Two weeks ago, the New York Times published its second annual “College Access Index,” which measures socioeconomic diversity and accessibility at America’s highest performing colleges and universities. Adjusting its methodology from last year, the 2015 College Access Index incorporated each institution’s average Pell Grant recipient graduation rate into its score (a new addition), along with the institution’s Pell enrollment rate and net price for low income students (both of which were used in the index’s 2014 iteration). Last week, Education Trust provided similar metrics for a much broader swath of colleges and universities, and released a summary and analysis of Pell enrollment and six-year graduation rates at 1,150 four-year institutions. The report found a national gap of 14 percentage points amongst Pell and non-Pell completers, and an average gap of 5.7 percentage points at the institutional level. Its authors argued that, in order for Pell dollars to be well-spent, “more attention must be focused on institutional roles and responsibilities” when it comes to graduating students who receive Pell Grant aid.

This attention paid to Pell graduation rates is important, but, until the release of the Ed Trust report and data last week, was hard to come by. Though institutions that receive Federal Title IV funds are required to report graduation rates for an entire cohort to the federal government, graduation rates for Pell Grant recipients are not made publicly available, and are only disclosed when requested. This means, to date, much of what we do know about Pell Grant recipients has to do with the shares that an institution enrolls rather than the shares that it graduates. This has led to much discussion about how to enroll more low-income students at institutions with universally high graduation rates, but contributed less to efforts to improve completion rates for low-income students at institutions where larger shares of them enroll and often struggle.  Alternatively, much of what we know about graduation rates—or what potential students and their families can know about graduation rates– reveals little about how well low-income students fare at specific institutions.

Though Ed Trust’s analysis revealed relatively small institutional gaps in Pell and non-Pell completion rates on average, its examination of outliers and its comparison of similar schools with vastly different graduation rates for Pell Grant recipients are provocative. These comparisons suggest that, even when factors like enrollment, net price, racial diversity, and median SAT score are held constant, institutional policies can play a large role in facilitating the success of low-income students.

Of course, this begs some questions: what have institutions with small or no graduation gaps done to ensure success for low-income students? What sort of resources does this entail? Which of these approaches might be more broadly applicable to other institutions?

The Ed Trust report looks underneath the hood at a couple of well-performing schools, but the availability of this data presents a much richer opportunity for research. Ithaka S+R’s own work on this topic suggests, perhaps not surprisingly, that different types of institutions find success with varied strategies. For example, Franklin & Marshall College, a small-liberal arts school that has a .5% six-year graduation gap between Pell Grant and non-Pell Grant recipients, eschews interventions that specifically target low-income students, and instead relies on a holistic recruiting and support strategy that benefits students of all backgrounds. Meanwhile, Georgia State University, which enrolls 25,000 students (50% Pell Grant recipients) and graduates Pell and non-Pell grant recipients at the same rate, relies heavily on data and predictive analytics to target specific interventions for at-risk students, many of whom are low-income.

New data and analysis of Pell Grant recipient graduation rates will be crucial to close the graduation gap between low-income students and their wealthier peers. A critical next step will be digging into this data, examining successful schools, and learning more about what different sorts of institutions—with different levels of funding, different demographics, and different tiers of selectivity—have done to ensure more equitable outcomes for students of all backgrounds.