The Census Bureau established the SIPP Methods Panel project to evaluate and redesign the core instrument for SIPP, a recurring, nationally representative, longitudinal survey of people and their socio-economic characteristics. The objectives of the project are to improve response rates in SIPP, to reduce burden, and to improve data quality. It is a research project consisting of analysis of extant data as well as experimental research. The data analysis component includes examining patterns of nonresponse, examining reporting patterns across waves of interviewing, and analyzing patterns of income receipt. The program of experimental research consists of three phases, designed to allow for three iterations of testing and refining the Wave 1 core instrument and two iterations for Wave 2. Each experiment involves the selection of an independent sample of the population in six regional offices—half of which are to be randomly assigned to a control group, and half assigned to a treatment group. Each sample is designed to have 2000 interviewed cases, with the half in the control group receiving the SIPP instrument in the field at the time of the experiment and the other half in the treatment group receiving an experimental instrument.

The AAPOR roundtable discussion presents findings from the first two field tests conducted in summer 2000 and summer and fall 2001 and uses that to promote dialogue on successful approaches to designing instruments for complex longitudinal surveys. Topics to address include approaches to reducing item nonresponse and income underreporting, methods of collecting earnings and their impact on response rates and data quality, nonresponse follow up techniques and their success, methods of improving interview efficiency and assessment of interviewer satisfaction, the success of new methods of assessing within household coverage, and a cognitive assessment of dependent interviewing techniques.