Neighborhood design and walking - A quasi-experimental longitudinal study

Background: Few studies have employed longitudinal data to examine associations between the physical environment and walking. Methods: Using cross-sectional (n=70) and longitudinal (n=32) data (collected 2003-2006), associations of neighborhood design and demographics with walking were examined. Participants were low-income, primarily African-American women in the southeastern U.S. Through a natural experiment, some women relocated to neo-traditional communities (experimental group) and others moved to conventional suburban neighborhoods (control group). Results: Post-move cross-sectional comparisons indicated that women in neo-traditional neighborhoods did not, on average, walk more than women in suburban neighborhoods. Race and household size were significant predictors of physical activity. Additionally, using longitudinal data, this study controlled for the effects of pre-move walking and demographics. Analyses examined the effects of environmental factors (e.g., density, land-use mix, street-network patterns) on post-move walking. Women who moved to places with fewer culs-de-sac, on average, walked more. Unexpectedly, increases in land-use mix were associated with less walking. Conclusions: Results suggest that neo-traditional neighborhood features alone (e.g., sidewalks, front porches, small set-back distances) may not be enough to affect walking; however, changes in street patterns may play a role.