Do active communities support activity or support active people

DESCRIPTION (provided by applicant):

Existing research shows promising results regarding the association between the built environment and physical activity. However, little is known about how built environment features are interrelated over time, and the predominant literature is cross-sectional and thus vulnerable to residential selection bias due to lack of control for individual choice of environments on the basis of their activity-related amenities. Objectives include (1) Description of built environmental features patterning across key life stages, and (2) Estimation of the longitudinal effects of the built environment on physical activity using several methods to control for and quantify residential selection bias due to: (2a) observed characteristics, (2b) time invariant unobserved characteristics, and (2c) time invariant unobserved characteristics. We will use data from a unique large scale Geographic Information System (GIS) that links community- and individual-level data in both space and time in two large ethnically diverse samples. The National Longitudinal Study of Adolescent Health (Add Health), a prospective cohort followed from adolescence to young adulthood, and the Coronary Artery Risk Development in young Adults (CARDIA), a prospective cohort followed from young adulthood, provide extensive health and behavioral data. A wide range of community-level factors include land use, street connectivity, recreation facilities, economic, climate, crime data, and others. Analyses will include: (1) Using descriptive methods and confirmatory factor analysis, we will describe patterning of built environment characteristics and how this patterning differs from adolescence to young and middle adulthood. Based on these findings, we will construct built environment summary variables to be used to estimate built environment effects on physical activity in subsequent aims. (2) By comparing estimates of the longitudinal effects of the built environment on physical activity using several methods for controlling for residential selection bias, we will quantify residential selection bias due to different types of factors: (2a) observed characteristics: compare random effects models (na?ve estimate) vs. propensity score weighting (control for observed characteristics), (2b) time invariant observed and unobserved characteristics: compare 2a estimates versus fixed effects models, and (2c) time invariant and variant, observed and unobserved characteristics: compare 2a-b estimates versus instrumental variables estimates. This study will improve the ability to make causal conclusions regarding the effects of the built environment on physical activity. Findings will improve the scientific knowledge base regarding "built environments that promote physical…health by encouraging healthy behaviors…," one objective of CDC's "Healthy People in Healthy Places" health impact goal. It also contributes to the "Healthy People in Every Stage of Life" goal by investigating health-promoting environments throughout several critical life stages.