Panel Paper: The Effects of Physical Education on Child Body Weight and Academic Achievement: New Evidence from the ECLS-K:2011

Friday, November 9, 2018
8222 - Lobby Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Kathryn Rouse and Steven Bednar, Elon University


In this paper, we provide the first post-NCLB era estimates of PE effects on child body weight and achievement using data from the Early Child Longitudinal Survey Kindergarten Class of 2010-2011 (ECLS-K: 2011). The ECLS-K:2011 survey provides data on a cohort of kindergarteners beginning their schooling about a decade after the students in the original ECLS-K cohort used in the earlier studies. By the time students in this latter cohort entered kindergarten NCLB had been in place for eight years and the effects of NCLB should have been fully implemented and realized. As a baseline, we first estimate OLS models including in the models a rich set of observable control variables. We estimate the OLS models with and without state fixed effects. To control for endogeneity concerns, we then follow the instrumenting strategy used by both Cawley et al. (2013) and Dills et al. (2011). We instrument weekly PE time using state PE regulation data from the Classification of Laws Associated with School Students (CLASS)[1] compiled by the National Institute for Health. This allows us to identify the impacts of PE separate from other school attributes that might influence a student’s actual time spent in PE. Finally, we exploit the longitudinal nature of the ECLS-K: 2011 and estimate a panel model, using individual fixed effects to control for potential within-student unobserved heterogeneity.

While students in the latter ECLS-K:2011 cohort were exposed to increased accountability pressures by NCLB, we find average PE time is higher than that reported by either Dills et al. (2011) or Cawley (2013) using the original ECLS-K cohort. Despite the difference in PE time across the two cohorts, our results suggest weekly PE time has little effect on the average child’s body weight over the span of kindergarten to third grade. These BMI and obesity results are consistent with Cawley et al. (2013)’s results for kindergarteners through third graders, but differ somewhat from what they find in their fifth grade sample. However, when examining whether there are differential effects by gender, we find evidence to suggest PE increases the likelihood girls are overweight or obese in kindergarten. This somewhat puzzling finding may be a result of PE being a substitute for other forms of physical activity for girls (Cawley et al., 2013). Our findings on achievement outcomes are consistent with those reported by both Dills et al. (2011) and Cawley at al. (2013). Across all models and for both genders, we find there is no economically meaningful impact of PE on either student reading or math achievement gains as measured by scores on standardized cognitive exams administered by the ECLS-K: 2011. These results confirm those found in previous studies, suggesting increased PE time does not come at the expense of diminished academic performance.

[1] https://class.cancer.gov/about.aspx (Accessed February 5, 2018).

Full Paper: