Non-linear effects of environmental and social exposures and interactions on early childhood academic achievement

Abstract

Non-linear effects of environmental and social exposures and interactions on early childhood academic achievementIntroduction. Previous research has shown that environmental and social exposures are associated with academic achievement. However, the interactions between environmental mixtures and social conditions, and the scales of the associations with academic achievement remain unclear.In our study we seek to:1.Examine childhood co-exposures to environmental and social conditions; and2.Capture the non-linear effects of environmental and social exposures (and interactions) that occur across the life course on academic achievement, as measured by end of grade (EOG) test scores.Methods. We use North Carolina birth records from the year 2000, linked to data on lead, education, neighborhood deprivation index (NDI), residential isolation (RI), air quality (PM2.5), and temperature. We use EOG standardized test scores in reading and mathematics for 4th grade as a proxy for academic achievement. We apply a generalized additive model (GAM) to measure the scales of associations, as well as interactions, between environmental and social exposures and educational outcomes.Results: Main effects: environmental exposure in trimesters 2 and 3, social conditions, birthweight percentile for gestational age, blood lead level, mother’s race/ethnicity group, and NDI are significant predictors of educational outcomes. Interactions between RI/NDI and environmental exposures, as well as blood lead level and RI, are also significant predictors. We also find non-linear effects of environmental exposures in trimesters 2 and 3 on educational outcomes, with higher pollution levels predicting worse test scores.Conclusion. Childhood blood lead level, RI, and NDI are associated with academic achievement both individually and interactively. Also, we find clear evidence of non-linear effects of environmental and social exposures (and interactions) on educational outcomes, which has implications for how multiple exposures should be modeled in future research.