Can Geographically Weighted Regression improve our contextual understanding of obesity in the US? Findings from the USDA Food Atlas

• We examined the influence of food environment on obesity in the US.
• We also included demographic and socio-economic factors at a county level.
• GWR was used to investigate the heterogeneous processes.
• Ratios of convenience-to-grocery stores were positively associated with obesity rate.
• The association between obesity and explanatory variables varied over space.

There is growing interest in the role the food environment as well as demographic and socio-economic factors play in the prevalence of obesity in the US. Existing empirical evidence examining the association between the food environment and obesity risk, however, remains equivocal. We hypothesized that spatial heterogeneity may account for the conflicting results. Using Geographically Weighted Regression, we examined how the associations between the food environment, and demographic and socio-economic variables associated with obesity vary over space at the county level in the US. The analysis shows that higher ratios of convenience-to-grocery stores, poverty rate, and urban environments were positively associated with obesity risk in the US. Conversely, areas with better physical environments were negatively associated with obesity risk. Most importantly, the association between obesity and all major explanatory variables in our analysis significantly varied over space.