A widely reprinted AP story, based on a recent AP/GfK poll, is entitled, "Opposition to health care law eases." Don't believe it. What has eased isn't the level of opposition to Obamacare, but rather the level of effort that AP/GfK has made to ensure that its polling sample is representative of American voters.

When the AP/GfK poll screened for likely voters a couple of weeks before the election, it estimated that 48 percent of voters leaned Republican and that 42 percent leaned Democratic (which the election showed to be about right). In its latest survey -- the one that serves as the basis for the AP story -- AP/GfK didn't screen for likely voters and didn't screen for registered voters. Instead, it merely surveyed 1,001 adults. The result?  The percentage of Democratic-leaning respondents stayed the same (42 percent), but the percentage of Republican-leaning respondents dropped by 12 points, to 36 percent.

As one would expect -- with the same percentage of Democratic-leaning respondents having been surveyed both times -- the level of support for Obamacare remained essentially unchanged: 41 percent supported it previously; 40 percent support it now. Just as unsurprisingly, when the percentage of Republican-leaning respondents dropped by 12 points, the level of opposition to Obamacare dropped by 11 points (from 52 to 41 percent). 

Moreover, despite its gross under-representation of Republican-leaning respondents, the current AP/GfK poll still shows more people opposing Obamacare than supporting it, even greater opposition among those who feel strongly, overwhelming opposition to the individual mandate, and more support for repealing Obamacare in its entirety than for keeping Obamacare as it is. There's no story here, even though AP wrote one.