Yesterday, I participated in an event hosted by the Census Bureau and the National Center for Health Statistics on federal statistics on health insurance coverage. One of the key topics we discussed at the event was methodological changes to the 2014 Current Population Survey (CPS) Annual Social and Economic Supplement (ASEC) and the statistics from that survey, which will be released in September.

I am committed to engaging the public on the methodology and the results of our surveys, so I’m glad we had the opportunity to co-host yesterday’s discussion and answer questions on our health insurance coverage statistics. If you weren’t able to attend the event in person or via live streaming, let me update you on what we discussed.

At the Census Bureau, we are constantly working to ensure our data are relevant, and accurately measure major changes in society. With this in mind, we designed the recent changes to the CPS to better measure health insurance coverage. The changes will provide a good baseline for coverage before the Affordable Care Act (ACA) went into effect, so that we can assess the impact of the law after it is in effect. Research shows that some CPS respondents who had health insurance coverage did not report it during the survey interview. This resulted in higher estimates of uninsured than from other sources. I believe the new questions will improve our measurement of health insurance coverage.

We administer the CPS every year in the early spring, and ask respondents about the prior calendar year. The new questions provide more accurate, precise, and detailed measurement by asking the questions in a manner that is easier for the respondents to answer. We begin by asking respondents about their current coverage, and whether it started before or after January of the prior year. Follow-up questions determine if the coverage was continuous; if not, whether there was any other kind of coverage during the gaps; and whether there was any additional coverage during that time. The CPS now includes additional questions to provide information on health insurance exchanges and marketplaces.

By instituting these changes now, the Census Bureau will have one year of data (2013) as a baseline. Next year, we can compare this baseline with the data about coverage in 2014, the first year of implementation of the ACA, to see what effects the Act has had.

The Census Bureau implements changes in questions to improve accuracy, and we base all changes on research, testing, and peer review. The Census Bureau implemented these changes based on more than a decade of research, which indicated that respondents often had difficulty recalling whether they had health insurance coverage and the type of coverage for a calendar-year period.

There is rarely an ideal time to make changes in the CPS, because they usually result in difficulties with year-to-year comparisons. We timed these changes to provide an accurate baseline of coverage before the effective date of key provisions of ACA. If we had waited another year – collecting this data for the first time in 2014 – the data would not show the effect of the ACA. Next year, once we have collected the CPS data about 2014, we will be able to provide estimates of year-to-year change in coverage using a consistent methodology.