School Participation in the Community Eligibility Provision and Obesity Prevalence: Evaluation of Treatment Effect Heterogeneity

• In Event: Differential Implementation and Impacts of Policies to Combat Poverty and Food Insecurity

Abstract

Introduction/Background: School meals are childrens’ most nutritious food source – particularly for those from low-income households. Previous research found that school participation in the Community Eligibility Provision, a universal free school meals policy, was associated with a net reduction in obesity prevalence in California schools.

Purpose/Research Question: This study aims to assess whether the association between school participation in the Community Eligibility Provision and obesity prevalence differs by the following key demographics: sex, race and ethnicity, socioeconomic status, and rurality.

Methods: California public schools eligible to participate in the Community Eligibility Provision between 2014 and 2019 were included in the analyses. Obesity prevalence data came from 5th, 7th, and 9th grade students who participated in the Physical Fitness Testing program; child obesity was defined as body mass index at or above the 95th percentile for age and sex. We conducted stratified analyses using a novel difference-in-differences approach for staggered policy adoption to obtain post-policy treatment effect estimates by demographic group. A secondary analysis assessed associations by race/ethnicity among plurality Hispanic schools.

Results/Findings: A total of 3,531 schools were included in the analyses. Schools were half male and predominantly Hispanic and economically disadvantaged. Sixty percent (60%) of schools were elementary schools, 46% were urban and 40% were suburban. At the group level, school participation in the Community Eligibility Provision was associated with a 1.28 percentage point net decrease in obesity prevalence among White students (95% CI: -2.41, -0.15, p=0.03) and a 0.61 percentage point net decrease among female students (95% CI: -1.14, -0.09, p=0.02). At the school level, policy participation was associated with a 0.69 percentage point net decrease in obesity prevalence among plurality Hispanic schools (95% CI: -1.19, -0.20, p=0.01), a 0.52 percentage point net decrease among high-poverty schools (95% CI: -1.11, 0.06, p=0.08), and a 0.66 percentage point net decrease among suburban schools (95% CI: -1.42, 0.09, p=0.09). The observed association in plurality Hispanic schools may have been driven by White students (1.40 percentage point net decrease; 95% CI: -2.87, 0.08, p=0.06).

Conclusion/Implications: Among eligible California schools, participation in the Community Eligibility Provision was associated with a net decrease in obesity prevalence among White students, female students, and schools that were high-poverty, suburban, and had a plurality of Hispanic students. The latter association may have been driven by White students within these schools. Future research should assess mechanisms potentially contributing to these associations.

Author

• Anna Localio, University of Washington
  Presenting Author