**2. Materials and Methods**

In this study, we were interested in examining the effects of the FAFSA completion intervention on the submission of financial aid applications. Our primary challenge was isolating the FAFSA Completion Project's effects from the scholarship announcement. Both interventions were likely to impact financial aid application behavior, and it can be challenging to assess the contributions of each. As with most education research, we could not randomly assign students to treatment and control groups in this situation, and so, we relied on several comparisons to help establish the FAFSA project's effects. First, we compared FAFSA completion rates from Buffalo to those of comparable, mid-sized urban centers across New York State. The tuition guarantee applied primarily to New York State public colleges and universities, and so, this contrast helped to assess the extent to which financial aid application behavior changed in Buffalo compared to peer cities. It did not indicate the relative contribution of the tuition guarantee or the application support. However, it did illustrate that something different was happening in Buffalo, and it was not the result of changes happening across the state.

The second set of analyses compared the FAFSA completion behavior in BPS to their public charter school peers. The advantage of this comparison was that both BPS and charter school students were eligible for the tuition guarantee. However, in the first year of the project, the FAFSA Completion project only assisted students in BPS, meaning charter school students provided a useful counterfactual. In later years, we expanded the project to include charter schools. However, the first year allowed us to examine the extent to which the FAFSA project had an independent impact on financial aid application behaviors beyond the announcement of the tuition guarantee. It is important to note that the study reports the total number of applications completed per school rather than the percentage completed because it was difficult to identify the appropriate denominator. At the time of the intervention, all seniors and any juniors identified as eligible to graduate early could utilize the services of the FAFSA Completion project. However, not all eligible students graduated from high school, and not all graduates intended to go to college. So, the possible denominators would be the number of currently enrolled seniors, the number of high school graduates in the cohort, or the number of graduates who intend to enroll in college—a measure for which reliable data were unavailable.
