New Findings on Welfare and Children's Development: Summary of a Research Briefing (1997)

return relatively quickly—close to one-half return within 2 years of leaving (Pavetti, 1996; Harris, 1994). Overall, over time, only about one-third of women who ever receive welfare will receive assistance for longer than five years. Yet a very different picture of welfare emerges if one looks only at families currently receiving welfare: about 90 percent of those currently receiving assistance will eventually spend more than 2 years on AFDC, and 76 percent will receive welfare for longer than 5 years (Pavetti, 1996).

Trying to determine how welfare receipt affects the course of children's development is therefore complicated by the fact that any amount of time a child spends on welfare is likely to be embedded in sizable amounts of time spent not on welfare. So for many children, the effects of time on welfare will be confounded by the effects of time not on welfare over the course of their development. Trying to determine the effects of welfare is further complicated by the fact that many parents both receive AFDC and work. These patterns of welfare receipt highlight the need to understand how the timing of welfare receipt affects children's development.

As you read this report, it is critical to recognize that the data that are available for examining questions about welfare assistance, family income, and children's development were collected before the enactment of major changes in welfare policy, and they provide a fairly limited range of children's developmental outcomes for analysis. When these data were collected, families' decisions to go off welfare (or not to accept welfare) were voluntary. Under welfare reform, the situation is quite different. One cannot assume that families who are now involuntarily removed from welfare as a result of time limits or changes in eligibility criteria will have the same characteristics, behave in the same ways, or have children with the same outcomes as those found for nonrecipients who have been studied to date. It is also possible that families who continue to receive assistance amid the current reforms are more disadvantaged and require more intensive services than those who do not, and may differ from families who are on welfare in the current data sets.

The research presented at the April 1996 briefing dealt directly with these challenges to trying to understand how welfare affects children. The investigators used state-of-the-art methods to disentangle the developmental effects of welfare from other factors that characterize recipient families, looked at questions of the timing of welfare receipt and its effects, and drew on the large longitudinal studies that constitute the most sophisticated national data sets now available for analysis. Although the findings on the effects on children of welfare receipt vary by data set, age, and outcome mea-

sured, the research begins to increase understanding of for whom, when, and in what contexts welfare receipt affects children, with special attention to children's and youth's intellectual performance, behavior problems, school completion, and non-marital birth rates.

The research briefing also included presentations of several current demonstrations involving random assignment that have the potential to provide major insights into the relative importance of welfare itself versus simple selection. These demonstrations are summarized in boxes that appear throughout this document.

This report is intended to supplement the literature on children and families living in poverty and receiving welfare, rather than to offer a comprehensive review of the research on family income, welfare, or poverty. It offers important, new, incremental evidence on these topics, from one of a series of research briefings of the Board on Children, Youth, and Families that present findings on various aspects of welfare and children's development.

The report is organized around the following three questions that emerged from the presentations and discussions at the briefing:

• What role does welfare receipt play in children's development?

• When and for whom does welfare have long-term impacts?

• How do children fare when their mothers combine work with welfare receipt?

The briefing also featured a rich discussion by state welfare administrators on assessing child outcomes in the context of evaluation and monitoring efforts. Comments from this discussion are presented in the last section of this report. Participants' ideas about topics that warrant further research are included at the end of each section.

Participants at the 1996 briefing echoed comments made at the 1995 briefing—that research has a critical role to play in sharpening understanding of the implications for children of changes in welfare and, more generally, of changes in families' economic status. But participants also cautioned against overstating the nature of the knowledge, and, in particular, were wary of the ability to make causal inferences about the relationship between welfare and children's development. Not only will it be critical to place the research presented at the briefing in the context of accumulated evidence based on different data and techniques, but it may also be necessary—due to the fast-moving pace of change—to modify the way research is carried out in order to measure the effects of the changes on children and to inform the continuing policy debate on welfare.