Skip Header

Report Number SEHSD-WP2014-17
Matthew W. Brault


In March 2013, the Census Bureau fielded a content test to evaluate new income and health insurance questions for the Current Population Survey (CPS) Annual Social and Economic Supplement. The response rate to the survey was 43.1 percent, which raised questions about bias in key estimates. The design of the Content Test was similar to that of the American Time Use Survey (ATUS), which also suffered from low response. However, research exploiting the sample design showed that non-response did not greatly bias key ATUS estimates (Abraham, Maitland, and Bianchi 2006). Using the ATUS analysis as a framework, I modeled survey response, contact, and cooperation rates from sample characteristics obtained during households’ last CPS interview. Results suggest that age, educational attainment, and household size appear to have the greatest effect on response. Non-response appears to bias uninsurance in general; employer-sponsored insurance and Medicare specifically; and wage, retirement, and interest income recipiency, but has little effect on mean income amounts. For instance, non-response exerted a 2.8 percentage point downward bias on the percentage of households with a person uninsured for the full year.

Related Information

Back to Header