This paper reports the general results of research undertaken by the Census Bureau staff. The view expressed are attributable to the authors and do not necessarily reflect those of the Census Bureau.

The Survey of Income and Program Participation (SIPP) is completing implementation of a major redesign. Starting with the 1996 panel, we will be changing the pattern of how we interview households. Instead of overlapping panels with a new panel being introduced each year, we will be changing to a non-overlapping design with a new panel being introduced every 4 years. The sample size for each panel will also be increased. The 1996 panel will contain about 37,000 household compared to 20,000 for previous panels. The 1996 panel will also include an oversample of the low income population.

The change in design supports the primary objectives of the SIPP which is producing longitudinal estimates of income and program participation, paying most attention to improving the information for people who are economically at risk, and improving the capability to respond to current policy needs in topical areas.

The redesign embraces all aspects of the SIPP program including sample design, questionnaire design, a move to computer assisted personal interviewing (CAPI), and automation. This paper will focus on the effects of the questionnaire changes, the move to CAPI, and the automation enhancements.

As part of the CASIC, computer assisted survey information collection, initiative at the Census Bureau, we will be using a CAPI questionnaire for the 1996 panel. Rather than simply automate the paper documents, we changed some question wording and content, moved some data consistency checks from post collection processing to the CAPI questionnaire, automated survey management and redesigned the data processing system. The areas of concentration were the labor force questions, follow-up on amounts in the assets income questions, and clarification/enhancement of general income questions and health insurance questions (Singh and Huggins, 1994).

The Census Bureau conducted a Dress Rehearsal (DR) during 1995 to prepare for full-scale implementation in 1996. The DR represents our first opportunity to compare data collected from the automated SIPP/CAPI instrument with data collected on a paper and pencil questionnaire for concurrent time periods.

This paper will identify the incidence and estimate the magnitude of data differences between the 1995 DR data collected from the SIPP/CAPI and data collected by paper and pencil interviewing (PAPI) in the concurrent 1993 panel. Our effort here is only to identify data differences because the transition of SIPP from paper and pencil to CAPI is overshadowed by other important changes in the survey design. Our reference to CAPI throughout the paper will refer to all aspects of the survey redesign. Also, our reference to PAPI will refer to the 1980 SIPP design. Results from this research will provide some information about the redesign's effect on the SIPP data and expectations for 1996. Therefore, the subject of study is the redesign of SIPP.