Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Economics (PhD)

Degree Level





Rodolfo Nayga

Committee Member

Arya Gaduh

Second Committee Member

Michael Thomsen


Childhood Obesity


The dissertation consists of three chapters exploring the effect of school program and neighborhood environment on childhood obesity outcome using individual panel data set of Arkansas public schoolchildren.

The first chapter (Section 2) investigates how the Fresh Fruit and Vegetable Program (FFVP), a program that provides funding for the distribution of free fresh fruits and vegetables to students in participating schools, affects childhood obesity. We combine matching methodology and difference-in-differences analysis to estimate the effect of the FFVP on childhood BMI outcomes. Estimates of the FFVP effect are sensitive to different matching methods. Methods that provide a good balance between treatment and control samples show that the FFVP program causes an economically meaningful reduction in the body mass index of participating children. Less strict matching methods yield insignificant results.

The second chapter (Section 3) measures the effect of fast-food restaurant density around the residences of Arkansas public schoolchildren on BMI outcomes. We use the distance from the child's residence to the nearest US highway or interstate highway as an instrument for the density of fast food restaurants. The results show that the exposure of fast food restaurants around the home environment does have significant and positive effects on children's BMI z-scores. Our results also indicate that some subpopulations -- children who are more affluent, rural, non-minority and female -- are disproportionately affected by fast food proximity.

Finally, the third chapter (Section 4) analyzes the effect of neighborhood parks around residences of northwestern Arkansas children on BMI outcomes. Our dataset covers the 2004 through 2007 period. To build comparative groups, we employ propensity score matching to measure the average treatment effect on the treatment group. The results indicate that proximity of neighborhood parks from the residence have a significant and negative effect on children's BMI z-scores in both the rural and urban areas, with some heterogeneity in the effects across gender. Specifically, our results show that girls in urban and rural areas are significantly influenced by neighborhood parks. The park effect is significant for boys in rural areas but not for boys in urban areas.