Don't Believe the Latest Elizabeth Warren-Hillary Clinton Poll

It's a lesson in how polls can artificially puff up candidates.

On Wednesday, the group Run Warren Run—which is fighting a quixotic campaign to convince Sen. Elizabeth Warren to run for president in 2016—released a much-hyped poll. "Huge Opening for Elizabeth Warren in Iowa and New Hampshire," a press release from Democracy for America, one of the groups supporting Run Warren Run, crowed.

And the results of the poll, at first blush, are shocking: The pooled results from Iowa and New Hampshire show 31 percent of respondents supporting Warren in a Democratic primary or caucus, versus just 26 percent supporting Hillary Clinton. Compare that result with those of a Bloomberg Politics poll released earlier this month, in which 56 percent of Democratic primary voters said Clinton would be their first pick, over Warren's 15 percent.

What's going on here?

Like many seemingly crazy poll results, it comes down to methodology. In this case, the Run Warren Run poll includes a string of ten leading questions that paint Warren in an exceedingly favorable light. Take the 16th question, which sounds like it was written in soft-focus:

16. Elizabeth Warren says she grew up "on the ragged edge of the middle class." After her dad suffered a heart attack and the bills piled up, she started waiting tables at age 13. Her mom got a job at Sears, and Warren says, "That minimum-wage job saved our home—and saved our family. ... Sure, I worked hard, but I grew up in an America that invested in kids like me, an America that built opportunities for kids to compete in a changing world, an America where a janitor's kid could become a United States Senator. I believe in that America." Is this a convincing reason to support Sen. Warren?

It's not so surprising, then, why so many respondents responded favorably to the idea of a Warren presidential run: They were primed to do so. On top of the leading questions, the poll's margin of error is 4.6 percent—meaning that the gap between Clinton and Warren is still a negligible one.

The poll, which was administered by the polling service YouGov, does admit what it's doing. On the sixth page of its polling memo, Run Warren Run says the poll does not put candidates on even footing.

"We should note—this is not a so-called 'clean' head-to-head ballot question, as voters were provided positive information about Warren but not other potential candidates," the memo reads. "It should not be read as reflecting how Iowans or Granite Staters would vote if the caucuses or primary were held today. Rather, it should be read as an indicator that many voters in these states are 'moveable,' open to supporting Elizabeth Warren when they learn about her and like what she has to say."

"The movement that we saw toward Elizabeth Warren once her message gets out there was incredibly impressive," Neil Sroka, a spokesman for Democracy for America, told National Journal. "It was surprising to us how movable voters were in New Hampshire and Iowa when they started hearing about what Elizabeth Warren stands for, what her values are, and what she would likely run on if she were a presidential candidate."

Still, the way the poll has been marketed to political reporters since its release has been much less subtle.

"These poll results prove that Senator Warren, should she decide to enter the 2016 presidential race, would be a clear favorite with voters on the critical, income-inequality issues that will define the election," Charles Chamberlain, executive director of Democracy for America, said in the poll's original press release.

Some polls are conducted with the intended effect of inflating the favorability of a group's preferred candidate and generating positive headlines about his or her momentum in early primary states. This appears to be one of those. Misleading polling may generate some positive headlines in the short term, but reality can quickly prick those inflated hopes—just ask Eric Cantor.