Air Pollution and Children's Cognitive and Health Outcomes

Air pollution remains a serious public health concern despite a general decline of pollutants since the amended Clean Air Act of 1990.  Half of the U.S. population still reside in counties that have unhealthful levels of ozone or particle pollution.  Children are more vulnerable to the adverse effects of ambient air pollution due to more time spent outdoors, higher incidence of outdoor physical activity, more rapid breathing (relative to adults), and still developing biological and immune systems.  A substantial body of literature has linked exposure to criteria pollutants with adverse health outcomes in children including respiratory systems, asthma exacerbations and hospitalizations, preterm birth, infant mortality, and deficits in lung growth.  However, much of this literature is based on cross-sectional studies and involves geographically-limited samples, limiting causal inference and generalizability. 

In this study, we examine the relationship between air pollution and children’s health and cognitive outcomes using panel data from the Early Childhood Longitudinal Study – Kindergarten Class (ECLS-K). First, we examine the influence of air pollution on children’s health outcomes including respiratory health problems, general health status, and school absences.  Second, we examine the relationship between air pollution exposure and cognitive outcomes including cognitive learning disabilities and cognitive development.  These cognitive outcomes allow us to further examine preliminary evidence identifying a potential link between cognitive dysfunction and chronic exposure to significant concentrations of air pollutants.  Finally, we examine whether these estimated relationships between air pollution and children’s outcomes vary with observable child characteristics including gender, race-ethnicity, socioeconomic status, participation in outdoor activities, and health characteristics.

We use data from the kindergarten, 1st, 3rd, 5th, and 8th grade waves of the ECLS-K, which is a panel dataset on a nationally representative sample of kindergarteners in the U.S. The ECLS-K data contain census tract identifiers for each child’s home and latitude and longitude for their school location.  We link individual level data from the ECLS-K to the Environmental Protection Agency’s (EPA) criteria pollutant data collected through 4000 monitors across the country.  We address the endogeneity of pollution exposure in several ways. First, we estimate child fixed effect models to control for time-invariant unobserved heterogeneity that may bias estimates of pollution effects. Further, we exploit the rich information available in the ECLS-K to implement flexible propensity score weighting and instrumental variables techniques to address potential selection and confounders that vary with pollution exposure.