Major Papers


Obesity, quantile regression, fruits and vegetables, BMI, physical exercise


Objective: This study examines the relationship between BMI and some life-style variables, socio-economic status (SES) variables, and some socio-demographic variables related to behavior of individuals along different points of the BMI distribution by using quantile regression. Methods: A representative sample of 34,225 individuals of Canada form the Canadian Community Health survey 2014 is selected to conduct this study. Ordinary least squares (OLS) method is used at first to differentiate the results between conditional mean framework and conditional quantile framework. Quantile regression is estimated to analyze the heterogeneous relationship among fruits and vegetables, physical activity and BMI. Results: Analyses expose that fruits and vegetables intake and physical activities are negatively associated with BMI and statistically significant both for male and female. The estimates are larger in the higher quantiles for individuals. OLS overstates these associations at the lower quantile and understates at the higher quantile of the distribution. Conclusion: Findings of OLS that assumes equal responses may be misleading. The study finding suggest that effective dietary strategy and appropriate physical consciousness strategy may be helpful to reduce the risk of obesity and overweight.

Primary Advisor

D. Li

Program Reader

Y. Wang

Degree Name

Master of Arts



Document Type

Major Research Paper

Convocation Year