U.S. flag

An official website of the United States government

Skip Header

Modifications to the Imputation Routine for Health Insurance in the CPS ASEC: Description and Evaluation

Written by:


The Annual Social and Economic Supplement to the Current Population Survey (CPS ASEC) is an important source of information about health insurance coverage in the U.S. Due to its long time series, its state representative sample and its detailed series on health insurance, the CPS ASEC is a critical data source for federal and state policy making and health policy research (Blewett et al., 2004). It is routinely used in surveillance activities and in policy evaluations; to project the cost of proposed legislation; and was historically used as an input in the federal allocation formula of the State Children’s Health Insurance Program (SCHIP).

The U.S. Census Bureau is engaged in an on-going effort to improve the quality of health insurance information in the CPS ASEC (Ziegenfuss and Davern, 2011). The purpose of this paper is to explain the latest development in their quality improvement effort – a change to the imputation routine for health insurance coverage. In an experimental version of the 2009 CPS ASEC, the change to the imputation routine and a simultaneous correction to the coding of directly purchased coverage increased the percent of people with health insurance coverage by about 0.5 percentage points (1.5 million people), primarily through an increase in private coverage. The Census Bureau implemented the new method with the 2011 CPS ASEC and retroactively applied the new routine for the 2000 to 2010 CPS ASECs. In addition to the new allocation procedures, the new data files reflect all data processing adjustments that have occurred since 2000. The data were released as supplementary files available from the Census Bureau’s health insurance web page.

The rest of this paper is organized as follows. First, we describe the CPS ASEC, review its imputation procedures and discuss previous work that identified problems with the way health insurance was imputed. Next, we describe the modifications that the Census Bureau made to correct the problem. Finally, we empirically evaluate the new procedure and discuss the implications of our findings.


Back to Header