are very careful to say that they’re not trying to get students to go to any particular college, or even a certain type of college. After all, academics aren’t the only consideration in college choice. “I don’t want to tell them where to go,” says Mr. Colón. “I want to tell them where they could go.”

Because the message is simple, and successful high-school students are probably attentive to written materials, sending them information is likely to be particularly effective. And those students are arguably an important group. If the country’s leaders tend to come out of a handful of elite colleges, the thinking goes, then making sure low-income students are represented there is one good way to get a broader group of leaders.

Top students also send a signal to their peers, says Melissa Roderick, the lead researcher on the Chicago consortium’s undermatching work. A district like the Chicago Public Schools wants to persuade students that they should work hard and do well in their classes. If students see that the class valedictorian heads off to the same unselective college as any other collegebound classmate, that argument is a lot harder.

But the very top students—a tiny fraction of all potential college-goers—are not the only ones who undermatch. In fact, the Chicago study found that students with the credentials to attend very selective colleges actually undermatch at a lower rate than those with the credentials for colleges on the next rung down. The College Board is also testing a different outreach program for a broader population of students called “Apply to 4 or More,” encouraging them to do exactly what it sounds like. The program includes fee waivers, as well as training for school counselors.

And in their email to The Chronicle, Ms. Hoxby and Ms. Turner hint at broader research: “Developing interventions that work for low-income high-achievers is a logical first step to developing information-based interventions that may improve college choice for students of all incomes and degrees of college readiness.”

But moving down the selectivity scale—away from the highest graduation rates and most-generous need-based financial aid—choices aren’t as straightforward. Say a student’s options are a regional public college with a low graduation rate (where she’ll have to borrow a bit) and a midtier private institution, where the graduation rate is somewhat higher but she’d have to borrow more. Given such trade-offs—pretty typical for the bulk of students bound for four-year colleges—an informed decision might require more help than a packet of information can offer.

So what would happen if all existing students and college seats were reshuffled to achieve perfect academic match? Graduation would barely budge, says Eleanor W. Dillon, an assistant professor of economics at Arizona State University.

Her work with Jeffrey Smith at Michigan has found a good deal of both undermatching and overmatching throughout higher education. They look at student ability through the Armed Services Vocational Aptitude Battery and sort colleges using a quality index Mr. Smith helped create. That system, the researchers say, allows for finer distinctions among colleges.

Fixing match, they will argue in a forthcoming paper, would replace overmatched affluent students at top colleges with currently undermatched low-income students. Even so, across the system, high-income students would still be concentrated at top colleges, and their low-income peers at less-selective institutions, because of the unequal preparation available to different populations.

With every student placed at a match college, Ms. Dillon says, the probability that all students starting at four-year colleges would graduate within six years would change by less than one percentage point. Eliminating undermatching does improve graduation rates, the researchers found. But wiping out overmatching hurts overall completion—by about the same amount.

And what would shuffling students around do to the peer effects of being on a particular campus? One component of college quality, after all, is the student body. The researchers are continuing to explore how that would play out.

Their evidence suggests that improving completion at big, less-selective institutions like Ferris State and Eastern Michigan Universities would “make a bigger difference,” says Mr. Smith, “than reshuffling people.”

Mr. Chingos, an author of Crossing the Finish Line, has also run an estimate on how perfect matching would affect degree attainment. In a blog post ahead of the White House summit, he argued that while reducing undermatching is a “laudable goal,” it would not improve graduation rates very much. But that doesn’t mean he considers the work unimportant. In fact, given the low cost of the intervention designed by Ms. Hoxby and Ms. Turner, he thinks it should be pursued.

“If you can pay six bucks and move the needle,” Mr. Chingos says, “you should do it tomorrow.”

Ms. Hoxby and Ms. Turner agree. Besides, they wrote to The Chronicle, “improving information is so inexpensive that efforts on it do not crowd out other efforts” to tackle higher ed’s big challenges.

Ms. Hoxby has found that the driving reason so few high-achieving, low-income students land at top colleges is not because they don’t get in, but because they don’t even apply. That does not necessarily suggest what would happen if, suddenly, they did. Selectivity, after all, comes from rejecting applicants. Not everyone with great grades and test scores gets in. And right now, there are “many more students qualified to do well than colleges with good graduation rates to support their attendance,” says Mr. Avery, one of Ms. Hoxby’s co-authors. “There aren’t enough spots at those colleges we’re currently labeling selective.”

Some undermatching researchers are convinced that elite colleges are working hard to ensure that more of their slots go to low-income students. Others are skeptical. Elite colleges don’t have much incentive to bring in more low-income students, says Ms. Roderick, of the Chicago consortium. “I’d like to see the commitment on the part of the colleges,” she says.

If undermatching among top students were reduced to some acceptable level, then what would happen? So far, that’s uncertain.

The results of the College Board’s project may offer the beginnings of an answer. As soon as this fall, researchers there will know where the students who received information packets applied to college, and where they enrolled. A fuller picture of how they fare, of course, is still years away.

The Rise of Undermatching Research

Caroline Hoxby

February 2006: “Cost Should Be No Barrier: An Evaluation of the First Year of Harvard’s Financial Aid Initiative,” by Christopher Avery, Caroline M. Hoxby, and others, is issued as a National Bureau of Economic Research working paper. The researchers found that Harvard University’s new effort drew a group of applicants that was larger and somewhat poorer than before.

Melissa Roderick

March 2008: The University of Chicago Consortium on Chicago School Research releases a research report, “From High School to the Future: Potholes on the Road to College,” that sheds light on the issue of college match among students in Chicago Public Schools.

September 2009: Crossing the Finish Line, by William G. Bowen, Matthew M. Chingos, and Michael S. McPherson, is published. The book, believed to contain the first use of the term “undermatching,” presents evidence that students have a better shot at graduation if they go to a more selective college.

Christopher Avery

December 2012: “The Missing ‘One-Offs’: The Hidden Supply of High-Achieving, Low-Income Students,” by Ms. Hoxby and Mr. Avery, is released as a working paper by the National Bureau of Economic Research. It estimates the population of high-achieving, low-income students and explains that many of them are not located in the places where college recruiters are likely to be looking.

Sarah Turner

March 2013: “Expanding College Opportunities for High-Achieving, Low Income Students,” by Ms. Hoxby and Sarah E. Turner, is released as a discussion paper by the Stanford Institute for Economic Policy Research. It shares the results of an experiment in which the authors mailed information on colleges and application fee waivers to such students.

March 2013: Ms. Hoxby and Ms. Turner announce that the College Board will expand their experiment.

June 2013: The Brookings Institution’s Hamilton Project releases a proposal from Ms. Hoxby and Ms. Turner saying they plan to work with the College Board and ACT to adapt their experiment and asking for federal data that would help them further customize their outreach efforts.

September 2013: The State of Delaware announces its effort to combat undermatching.

December 2013: Ms. Hoxby and Ms. Turner are included in The Chronicle’s 2013 Influence List.

January 2014: Undermatching is discussed at the White House college opportunity summit.

Michael Bastedo

February 2014: Educational Researcher releases “Conceptual and Methodological Problems in Research on College Undermatch,” by Michael Bastedo and Allyson Flaster, which describes assumptions made in undermatching work and argues that they may not hold.