Working with panel data such as that collected by the Survey of Income and Program Participation (SIPP) can be relatively expensive, both in terms of manpower and computing resources. The costs of using two or more panel data sets for comparative research increases, nearly proportionately, with the number of data sets. Experience with one panel data set is an advantage for working with another. Still there is no substitute for the detailed technical and substantive knowledge needed to use microdata, particularly a complex hierarchical longitudinal data set such as SIPP. Likewise, the data processing expenses (hardware, software and programmer time) associated with the use of multiple panels is considerable.

For these reasons, researchers are only likely to conduct analyses based on two or more data sets of this type, if a convincing case can be made for the methodological advantages and practical feasibility of such research. The opening section of this paper presents the methodological rationale for comparative panel research. Assuming that the reader is familiar with the Survey of Income and Program Participation (SIPP)1 , the second section begins with a general introduction to the history, organization and design of the German Socio-economic Panel (SEP). The discussion goes on to consider the SEP survey instruments and their content, field procedures, and data management and dissemination. The final section of the paper reviews the most important differences between SIPP and SEP and assesses their potential for comparative studies in specific substantive areas (including: employment and earnings, program participation and benefits, household composition and family events, education, and health and disability). This section is intended to aid SIPP users in determining whether they should look more closely into the SEP as an additional source of data.