Table 4.

Correlation Between Individual-Level BMI and County-Level Variables, North Carolina^a

County-Level Variable	Regression Modelb,c
	1	2	3	4
2000 Commute time	0.0847	0.0730	0.0733	0.0660
Commute time difference (2000 – 1990)	0.2719	0.1812	0.1791	0.1572
Retail gap per capita (NAICS code 445)d	0.0004	0.0002	0.0003	0.0002
Retail gap per capita (NAICS code 722)e	0.0004	0.0004	0.0004	0.0004
Combined retail gap per capita (NAICS codes 445 + 722)	0.0002	0.0002	0.0002	0.0002

Abbreviations: BMI, body mass index; NAICS, North American Industry Classification System.

^a Individual-level BMI was the dependent variable and county-level commute times and food retail gap per capita were independent variables. Individual covariates were age, age squared, sex, education, and race/ethnicity

^b Model 1: no additional covariates; model 2: individual covariates only; model 3: individual covariates + Rural to Urban Continuum Codes (RUCC) (24); model 4: individual covariates + RUCC + regional dummy variables.

^c P values ranged from <.001 to .048 using a t test, except those for Model 2 for retail gap per capita (NAICS code 445 [P = .08] and NAICS code 722 [P = .06]) and for Model 4 for retail gap per capita (NAICS code 445 [P = .06]).

^d Retail gap per capita calculated by subtracting county-level sales of products for a NAICS category in 2008 from county-level demand for products in that category in 2008. NAICS code 445 defined as stores that sell food and beverages from fixed point-of-sale locations, including supermarkets, grocery stores, convenience stores, meat markets, produce markets, and specialty food stores.

^e NAICS code 722 defined as food services and drinking places that prepare meals, snacks, and beverages to customer order for consumption on and off the premises, including full-service restaurants, limited-service eating places (fast-food restaurants), special food services, and drinking places.