Neighborhood Environments and Physical Activity: A Longitudinal Study of Adolescents in a Natural Experiment

Experimental and quasi-experimental evidence on the relationship between adolescents' physical activity and their physical activity environments is scarce. This study provides natural experimental evidence using within-person longitudinal variation in physical activity environments resulting from the compulsory re-assignment of military families to new installations, termed permanent changes of station.

Adolescents in Army families (N=749) reported usual weekly minutes of moderate and vigorous physical activity in 2013-2015. Objective measures of the physical activity environment, including the number of fitness and recreation facilities within 2 miles, were constructed for adolescents' neighborhoods using GIS methods. In 2017, individual-level fixed-effects models with and without a comparison group estimated the relationship between usual weekly minutes of physical activity and physical activity environments among permanent changes of station movers using within-person variation.

Increases in opportunities for physical activity were significantly and positively associated with increases in total (p<0.05) and vigorous physical activity (p<0.05) among adolescents who experienced permanent changes of station moves. The relationships were statistically significant for permanent changes of station movers living off-installation (p<0.05) and hence subject to greater variation in physical activity environments and those with more time to adjust to their new environments (p<0.05). Significant findings persisted when broader measures of physical activity environments were utilized.

The decline in physical activity and alarming obesity levels during adolescence suggest that this age may represent an important opportunity to address the obesity epidemic. This study provides evidence that increasing opportunities for physical activity may be an important pathway to improving their levels of physical activity and, consequently, obesity.