Skip to main content
Lifestyles, Demographics, Dietary Behavior, and Obesity: A Switching Regression Analysis
Health services research (2009)
  • Steven T Yen, University of Tennessee - Knoxville
  • Zhuo Chen, Centers for Disease Control and Prevention
  • David B Eastwood, University of Tennessee - Knoxville
Objectives. To investigate the effects of lifestyles, demographics, and dietary behavior on overweight and obesity. Data Source. Continuing Survey of Food Intakes by Individuals 1994-1996, U.S. Department of Agriculture. Study Design. We developed a three-regime switching regression model to examine the effects of lifestyle, dietary behavior, and sociodemographic factors on body mass index (BMI) by weight category and accommodating endogeneity of exercise and food intake to avoid simultaneous equation bias. Marginal effects are calculated to assess the impacts of explanatory variables on the probabilities of weight categories and BMI levels. Principal Findings. Weight categories and exercise are found to be endogenous. Lifestyle, dietary behavior, social status, and other sociodemographic factors affect BMI differently across weight categories. Education, employment, and income have strong impacts on the likelihood of overweight and obesity. Exercise reduces the probabilities of being overweight and obese and the level of BMI among overweight individuals. Conclusion. Health education programs can be targeted at individuals susceptible to overweight and obesity. Social status variables, along with genetic and geographic factors, such as region, urbanization, age, and race, can be used to pinpoint these individuals.
  • Lifestyle,
  • obesity,
  • ordinal probit,
  • overweight,
  • switching regression
Publication Date
Fall August, 2009
Citation Information
Steven T Yen, Zhuo Chen and David B Eastwood. "Lifestyles, Demographics, Dietary Behavior, and Obesity: A Switching Regression Analysis" Health services research Vol. 44 Iss. 6 (2009)
Available at: