Written by: John H. Thompson
Today, the U.S. Census Bureau hosted a webcast on income, poverty, and health insurance coverage statistics. One of the key discussion topics was the implementation of methodological changes to the 2014 Current Population Survey. In conjunction with the event, the Census Bureau released expanded information from last year’s survey; in two weeks, we will release the results from this year’s survey, the second year of data collection with the improved methodology.
The Current Population Survey is the primary source of labor force statistics for the U.S. population. Every spring, the survey asks respondents about their income and health insurance in the prior calendar year. As part of the Census Bureau’s commitment to continuous improvement in measuring changes in our society, we’ve been actively investigating new methodologies for several years – including more than a decade of research on the health insurance coverage questions, and through content tests for both income and health insurance coverage questions.
In 2014, we began asking the survey questions about income and health insurance coverage in a manner that is easier for the respondents to answer. Over the course of the past year, we continued our evaluation of the redesign, and sought out experts to review and provide feedback on our efforts.
Making changes to the Current Population Survey is not something we do lightly, because it can result in difficulties with year-to-year comparisons of the data. However, we needed to implement the changes in 2014 in order to establish a good baseline for health insurance coverage estimates before certain provisions of the Affordable Care Act went into effect.
With the idea of maintaining a consistent time series in mind, we introduced the redesigned income questions using a probability split panel design. Of the 98,000 households selected to participate in the 2014 Current Population Survey Annual Social and Economic Supplement, approximately 68,000 addresses received the traditional set of income questions, and the remaining 30,000 addresses received the redesigned income questions.
The split design functions like a bridge for year-to-year comparisons of the data. Last year, we used the traditional income questions to look at changes between 2012 and 2013. This year, we will use the redesigned income questions to measure changes between 2013 and 2014.
For the health insurance section of the questionnaire, a split panel design wasn’t necessary for year-to-year comparisons because data from the American Community Survey provided a consistent time series starting from 2008. We administered the redesigned health insurance questions to all households last year. As a result, we increased precision in the measurement of changes in health insurance coverage between 2013 and 2014 – before and after some of the major provisions of the Affordable Care Act were implemented. This year we will publish our first statistics showing the law’s effect.
For more information about the redesigned questions in the Current Population Survey, check out the recording of the webcast. You can also look to our Random Samplings blog over the next two weeks for more information about methodology changes. Check back with us on September 16 for the release of 2014 income, poverty and health insurance estimates.