Primary Submission Category: Place/Communities
Do Food Environments Shape Children’s Diet? Weak Evidence from a Poverty Reduction RCT
Authors: Samantha Gailey, Tim A. Bruckner,
Presenting Author: Samantha Gailey*
Poor diet remains a leading cause of chronic disease. In the US, low-income children adopt less healthy diets and suffer disproportionately from related cardiometabolic disorders. For example, obesity affects 1 in 4 low-income children, compared to 1 in 10 high-income children. A growing literature examines whether obesogenic food environments contribute to nutritional inequality, given evidence that low-income families are differentially concentrated into neighborhoods characterized by limited availability of healthy foods (i.e., food deserts). However, the role of food deserts in health inequities appears equivocal, owing in part to inconsistent measures and the predominance of cross-sectional designs. Here, we use a rigorous, randomized controlled trial (RCT) of poverty reduction among 1000 low-income families to test whether improvements in the food environment mediate the effect of unconditional cash transfers on children’s diet. We first replicate prior work showing a protective effect of treatment—i.e., random assignment to the “high cash” group (in which mothers of newborns received $333/month) vs. the control “low cash” group ($20/month)—on children’s diet at age 2. We then leverage inverse odds ratio weighting (IORW) to estimate natural direct and indirect effects using 3 popular food environment measures: (1) the Modified Retail Food Environment Index (tract level); (2) the number of food retailers accepting SNAP benefits (1-mile buffer around home); and (3) low-food-access areas designated by the Food Access Research Atlas (tract level). Results show null indirect effects for all 3 food environment measures. Taken together, improvements in the food environment (i.e., the combined mediating effect of all 3 measures) account for approximately 24% of the total effect of treatment on diet, but this combined indirect effect fails to reach statistical detection. Overall, we find weak evidence that improvements in the food environment, defined by 3 popular measures, mediate the effects of a poverty reduction trial on children’s diet.