In a session last night with Sarah Palin, Sean Hannity of Fox News posed a number of leading questions. One of them dealt with a new Washington Post-ABC News poll showing President Obama and Mitt Romney neck and neck in the presidential race. Hannity zeroed in on one stat:

With that nine-point split, by the way, when you break it down, there was a seven-point split in 2008. In other words, this polling sample used by The Washington Post assumes the Democratic advantage over Republicans in 2012 will be 20 percent better than it was in 2008. Do you think this is purposefully manipulative of the media to put out something like this?

Palin spent a half-second giving this matter massive consideration, before replying: “No question. Absolutely. That’s par for the course, though, with the ma­nipu­la­tion of the media’s message — like they did four years ago, Sean. They chose their one, and they did all that they could to usher him into the White House.”

Stunning that Palin didn’t dig into the specifics of the polling questions that Hannity raised. We here at the Erik Wemple Blog will fill in that void.

When Hannity cites the “seven-point split” of 2008, he appears to be referencing exit polling from that election, when 39 percent self-identified as Democratic and 32 percent self-identified as Republican.

Now for that “nine-point split” in the Washington Post-ABC poll: 33 percent of respondents self-identified as Democratic and 24 percent self-identified as Republican in the survey. The jump from 7 to 9 is actually an increase of 28 percent, though Hannity charitably characterizes it as 20 percent. Even so, he takes an Olympic-caliber leap of logic in alleging that the Washington Post-ABC sample “assumes” the Democratic advantage will be 20 percent greater than in 2008.

Washington Post polling chief Jon Cohen notes: “We don’t ‘use a sample’ or ‘assume’ anything; we randomly sample conventional land-line telephone numbers — including both listed and unlisted numbers — and do interviews with the growing numbers of adults in the U.S. who are ‘wireless-only.’ ”

Given his eagle eye for manipulation of Democratic-v.-Republican polling constituencies, Hannity must certainly have scolded Fox News’s in-house pollsters just several weeks ago. Those liberal-media masseurs of data, after all, produced a May poll showing an 8 percent gap between Democratic-identifying respondents and Republican-identifying respondents. That’s a 1 point jump from the 2008 exit numbers, or a boost of 14 percent. So plugging those results into the Hannity polling-rhetoric machine yields this formulation: “This polling sample used by Fox News assumes the Democratic advantage over Republicans in 2012 will be 14 percent better than it was in 2008.”

So manipulative. Good thing that the most recent Fox News poll, from late June, shows just a 5 percent gap in party identification. Far less manipulative. Notes Cohen: “Fox News . . . deploys a broadly similar methodology in its polling and has averaged a five-point Democratic advantage so far this year. Fox asks the partisanship question a bit differently, but given the sample sizes in question, those margins aren’t all that different.”