We propose that the Panel Study of Income Dynamics (PSID) be one of the key data bases in the network and that Greg J. Duncan, a PI on the PSID project, be one of the network PIs. By compiling longitudinal information on a large and continuously representative sample of U.S. families over a 25-year period, the PSID provides unparalleled longitudinal data on the poverty and welfare experiences of children, rich intergenerational data on childhood backgrounds (including neighborhood conditions) and adult attainments, unique data on intergenerational transfers, and oversamples of Blacks and Hispanics. Duncan's research record is replete with innovative studies of child poverty, welfare dependence, intergenerational processes, cross-national comparisons of family well-being, survey methods, public outreach and research network activities. Collaborating with Duncan on the proposed research project are Martha Hill of the Survey Research Center, who has designed and analyzed many innovative demographic survey modules concerning intergenerational issues, family composition and the economic fortunes of children, and Sandra Hofferth of the Urban Institute, who has made many scientific and policy-related contributions to research on kin support networks, child care and family structure. Two individual research projects are proposed. The first consists of a set of interrelated analyses of the effects of family background, neighborhood conditions and extended family networks on a series of important adolescent and early-adult outcomes -- dropping out of high school, attending college, early career attainments (including self- employment) and first home purchase. The second involves collaborators in Germany, France and Sweden and extends Duncan's comparative cross- national work on longitudinal aspects of poverty and welfare use to focus on the issue of why, despite apparently similar amounts of economic mobility, social-assistance recipients in different countries have such very different durations of receipt. Three network research projects are proposed: a system of descriptive longitudinal indicators of children's well-being and of trends in those indicators, building neighborhood indicators from the 1980 and 1990 Census into some of the other data sets in the network, and coordinated research on intergenerational exchanges.