Geographic determinants of individual obesity risk in Spain: A multilevel approach

•Both group and individual effects play a key role in understanding BMI/obesity.•Our proxies of social environment have a statistically significant effect on BMI.•MLM offers a more accurate description of relationships in clustered data than OLS.•Some effects can only be observed by examining interactions between levels.•Environmental and regional characteristics influence BMI and obesity prevalence.

This paper seeks to understand the determinants of individual body weight status and obesity risk in Spain by concurrently examining individual and regional characteristics. The data are drawn from the National Health Survey of Spain for the year 2011-2012 (INE-National Statistical Institute of Spain) and contain information for a representative sample of 12,671 adults across 50 provinces in Spain. A multilevel analysis is carried out to examine the determinants of individual weight status and obesity, controlling not only for the individual effects and those of the immediate environment but also for the broader setting to which individuals and their immediate environment belong. Our findings suggest that attributes from all three levels of analysis have an effect on individual weight status and obesity. Lack of green spaces and criminality taken as proxies of the social environment positively affect individual and women's BMI and obesity, respectively.