SNAP Distribution Schedules and Student Standardized Test Score Performance

• In Event: SNAP Program Design: Implications for Participants and Policy

Abstract

Apart from well-documented associations with child-level conditions such as sleep (Eide & Showalter, 2012) and dietary intake (Burrows et al., 2017; Figlio & Winicki, 2005), one factor that may affect students’ standardized test performance among students in low-income households is receipt of public assistance from programs like the Supplemental Nutrition Assistance Program (SNAP), which provides cash-like vouchers to purchase food. SNAP is a major anti-poverty program, but benefits are distributed once monthly, leading to within-month resource instability, which may harm children as resources run low at the end of a benefit month. Previous work has linked this benefit cycle to within-month variation in food and other purchases (Castellari et al., 2017), diet quality (Kuhn, 2018; Todd, 2015), conflict in the home (Carr & Packham, 2021), and health and achievement (Arteaga et al., 2018; Bond et al., 2022; Gassman-Pines & Bellows, 2018). How the SNAP benefit cycle contributes to achievement in elementary and middle school, however, remains unclear. Addressing this important gap has the potential to influence both the design and implementation of standardized testing regimes and SNAP policies.  

In this study, we link state-year data on SNAP benefit schedules with restricted-access child-level data from the National Assessment of Educational Progress (NAEP) on the reading and math test scores of fourth and eighth grade students from 2005 through 2015. We assess whether time since receipt of SNAP benefits are associated with test scores and heterogeneity by type of test and student grade. Because both determinants of SNAP distribution timing (date of NAEP assessment and date of state SNAP disbursement) are exogenous, we consider time since benefit receipt to be as good as randomly assigned, but nonetheless control for school, year, and day-of-week fixed effects in all models. First, using our main sample of students in states with single distribution days, we test the effects of distribution schedules using a standard regression framework, controlling for various student-, school-, and state-level controls, using a sample of children receiving free lunch (<130% FPL), and assess heterogeneity by student grade and assessment type (math/reading). We repeat these analyses on our broader sample including states with multiple days of benefits distribution, using probability of benefits distribution on thte test day as our main predictor. We will conduct a range of sensitivity analyses to assess the robustness of results. 

Preliminary results suggest that days since SNAP benefit receipt is associated with students’ test scores; however, effects vary by grade. In fourth grade, effects were consistent with hypotheses: in the first half of the benefit month, test scores increased with days since SNAP benefit distribution before decreasing in the second half. Eighth grade test scores followed the opposite pattern, decreasing in the first half of the benefit month then increasing in the second. Dfferential effects by student grade may reflect age differences in family meals and shielding children from food insecurity. Effects are stronger for math than reading scores. Further, findings may suggest changes to state SNAP benefits administration schedules, benefit levels, and testing schedules or policies.

Authors

• Daniel P. Miller, Boston University
  Non-Presenting Co-Author

• Taryn W. Morrissey, American University
  Non-Presenting Co-Author

• Madeleine Maylath, American University
  Non-Presenting Co-Author

• Sruti Mohanty, Boston University
  Presenting Author