Researchers Propose New Way to Gauge Colleges’ Success in Enrolling Low-Income Students

By Lindsay Broderick, Staff Writer

A recent study conducted by Caroline Hoxby and Sarah Turner of Stanford University and the University of Virginia, respectively, has received a great deal of traction from both higher education news sources, like The Chronicle and Inside Higher Ed, and mainstream media, like The Washington Post.

This study is getting so much attention because it offers an alternative to the parameters typically used to measure an institutions’ ability (or willingness) to enroll low-income students. The number of students who are Pell-eligible nationally or who are in the bottom quintile (20 percent) of the national distribution of income are common metrics for assessing how institutions serve low-income students.

While tracking the number of students who are Pell-eligible at a state’s public flagship institution is an easy, standardized method to measure low-income students’ enrollment rates, Hoxby and Turner argue that this is unfair to institutions in states with a lower overall percentage of Pell-eligible students in their “relevant pool.”

An institution’s “relevant pool,” according to Hoxby and Turner, is the population of students in a state whose SAT or ACT scores make them a reasonable match for the flagship institution (within the 25th-75th percentiles of the institution’s student profile). These vary between colleges and states and, therefore, alter the type of students the colleges are trying to recruit.

Today’s measurements of a school’s ability to serve low-income students largely disregard relevant pools. States with higher average incomes, such as Illinois and Connecticut, are punished for their relatively lower rates of Pell-eligible students enrolled at their public, flagship institutions; meanwhile, states with lower average incomes, like Montana and Maine, are rewarded by virtue of having relatively higher percentages of Pell-eligible students.

This chart from Hoxby and Turner’s study presents data on student representation from each income group at these states’ flagship, public four-year university. (Image via EducationNext.)

Hoxby and Turner reexamine how to identify low-income students. Rather than use national averages as many measurements do now, their approach compares each institution’s student population with its state’s income distribution across the “relevant pool” of students.

Their methodology breaks down the income distribution of the relevant pool into 20 groups, each comprising 5 percentiles. This allows for a more comprehensive view of enrollment rates from each income level at a given institution and not just from the extremes of the income distribution.

By comparing institutions with their state income distributions, observers are able to examine more holistically institutions’ support for low-income students based on the relevant pool of students from which they can draw. Although some states have lower rates of Pell-eligible students, their flagship institutions may still be working hard to enroll low-income students. In Connecticut, for example, students below the 40th income percentile are overrepresented on-campus while students from the 70th-100th percentiles are underrepresented. In other words, among students who are academic fits (roughly speaking) for the flagship institution, low-income students appear on-campus more often than higher-income students. This is despite the state’s overall income distribution, which skews higher than other states’.

Hoxby and Turner also note the importance of an institution’s mission when evaluating their student population in comparison to their relevant pool. They use Berea College, an NCAN member, as an example. Berea aims to support low-income students from the Appalachian region. Therefore, lower percentiles from their relevant pool will be overrepresented because of the institution’s mission to support these students before upper-income students.

National averages have dominated evaluations of public, flagship state institutions’ ability to support low-income students. Hoxby and Turner argue that these measurements are not effective and even have the inverse effects of what they intended. Inside Higher Ed reprinted two charts from the researchers’ work. The first, from 2008, shows two institutions’ enrollment compared to the relevant pool in 2008 and 2016. The second chart clearly shows a shift toward enrolling students just below the Pell eligibility threshold, which makes institutions look better on metrics emphasizing the percentage of Pell-eligible students enrolled. This behavior shift comes at the expense of students just above that threshold. Turner told Inside Higher Ed, “If you’re a student who’s just above [the Pell threshold], you still need just as much financial aid as somebody who is just below. We know that there’s distortion going on there."

Through Hoxby and Turner’s method, researchers and policymakers can better evaluate states’ flagship public institutions within the context of their own state, rather than in comparison to the national average. Each state has a different income distribution. It would be unfair to hold them all to the same metric, and doing so may have unintended consequences in other areas.