Measurement Error in Health Insurance Reporting

Abstract

In the United States, surveys serve as the only source of data for the number of uninsured people; they also provide rich data for exploring the relationships between health insurance coverage and individuals' life circumstances, such as employment, income, and health status, enabling researchers to assess the effectiveness of various aspects of the health care system. The Current Population Survey (CPS) is one of the most influential surveys measuring health insurance, but it is not without critics. To address outstanding questions about the data quality of the CPS health insurance questions, qualitative testing was conducted to assess various aspects of the questionnaire from the respondent's perspective. A testing protocol was developed largely based on previous health survey methods literature, and test subjects were probed about their comprehension of the questions, particular terms and phrases, and their strategies for formulating an answer. Several design features were identified as problematic, including the overall questionnaire structure, the calendar year reference period, the household-level design, and the wording of questions on public coverage.