Affiliations: [a] Department of Economics, Social Science Building,
University of Wisconsin, 1180 Observatory Drive, Madison, WI 53706, USA
Note: [1] Prepared for a conference of the National Academy of Science to evaluate the
Survey of Income and Program Participation, April 11–12, 1991. I am grateful
to Philip M. Gleason, Gary Burtless, and Clifford Clogg for their help and
comments.
Abstract: SIPP is examined for its uses for economic research, particularly as a basis for
estimating causal relationships between labor market outcomes and potential
policy variables. The Current Population Survey dominates SIPP as a provider of
descriptive statistics for most, but not all, purposes. As a large longitudinal
survey, extending for up to 32 months, and with rich data on income, wages, and
other variables, SIPP has unique advantages for dynamic analysis of labor force
behavior that is described by short or moderate durations, such as unemployment.
Examples of historical and contemporary research studies are discussed to
illustrate the strengths and weaknesses of SIPP. Among several recommendations
for using SIPP, it is argued that the most important is to obtain market
characteristics for the city or county locations of the survey respondents.
