Given growing disparities in college enrollment by household income, policymakers and researchers often are interested in understanding whether policies expand access for low-income students. In this brief, we highlight the limitations of a commonly available measure of low-income status--whether students receive a federal Pell grant--and compare it to new data on enrollment by income quintile to evaluate a recent policy effort within elite colleges aimed at expanding access. We demonstrate that Pell is a rough measure of low-income status and that without more detailed data on colleges' economic diversity, policy evaluations focusing on existing Pell data will suffer from measurement error and potentially miss enrollment effects for moderate- and high-income students.