Could strength of exposure to the residential neighbourhood modify associations between walkability and physical activity?