Following the ongoing increase in nonmarital fertility, policy makers have looked for ways to limit the disadvantages faced by children of unmarried mothers. Recent initiatives included marriage promotion and welfare-to-work programs. Yet policy might also consider the promotion of three generational households. We know little about whether multigenerational households benefit children of unwed mothers, although they are mandated for unmarried teen mothers applying for welfare benefits. Multigenerational households are also becoming increasingly common. Thus, using data from the Panel Study of Income Dynamics (N = 217), this study examines whether grandparent-headed coresidential households benefit preschool-aged children's school readiness, employing propensity score techniques to account for selection into these households. Findings reveal living with a grandparent is not associated with child outcomes for families that select into such arrangements but is positively associated with reading scores and behavior problems for families with a low propensity to coreside. The implications of these findings for policy are discussed. (author abstract)