| Background. Physical activity (PA) may be influenced by environmental factors. As part of the socioeconomic environment, gasoline price is a key component of the cost of driving and may influence individuals' transportation modes and PA; as part of the built-environment, street attributes such as connectivity are hypothesized to be supportive of PA, particularly street-based PA (SBPA). Methods. This research used secondary data from CARDIA study, a prospective cohort of young adults (N=5115 at baseline, 1985--86) followed through 2000--01 with three repeated examinations. Based on a PA history questionnaire administered at each examination, we calculated PA scores in exercise units (EU) by intensity and frequency of 13 PA categories, and characterized SBPA as total frequency of walking, bicycling, and jogging/running. The individual-level CARDIA data were spatially and temporally linked to multiple environmental datasets by participants' time-varying residential locations, using Geographic Information Systems technology. This dissertation follows two aims. Aim 1 consists of analysis examining longitudinal association between inflation-adjusted, county-level gasoline price and PA, using a random-effect longitudinal regression model and two-part marginal effect models. Aim 2 is to investigate longitudinal association between neighborhood street attributes (intersection density, link-node ratio, and characteristics of local roads) and SBPA, using the two-part marginal effect modeling, by urbanicity and gender. Results. A... |
