Background : Growing evidence supports the associations of neighborhood environments with physical activity. Understanding moderators of such associations is a critical gap in current research. Objective : The dissertation aims to examine a series of potential moderators of the associations between neighborhood environments and physical activity. Selection of moderators was based on ecological models.Methods : Secondary data from three separate studies were used. Chapter 2 used data from the Neighborhood Quality of Life study (NQLS), an observational epidemiological study of 2199 adults from selected neighborhoods in two U.S. regions. The moderators examined were psychosocial attributes about physical activity. Chapter 3 used data from the Senior NQLS study, a study of 880 seniors with similar design to that of NQLS. The moderator examined was the driving ability of seniors. Chapter 4 used data from the International Prevalence Study. Data were collected in 11 countries with standardized methodologies and instruments. The moderator examined was country. With each study, appropriate statistical models were selected based on the nature of data and the distribution of outcomes. Results : In Chapter 2, psychosocial attributes (self- efficacy, social support, enjoyment, benefits, and barriers) were found to be moderators of neighborhood environments with leisure walking as the outcome, but not with transport walking or accelerometer-based physical activity as the outcome. All interactions consistently supported stronger neighborhood environments-leisure walking associations among those with less favorable psychosocial attributes. In Chapter 3, driving ability of seniors was found to be a moderator with leisure walking as the outcome, but not with transport walking or accelerometer-based physical activity or sedentary behavior as the outcome. Patterns of interactions were consistent, suggesting neighborhood environments-leisure walking associations among driving-able seniors, but not non-driving-able seniors. In Chapter 4, the associations of physical activity/walking with land use mix, sidewalks, and bicycle facilities were more consistent across countries, suggesting generalizability. Associations involving other neighborhood attributes were more variable across countries, suggesting country as a moderator of the association. Discussion : There was evidence suggesting that psychosocial attributes, driving-ability, and country could modify associations between neighborhood environment attributes. Future studies should continue this inquiry with improved theories, conceptualization, and measurement instruments