A sibling study of stepchild well-being

Examining 33 indicators of well-being from the National Longitudinal Study of Adolescent Health, we conclude that stepchildren's inferior outcomes are not entirely explained by sample selection. Using sibling comparisons to control for unobserved family characteristics, we identify stepparent effects by comparing half-siblings in families in which one child has both parents and the other a parent and stepparent. Most estimated effects retain their sign after differencing across siblings, and a third remain statistically significant. The estimates' sensitivity to the choice of indicator suggests that studies based on a single measure of child wellbeing may be misleading.