Fig. 1

Relationships Between Built Environment Characteristics and Obesity: Results From Six Multivariable Logistic Regression Models
Each of the six models includes demographic variables (age, sex, personal income, education) as adjustors
The ‘low’ tertile includes counties in the lowest third of each built environment variable, the ‘med’ tertile includes counties in the middle third, and the ‘high’ tertile includes counties in the highest third
The reference tertile for each model is indicated by an odds ratio equal to 1