Panel Paper: School Breakfast and Children's Attendance: New Evidence from Wisconsin

Saturday, November 4, 2017
Field (Hyatt Regency Chicago)

*Names in bold indicate Presenter

Judi Bartfeld, Lawrence Berger and Yiyu Chen, University of Wisconsin - Madison


The School Breakfast Program (SBP) is a critical part of the food assistance safety net, providing meals to almost 12 million children nationwide. The availability of the SBP has grown considerably over the past two decades, and is currently available in about 91% of schools, ranging from 74% to 100% across states. Along with growth in the reach of the program, there have also been changes in program delivery aimed at increasing participation, including alternative delivery models such as breakfasts served in the classroom (BIC) rather than prior to the start of school. Evidence suggests that the SBP is linked to benefits ranging from reduced food insecurity to improved nutrition to, in at least some studies, improvements in educational outcomes.

This paper uses harmonized administrative data from multiple agencies in Wisconsin to explore how children’s attendance in elementary school varies according to the availability of the SBP as well as the use of BIC as a delivery model. Existing studies in this area are typically based on survey data, on pilot expansions of breakfast programs or delivery models in a limited number of schools, or on school records from a single district. Here, we capitalize on an extensive administrative data infrastructure, housed at the Institute for Research on Poverty, that includes standardized test scores and attendance rates for all public school children statewide, as well as detailed information about current and past program participation and parental employment and earnings. Thus, we have a comprehensive statewide sample with detailed information about economic circumstances, with a large enough sample to examine how program impacts vary across subgroups. Wisconsin is an optimal location for this study because it routinely ranks near the bottom of states in SBP availability, yet has seen increases in SBP availability in recent years, as well as increased use of BIC, thus providing a natural experiment that supports comparisons both across schools and over time.

We compare students in schools that are late-adopters versus non-adopters of school breakfast, and use a variety of modeling strategies to explore the relationship between SBP and attendance. Across models, we find robust evidence that access to school breakfast has beneficial impacts on attendance—in terms of overall attendance rates as well as decreased risk of low attendance (<90% of days), with the magnitudes larger for BIC than for traditional breakfast models. Looking across children, the magnitudes are largest for SNAP recipients, for whom participation in school breakfast is most common, and smallest (often zero) for higher income children, for whom school breakfast participation is uncommon. Results provide substantial new information about the impacts of the SBP on school attendance, and are relevant to ongoing policy debates about the impacts of the program on children’s wellbeing.