ABSTRACT Unhealthy diets are associated with numerous deleterious health outcomes. Yet, most children do not consume a healthy diet, in particular Black/African American and Hispanic children, and children of low socio-economic status (SES). Previous studies have found that community characteristics influence diet but the effectiveness of community programs and policies (CPPs) to influence diet is not consistent across communities by racial/ethnic composition and community SES. The contextual food environment is an important consideration in understanding the effectiveness of CPPs on diet and may explain, in part, the differential effect of CPPs across subgroups. However, very few studies have examined what interventions are most effective for improving diet within community settings and there are even fewer national studies of how contextual factors influence food environment intervention outcomes. The proposed study will leverage existing NHLBI datasets with information on CPPs, contextual food environments, and child, household, neighborhood, school, and community characteristics that can affect child diet. The study includes over 4,500 children in grades K – 8 from diverse racial/ethnic and SES backgrounds in 130 communities in the United States. Nearly 10,000 CPPs were documented from over 1,400 key informant interviews within communities. Contextual food environment measures (e.g., density of supermarkets) in 1, 3, 5, and 8 km buffers around the home, and in a school-centric buffer, were derived from over 58 million business listings. The specific aims in this study are to 1) determine the effect of CPPs and the contextual food environment on diet including intake of energy-dense foods of minimal nutritional value; total added sugar intake; sugar intake from SSBs; fruit and vegetable intake; whole grain consumption; and fast food consumption, and 2) determine whether differences persist by race, ethnicity, and SES in the relationship between CPPs and diet once contextual food environment data are considered. To assess our aims, a generalized linear mixed model with random intercepts and slopes will be estimated controlling for clustering at the community level. The proposed study is cost-effective in that it entails secondary data analyses using linked data from two previous studies. The study is unique in its use of CPPs and contextual food environment measures at multiple geographic scales in a large, diverse national sample of children along with control variables at the child, household, neighborhood, school, and community levels. The findings are expected to help shape how future community investments in programs, interventions, and policies are tailored and implemented to improve children’s diets and to address diet-related health disparities early in life.