National Archive of Computerized Data on Aging 1996 Summer Workshop


              Using Multiple Surveys in the Study of Aging

                          July 15-19, 1996
         ICPSR Summer Training Program in Quantitative Methods
                         Ann Arbor, Michigan

                   Instructor: Richard T. Campbell
                               University of Illinois, Chicago

Until very recently, most research on human aging was conducted on a specific set of birth cohorts measured at particular points in time. It was quite uncommon to have data available for birth cohorts separated by widely varying historical circumstances. Although researchers understood the fact that virtually all findings from such studies have to be taken as provisional, subject to cross-cohort replication, there was little one could do beyond the data at hand. As the data base for the study of aging becomes ever broader and deeper, opportunities for replication and for analysis involving multiple data sets collected at varying points in time have become available. Some repeated cross sectional studies, such as the Current Population Survey, the General Social Survey and the National Election Studies consist of carefully replicated data observations over an extended period of time. Other studies, such as the series of three cross sectional studies of aging carried out by Ethel Shanas between 1957 and 1975, while not exact replications, are extremely useful. Exact replications of longitudinal studies are less common, but several rough replications are available. For example, the Health and Retirement Survey currently being conducted on a sample of persons 51-61 in 1993 is a rough replication of the Retirement History Survey carried out between 1969 and 1979 on the birth cohorts of 1905-1911, and the Panel Study of Income Dynamics overlaps both of these panels.

The analysis of multiple surveys presents the analyst with serious conceptual and technical problems. How does one deal with lack of exact comparability across measures and samples? How does one relate macro level variables that change over time to relationships at the micro level? How does one compare models across surveys when one of the surveys lacks variables measured in the other? The workshop will focus on these and related issues. The workshop is appropriate for researchers who are currently engaged in multiple data set analysis or who are contemplating such work. Students should have a good knowledge of regression analysis and some computing experience. A number of data sets will be made available for analysis, including those on NACDA's forthcoming CD focusing on the Health and Well Being of Older Adults. Among other data sets, the CD contains The National Survey of Black Americans, the 1957, 1962 and 1975 National Surveys of the Aged conducted by Ethel Shanas, along with similar surveys done in 1974 and 1981, and Waves I and II of the Americans' Changing Lives survey. ICPSR's computing lab, which contains a number of internet-linked high-end micro computers will be available to participants in the workshop.

The course will be led by Richard T. Campbell of the University of Illinois at Chicago, who is the outgoing NACDA Resident Scientist, with contributions by Fredric D. Wolinsky of Indiana University, who is the incoming Resident Scientist, and other prominent scholars.

Preliminary List of Topics:

Understanding the data base for the study of aging
Analyzing repeated cross sections
Pooling disparate surveys
Dealing with non-comparable measures
Comparing models at different time points
Linking micro and macro variables

Enrollment in this workshop is limited. Applications should include a vitae, a cover letter describing the applicant's background and research interests, and a completed application form. Forms are available from the ICPSR Summer Program Office. Modest stipends are available for participants. Send all materials and inquiries to:

Henry Heitowit
ICPSR Summer Program
P.O. Box 1248
Ann Arbor, MI 48106-1248
Phone: (313) 764-8392
Fax: (313) 764-8041

Applications are due by April 19, 1996.

NACDA is a project of the Inter-university Consortium for Political and Social Research, funded by the National Institute on Aging.