摘要：The National School Lunch Program (NSLP) has been criticized for providing high calorie lunches to over 31 million school-age students. With nearly 17 percent of children considered obese, the relationship between NSLP participation and childhood obesity has been studied extensively without conclusive evidence. This paper presents a review of the literature thus far and examines identification issues when estimating the average treatment effect or local average treatment effect of the NSLP. Using data from the National Health and Nutrition Examination Survey (NHANES), the author finds that when restrictive parametric assumptions are removed, participation in the NSLP reduces the probability of being obese in school age children.
关键词：National School Lunch Program;Obesity;Regression Discontinuity;Monotone Instrumental Variable
