The Survey of Income and Program Participation (SIPP) collects longitudinal data on a variety of topics including employment and program participation. The strength of SIPP is that respondents are followed for two to four year periods, even when they change residences, enabling researchers to perform analyses not possible with cross-sectional Census surveys. The longitudinal design of SIPP means collecting accurate residence history information is vital to the success of the survey.

The Bureau is redesigning the SIPP survey instrument, and this redesign will have implications for how the survey collects residence history information. The current in production SIPP has a respondent recall period of four months and respondents are only allowed to report one move during each four month interview period, or wave. Detailed information about previous residences is collected early in the survey (Wave 2), but this information is not concurrent with data collected on respondents in later waves. The reengineered SIPP instrument uses an Event History Calendar (EHC) format with a respondent recall period of one year. Residence history information for the year is collected concurrently with data on jobs, program participation, and health insurance. Respondents can report living in up to five residences during the reference year of the survey, and detailed address information is collected for each residence reported.

We analyze residence history data collected in the 2010 and 2011 field tests of the Re-Engineered Survey of Income and Program Participation (SIPP-EHC). We use two methods to evaluate the quality of the SIPP-EHC residence history data. First, we match the addresses collected in the 2011 field test of SIPP-EHC to spatial extracts from the Census Bureau's Master Address File/TIGER database. We report the percentage of addresses successfully matched at the place, county, and state level. Second, we compare the monthly mover rates reported in the test 2011 SIPP-EHC (covering calendar year 20 10) to the 2010 SIPP-EHC (covering calendar year 2009) to the 2008 SIPP (covering calendar year 2009). We primarily focus on whether there are time trends in the likelihood of respondents reporting moving in the two SIPP-EHC field tests, indicating memory decay or misreporting.