Many public programs serve poor children. By setting budgets, benefit levels, and program rules, policymakers decide how many children will receive benefits and which benefits they will receive. Making these choices in a rational way is difficult, given the noncomparability of different types of program benefits and the limited information available about the effects these programs have on poor children. This article suggests four criteria (efficiency, return on investment, incentives, and equity) for evaluating and comparing public programs for poor children, and provides an overview of the patchwork of information that is currently available about the effects of eight large federal programs using these criteria. Some broad themes emerge. First, several programs that target specific benefits directly to children have been shown to have positive effects on a range of outcomes. Second, even before the current round of welfare reform, the mix of federal support available to poor children had changed in a way that put more emphasis on providing benefits in kind. Finally, more must be learned about the effects of programs for poor children before sweeping policy recommendations can be made. This article concludes with policy recommendations that can be supported by the available evidence.