There is growing interest in leveraging Title IV student aid to improve college completion. Advocates have proposed linking funding in the Pell and Campus-Based Programs to measures of college performance. However, to do so in an equitable and efficient manner, raw measures of college output, such as rates of graduation and academic progress, must be adjusted to reflect differences in factors that determine those rates--inputs such as college mission, student characteristics, and other constraints. Failing to account for inputs when measuring and evaluating college performance unfairly penalizes colleges that are efficiently serving large numbers of low-income students, particularly colleges doing so with limited resources. To demonstrate the impact that serving low-income students can have on raw measures of college performance, this bulletin explores the relationship between 6-year graduation rate and three inputs: the percentage of first-time students who are Pell recipients, average test score of the student body, and level of endowment per student. The analysis finds that these three inputs are powerful determinants of 6-year graduation rates at nonprofit 4-year public and private colleges. The empirical findings and policy implications in this brief are not meant to imply that 4-year colleges serving a large number of Pell recipients should not be evaluated or that graduation rates cannot or should not be improved. All colleges must meet minimum standards of performance, implement policies and practices shown to be effective in improving graduation rates at peer colleges, and provide adequate consumer protections, including accurate information for students and parents. The findings noted in this document have major implications for student aid policy, including HEA reauthorization: (1) Using raw output measures, such as rates of graduation or student academic progress, in the awarding or allocation of Title IV student aid will harm low-income students and the colleges that serve them; (2) To prevent such harm, output measures must be adjusted to adequately reflect differences in inputs, in particular, college mission, student characteristics, resources, and factors beyond colleges' control; (3) Because estimates of value added are vulnerable to both modeling and data limitations, such measures should not be used to support high-stakes decisions in either the Pell or Campus-Based Programs; and (4) The best approach to improve graduation rates at 4-year public and private colleges is to use well-designed case studies to identify policies and practices shown to be effective at peer colleges and incentivize colleges to implement them.In race-to-the-top competitions for additional funds, colleges should be required to compete only against peers with highly comparable inputs. These inputs include the percentage of students who are Pell recipients, average test score, endowment per student, and other factors shown to have a statistically significant relationship to the output measure in question.