It is not clear if relationships between GIS obtained environmental features and physical activity differ according to the method used to code GIS data.
Physical activity levels of 210 Boy Scouts were measured by accelerometer. Numbers of parks, trails, gymnasia, bus stops, grocery stores, and restaurants within the commonly used 400 m and 1-mile (1609.3 m) buffers of subject residences and distance to the nearest feature were calculated. Residential density, connectivity, and crime rate were calculated. Regression models with minutes of sedentary, light, or moderate-to-vigorous activity as dependent variables and environmental and demographics as independent variables were run with backward deletion of environmental variables.
Park, crime, and gym variables were associated with physical activity, but relationships varied according to whether a 400 m, 1 mile, or nearest criteria was used.
Environmental variables were associated with the physical activity of adolescent males, but the association was method dependent.