WORLDFAM · Towards a Unified Analysis of World Population: Family Patterns in Multilevel Perspective

The overarching aim of this proposal is to develop the conceptual and analytical instruments to establish a formal linkage between macro and micro level perspectives in demographic research, with an application to the study of worldwide patterns of family formation. Using census and survey microdata, we will conduct worldwide multilevel analyses that will allow us to investigate demographic trends at three levels of disaggregation: national, regional and individual. We will study the relationship between societal changes and three interrelated aspects of family formation: union formation, assortative mating, and intergenerational co-residence from the young cohort perspective. The societal effects will include phenomena such as educational expansion, women s economic activity, urbanization, as well as individual socio-economic characteristics. Analysis will be based on data from a vast new archive of international census microdata made available by the Integrated Public Use of Microdata Series international project (IPUMSi), with complementary use of Fertility and Family Surveys (FFS), Demographic Health Surveys (DHS) and Gender and Generations Surveys (GGS). The full dataset will amount to 124 countries, more than 1,400 regions and 305 million person records, statistically representing roughly 90% of the world population. This research raises complex theoretical and methodological questions. We do not contend that this project will be able to establish causality; rather, we will identify and illustrate differences between and within countries based on a rigorous and comprehensive set of variables exploiting microdata to develop systematic measures at different levels. Methodologically, the project will confront the challenges of combining datasets, providing meaningful measures of family formation, creating contextual variables, optimizing computational requirements, framing models that encompass different levels, time spans and regions.