The seemingly contradictory findings stem more from question wording and timing than from either poll doing something wrong or even misrepresenting public opinion.
The AP-GfK poll asks those taking the poll whether they "support, oppose or neither support nor oppose" the law, while the Washington Post-ABC News poll asks if they "support or oppose" it. The inclusion or absence of that neutral category is critical, as poll takers respond to the question they're asked rather than offering opinions unprompted.
AP-GfK polling has found significant shares of Americans choose that middle option. In surveys such as the Washington Post-ABC News one, in which a middle category isn't explicitly offered, less than 10 per cent typically say that they're unsure about the law. Poll takers are even more likely to choose a "neither" category when taking a survey online than they are when speaking with a live interviewer on the phone. But that difference existed even before the AP-GfK poll moved online in October 2013.
The difference between the two polls lies in that neutral group, which tends to be more Democratic than the overall public. About half of them in each of AP's recent polls identify as Democrats, and their demographics generally lean Democratic: mostly women, only about half white, less likely to have formal education and younger than those who explicitly support or oppose the law. Half in the latest poll said they approve of the way Obama is handling his job as president.
The March AP-GfK poll asked this group why they chose a neutral response and found the biggest subgroup to be Democrats who say they don't know enough about the law to say how they feel. The second-largest group consisted of Democrats who express mixed feelings about it. It makes sense that if that same group were asked how they felt about the law without a neutral position offered, they would tilt toward supporting a Democratic-backed law. That's consistent with the Washington Post's conclusion that the modest increase in support for the law is driven by Democrats who "somewhat" support it.
One further difference likely played a role in the divergent findings. The Post-ABC poll was conducted significantly after the Obama administration announced that the federal health insurance marketplaces had reached their revised target of signing up 6 million people on March 27. The poll launched on March 26 and wrapped up on March 30. A Quinnipiac University poll among registered voters conducted March 26-31 also found a slight uptick in support for the law, 41 per cent backed it, up three points since January. The AP-GfK poll was completed before the announcement.
On Tuesday, Obama announced that a late surge in sign-ups had pushed the number of health care enrollments to 7.1 million, exceeding the original target.
Associated Press-GfK Poll: http://www.ap-gfkpoll.com
Follow Jennifer Agiesta on Twitter: http://www.twitter.com/JennAgiesta
EDITOR'S NOTE _ Digits is Associated Press Director of Polling Jennifer Agiesta's take on the numbers that reflect our world and the survey research techniques used to find them.