Photographer: Saul Loeb/AFP/Getty Images

NPR Polling: Difference Without Distinction

Jonathan Bernstein is a Bloomberg View columnist. He taught political science at the University of Texas at San Antonio and DePauw University and wrote A Plain Blog About Politics.
Read More.
a | A

Oy, National Public Radio and Mara Liasson. Really?

NPR Poll: Obamacare More Popular Than President

A new bipartisan NPR poll shows approval numbers rising for Obamacare — which is now slightly more popular than its namesake.

That's the headline and the lead paragraph of a write-up of an NPR survey. But in the 13th paragraph, you find that Obamacare ranks 1 percentage point higher than President Barack Obama . One percentage point! That's not "more popular" or "slightly more popular." There's no way that a single survey with a 1 percentage-point difference justifies any finding of difference at all.

Worse, severalother news organizations ran similar headlines and stories on the NPR poll. Just embarrassing.

I'll mention, too, that both of the top-line numbers - 47 percent support for the Affordable Care Act, and a 46 percent approval rating for the president - are higher than the HuffPollster current averages. Best practices would require placing the poll in the context of those averages, rather than presenting a single survey result as some sort of absolute truth.

NPR deserves some credit for noting that part of the reason for the poor Obamacare number is that a small group doesn;t support the law because they believe it doesn't go far enough. Liasson also does a good job of looking at the differences between Democrats and Republicans in self-reported intensity of opinions.

But really, a 1 point difference doesn't mean anything. C'mon, we know this stuff: Look at polling averages, not single polls; don't make too much of very small differences that may be statistical noise; double that caution for subsamples; report trends against polling averages, not the previous iteration of any particular poll. This is basic, simple, stuff, and there's no excuse for getting it wrong in the age of HuffPollster, FiveThirtyEight, RealClearPolitics, The Monkey Cage and Vox.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

(Jonathan Bernstein covers U.S. politics for Bloomberg View.He is co-editor of "The Making of the Presidential Candidates 2012."Follow him onTwitter at @JBPlainblog.)

To contact the author on this story:
Jonathan Bernstein at

To contact the editor on this story:
Max Berley at