When confronted with a prediction that doesn't match your own, it's natural to assume you are right and to try and prove it. It's why, with polls showing President Obama headed toward a comfortable re-election two years back, many Republicans set about frantically trying to "unskew" the results. It looked great on paper — "Hey, Romney is actually winning!" — but proved flawed come November.
With the midterms on the way, the unskew movement is back. And this time, the poll in question is a New York Times/Kaiser Family Foundation survey, released Wednesday, that shows Democrats doing well in four crucial Senate races. In the most remarkable finding, the poll gave Sen. Mark Pryor (D-Ark.) — who was believed to be the most vulnerable Senate Democrat this cycle — a 10-point lead.
The Republican National Committee's response: "Desperate after losing Nate Silver, The New York Times published a poll taken from people they found outside the DSCC who confidently predicted they'd keep the Senate."
The RNC's beef is that the poll ostensibly overrepresented Obama supporters, thus boosting the Dem candidates. (The percentage of respondents who said they voted for Romney in 2012 didn't match the actual elections results.)
But as the Times' Nate Cohn explained in defending the survey, "there's a well-known bias toward the victor in post-election surveys," so it's likely some Romney voters either claimed to have backed Obama, or declined to say whom they voted for. And since the rest of the poll's demographics make sense, the results can't just be dismissed out of hand.