The polls weren't wrong. The pundits were
Published November 13, 2014
by John Wynne, Politics North Carolina, November 11, 2014.
Those who cover politics are scrambling to figure out why the polls were so off this year. Nobody saw it coming, they say. Maybe polling is no longer a reliable practice – we might be able to get the numbers in the right ballpark, but until Election Day it’s a big mystery.
But in North Carolina at least, the polls weren’t wrong. The pundits were. Yes, Hagan had a very consistent 1 or 2-point lead. The final RCP average for the Senate race was Hagan +1.1. It was baffling to have pundits acting like this was some sort of insurmountable lead and that Hagan was going into Election Day the clear favorite. I’ve never seen a cycle where people were so gung-ho about claiming candidates with 1 or 2-point leads had their races wrapped up. John Hood made the same point in his latest column.
Of course, the poll average needs to be taken seriously. Every cycle, there’s one party clinging to hope that the polls are wrong. This is a bad place to be. It happened to Republicans in 2006 and 2008, to Democrats in 2020, and again to Republicans in 2012. In the final weeks of the Senate campaign, Tar Heel GOPers were also clinging to hope that, in this case, the polls did not reflect reality. It turns out, they did reflect reality – but the topline numbers belied the real story.
What was that story? The story was that Hagan couldn’t eclipse 45% in the polling average. Sean Haugh was taking an unusual amount of support – 4.9% according to RealClearPolitics. Past history would indicate Haugh getting less on Election Day. That left about 8% of voters undecided. Hagan with a 1-point lead, Tillis likely to benefit from the Libertarian losing support, and a ton of undecideds. You do the math. That’s not a race with which Democrats should be comfortable. That’s a toss-up. (Tom Jensen of Public Policy Polling, by the way, says the undecided voters in the last NC poll only gave Obama a 10% approval rating.)
To be fair, predictions of a Hagan win were also based on her supposedly incredible ground game. Indeed, looking at past history you would expect her to outperform the polls, if the early voting was any indication. And the early voting looked pretty good for Democrats – reason for cautious optimism for Hagan supporters, but not jubilation. There were also good signs for Republicans there too, namely a much older electorate than expected. In the end, the vaunted Democratic ground game turned out to be a paper tiger. They turned out a lot of voters and they did especially well turning out African Americans. But mostly they succeeded in simply redistributing their Election Day voters to the early vote period. And they lost independents big.
The victory by Tillis, then, was not a surprise. The surprise was that it came as a surprise to so many people when it shouldn’t have. Yes, Hagan led virtually every poll taken through the summer and well into October, but elections aren’t won by how many polls in which you lead. Even in the midst of the September ad blitz, with Tillis being depicted as a guy who wanted to destroy education in North Carolina, Hagan’s numbers still didn’t look impressive. In the end, the only thing resilient about Hagan was her inability to crack 47% in the polls. And that’s exactly where she ended up on Election Day.