The welfare system has been transformed over the past two decades, notably through the introduction of stricter work requirements and time limits on cash assistance in the 1990s. At the same time, government at both the federal and the state level invested in offering financial work supports of unprecedented scope to low-income parents. A top priority on the national agenda was to understand how these changes in the safety net — and sharp rises in employment, particularly among single parents — are affecting children.

Long before the federal welfare reforms of 1996 drew attention to this question, MDRC — in collaboration with the U.S. Department of Health and Human Services and other government, foundation, and research partners — was systematically gathering data about children in its studies of pioneering efforts to promote employment among low-income parents. These studies laid the groundwork for the Next Generation project — a collaboration that began in 1999, involving MDRC, several other leading research institutions, and the project’s foundation funding partners. Drawing on 10 studies by MDRC of welfare, antipoverty, and employment policies piloted as early as the 1980s and widely adopted since then, Next Generation is synthesizing and reanalyzing data to learn how programs targeted at parents' economic circumstances affect children’s school performance, behavior, and health, as well as measures of family well-being.

Agenda, Scope, and Goals

The Next Generation project aims to inform policymakers, practitioners, and scholars by identifying lessons that cut across evaluations of individual welfare, antipoverty, and work programs. With a focus on how such programs can influence children’s and families’ well-being through their effects on employment, income, and child care, the project addresses questions such as:

  • How do programs that share certain effects on parents’ economic outcomes — for instance, that boost employment or income — affect children? What combinations of economic effects harm or benefit children most?
  • How do different types of child care assistance shape patterns of child care use among low-income parents? What are the consequences for children’s well-being?
  • Do the programs affect children of different ages differently? If so, how?
  • What can we learn from in-depth interviews with families subject to welfare reform about the daily challenges they face and how public policies can address their needs?

Design, Sites, and Data Sources

The Next Generation project draws data and perspectives from 10 rigorous studies conducted by MDRC, including the Project on Devolution and Urban Change and random assignment evaluations of the following programs:

  • Canada’s Self-Sufficiency Project
  • Connecticut’s Jobs First
  • Florida’s Family Transition Program
  • Indiana's Welfare Reform Evaluation, conducted by Abt Associates
  • Iowa's Family Investment Program, conducted by Mathematica Policy Research, Inc.
  • Los Angeles Jobs-First Greater Avenues for Independence
  • New Hope Project
  • Minnesota Family Investment Program
  • National Evaluation of Welfare-to-Work Strategies
  • New Chance
  • Vermont’s Welfare Restructuring Project

Because of the variety of policy approaches, data sources, and methodologies represented, Next Generation can delve deeper into the programs’ effects on children and families than could any single study. The project derives statistical reliability from the large size of its aggregate database. Together, the underlying evaluations encompass tens of thousands of parents and children across North America.

Owing to the richness of the underpinning research, the Next Generation team can make new discoveries without collecting new information. All the findings rest on secondary analysis of primary data, which include administrative records of employment and public assistance receipt, surveys of parents and teachers, and ethnographic interviews. The project’s synthesis approach is facilitated by the use of comparable measures of children’s and families’ well-being in many of the studies.