CONTEXTUAL INFLUENCES ON OBESITY PREVALENCE: A SPATIALLY EXPLICIT ANALYSIS

Open Access
Author:
Black, Nyesha Cheyenne
Graduate Program:
Sociology
Degree:
Master of Arts
Document Type:
Master Thesis
Date of Defense:
September 30, 2011
Committee Members:
  • Stephen Matthews, Thesis Advisor
  • Stephen Augustus Matthews, Thesis Advisor
Keywords:
  • health disparities
  • geographically weighted regression (GWR)
  • obesity
Abstract:
This study examines contextual influences on obesity prevalence in coterminous counties in the United States. Using a variety of secondary data sources, I constructed a dataset with a rich set of county-level contextual variables. This study employs an ecological, spatially explicit perspective by exploring the influence of macro-level processes and the environment on health disparities. Traditional regression methods (OLS), along with exploratory spatial data analysis (Moran’s I) and geographic weighted regression (GWR), were utilized to thoroughly examine the relationship between obesity and a rich set of predictors. Results from traditional regression methods show that rural vs. urban residence does not significantly contribute to differences in obesity prevalence by county; whereas minority composition, features of the built and natural environment, and physical inactivity among adult residents are all significantly associated with county-level obesity prevalence. Also, county-level income inequality was found to provide a protective barrier against higher obesity prevalence. This finding suggests that the relationship between relative deprivation and health should be further explored in the health disparities literature. Furthermore, GWR confirms that place matters and the relationship between contextual influences and obesity prevalence varies substantially across place. GWR also provides an empirical basis for the public health community to design interventions that effectively target predictors of obesity at the local level.