| dc.description.abstract | Most studies of family structure and child outcomes conclude that stepchildren fare little better than children in single-parent families, and substantially worse than children in intact families. Is this because adults treat biological children differently from stepchildren, or are stepchildren's inferior outcomes "explained" by selection into stepfamilies? To address this question we examine a rich array of child outcomes and parental investment measures drawn from the National Longitudinal Study of Adolescent Health (Add Health). Sibling comparisons control for unobserved parental and household characteristics, and the effect of living with a stepparent is identified by comparing outcomes across half-siblings in households in which one child is living with both biological parents and the other is living with their common parent and a stepparent. The results support the differential-treatment hypothesis. Despite the radical reduction in sample size associated with differencing across siblings, a third of the stepparent effects remain statistically significant. Over four-fifths of the point estimates retain their sign after differencing, and as many increase as shrink. Selection bias may be present in stepparent effects measured between families, but in the Add Health data the bias is as likely to be negative as positive. Furthermore, the sensitivity of the results to the choice of indicator suggests that the story cannot be told in full by studies focusing on only one or two indicators. | |
