To examine both genetic and environmental influences on children's behavior problems in households defined by marital status and sibling relatedness, this study applied behavioral genetic methodology to four groups totalling 1524 sibling pairs drawn from 796 households: (1) two-parent full siblings, (2) two-parent half siblings, (3) mother-only full siblings, and (4) mother-only half siblings. Model-fitting procedures found that within-group variation on four subscales from the Behavior Problems Index was best explained by a model including both genetic and shared environmental factors. This model was then fit to the behavior problems means of the four groups. Its successful fit to these mean structures suggested that mean-level differences between groups were explained with the same influences that accounted for within-group variation. Genetic influences accounted for 81% to 94% of the mean-level difference in behavior problems between the two-parent, full sibling and the mother-only, half sibling groups. In contrast, shared environmental influences accounted for 67% to 88% of the mean-level difference in behavior problems between the two-parent, full sibling and mother-only, full sibling groups. The genetic influences are interpreted in terms of genetic self-selection into family structures.