Determinants of Children's Mental Health: Relative Contributions When Accounting for Unobserved Heterogeneity, Endogeneity and Self-Selection

Abstract

While the neighborhood has been shown to contribute to children's mental health, the extent of this contribution is not yet sufficiently clear due to several methodological challenges. We investigate the relative contribution– sign, magnitude, and significance– of six neighborhood characteristics associated with children's mental health when accounting for unobserved heterogeneity, endogeneity, and self-selection in neighborhoods of different socio-economic status. Using a correlated random effects model in space and time and two waves of data extracted from the Lifelines Cohort Study of children between 7 and 15 years of age in northern Netherlands, we analyze both internalizing and externalizing behavior. Besides the importance of the environment for healthy child development, our empirical findings show that the magnitude and significance of neighborhood characteristics increase significantly, especially when we control for endogeneity of maternal family characteristics. Controlling for self-selection again reduces the magnitude slightly, but only for externalizing behavior. Controlling for unobserved child-invariant factors is also found to be significant, but not for unobserved heterogeneity.