Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: Many states are conducting surveys to estimate rates of uninsurance, as well as to examine the characteristics of their uninsured populations. Yet, state generated estimates are likely to differ from the annual estimates of uninsurance rates based on the Census Bureau’s Current Population Survey (CPS). In our analyses from seven states using a similar instrument we show that the state surveys conducted produced anywhere from 15 percent to 50 percent lower uninsruance estimates than the CPS. Because the CPS estimates are widely cited by the media and in the health policy literature, they can create potential confusion when states use their own survey data for policy development. Our paper examines key reasons for the differences and discusses their relevance for state health policy. We focus on sample design, weighting adjustments for non-coverage, survey administration, definitions of the uninsured, survey design, and imputation/editing procedures. Although we are able to explain portions of the differences between the CPS and state survey estimates there is still several outstanding issues for both the CPS and state surveys.