via Wall Street Journal
Sometimes, fixing one statistical problem creates a new one. For years, the Current Population Surveyconducted by the U.S. Census Bureau overestimated the number of people without health insurance. Last year, the Census Bureau revised the survey to correct the error, which was likely caused by the way the government asked Americans about their coverage.
The fix, which had been in the works since 1999, should produce more-accurate estimates. But responses obtained with the new methodology aren’t compatible with responses collected using the previous approach, making it harder to examine uninsured rates over the years at a crucial moment—as Americans try to gauge the success of the Affordable Care Act…
To read the full article please CLICK HERE