Response Error and the Medicaid Undercount in the 2009 Current Population Survey

James Noon, U.S. Census Bureau
Leticia Fernandez, U.S. Census Bureau
Sonya Rastogi, U.S. Census Bureau

The Current Population Survey Annual Social and Economic Supplement (CPS ASEC) is an important source for estimates of the uninsured population. Medicaid coverage is one component of uninsured estimates, and previous research has shown that survey estimates consistently produce an undercount of beneficiaries compared to Medicaid enrollment records. This paper examines reports of Medicaid coverage in the 2009 CPS ASEC compared to enrollment data from the National Medicaid Statistical Information System for 2008. Linking individuals across datasets, we analyze two types of response error regarding Medicaid enrollment. First, some persons have Medicaid coverage but report no coverage in the CPS ASEC (false negatives). Second, some persons report Medicaid coverage in the CPS ASEC but can not be linked to Medicaid enrollment data (false positives). We use regression analysis to analyze factors associated with false negatives and false positives in the CPS ASEC and discuss implications for estimating the uninsured population.