The views expressed are attributable to the authors and do not necessarily reflect those of the Census Bureau.

Beginning with the 1996 Panel, the Survey of Income and Program Participation (SIPP) has undergone a comprehensive redesign which included the introduction of a new survey design, the use of computer-assisted personal interviewing (CAPI), and the development of a new longitudinally oriented processing system. 1 As part of the conversion from paper-and-pencil personal interviewing (PAPI) to CAPI, several content changes were introduced. Many of these content changes became feasible because of the new collectionmode. In addition, there were changes in the approach to gathering data on certain topics, in question wording, and additional new questions added to expand coverage of existing topics.

As part of the 1996 Panel redesign plan, there was a series of field tests to develop the new CAPI instrument. A dress rehearsal of the survey was conducted in 1995 consisting of 2 waves administered in February through May and June through August. During that period wave 7 and wave 8 of the 1993 Panel of SIPP were still in the field. As a result, data for the same time period are available based on the old (PAPI) and new (CAPI) versions of the instrument.

This paper describes the major changes introduced in the 1996 Panel CAPI instrument, focusing on the sections on employment status, job and employer characteristics, and earnings. We also compare the data from the dress rehearsal and the 1993 panel in an attempt to gauge the impact of the redesign changes. Since the dress rehearsal incorporated all aspects of the redesign, the impact of the changes in both mode of collection and content relative to employment and earnings are examined. We describe the differences between the SIPP PAPI and CAPI instruments (that is, between the pre- and post- redesigned instrument) and we measure the total effect of the redesign on employment and earnings data.