Pennsylvania is a lot like Michigan, in that it has been reliably Democratic state in recent elections, has a history of fairly conservative Democrats holding office as well as comparatively moderate Republicans, is very pro-union, and pro-gun, and pretty divided on social issues. That’s why I knew Democrats were in trouble in Pennsylvania when I started seeing really bad polling numbers out of Michigan. Then the numbers in Ohio, Indiana, Illinois, and Wisconsin started to go south, and I realized that something very bad was happening in the Rust Belt, and that we could be in for a punishing midterm election that goes far beyond anything I had imagined possible.
At the same time, I really had trouble believing the polls because I couldn’t identify any real support for the Republicans. In all national polling they remained less popular than the Democrats, less supported on the issues, and far less popular than the president. In fact, they have been polling at record lows all year long. Then it became clear that poll respondents weren’t actually signaling support for Republicans in many cases, but pollsters were weighting the polls in the Republicans’ favor because their voters were expressing more certainty about whether or not they would actually vote. This became known as the enthusiasm gap.
I’ve been pondering the enthusiasm gap and polling practices for a while now. I was unsettled in 2008 when the polls (of House races) proved to be surprisingly accurate. I had hoped that Obama’s massive turnout operation would lead to Democrats further down the ballot outperfoming the polls. But it turned out that people did more ticket-splitting than I had predicted, and the polls seemed to have gotten the electorate right. So, all year I have had the sick feeling that the polls (excepting Rasmussen) are correct.
Yet, on the other hand, since the polling was only showing a Republican advantage in commitment and not in outright support, I have been optimistic that the polls would start to close as the election grew closer and more Democrats were exposed to media coverage of and advertising from the campaigns. That seems to be happening now, and we can see concrete examples of Democrats who have trailed all year pulling into the lead.
One other reason for optimism is the rising level of early voting. If everyone has to vote on the same day (which is the case in Pennsylvania, unless you certify that you will not be home on election day) then the enthusiasm gap is very important. But if you have five weeks of early voting (as is the case in Ohio, for example) then the enthusiasm gap seems much less meaningful. Take this example from Nevada:
Walter Grimsley woke up Saturday morning and remembered he had an errand to run. He had to go vote.
The government retiree’s phone had been ringing every few hours with reminder calls from Democratic volunteers. Canvassing teams appeared at his door at least 10 times. At a voting booth set up next to the produce section of his neighborhood grocery store, Grimsley cast his ballot for Sen. Harry M. Reid and the rest of the Democratic ticket, just as he had always intended.
The Democrats are well positioned to push, pull, and drag all their voters to the polls in states with early voting, so how important is that enthusiasm gap? I think we’ll have to compare turnout in Ohio and Pennsylvania after the election and see how much it mattered. In fact, we’ll have to look at a lot of data:
Democrats across the country know they face a daunting enthusiasm gap that veteran politicians such as Reid can’t possibly overcome. What they can do is try to outperform their Republican opponents by taking advantage of the longer window to get folks to the polls.
Election analysts expect more people to cast early ballots this year than in any previous midterm election. A decade ago, early voting was an obscure practice allowed in just a handful of states. This year, the District and 32 states, including Maryland, allow some form of early voting.
So, I have also been wondering all year whether the polls are accounting for this wave of early voting. And early voting isn’t necessarily a complete positive for the Democrats. They saved money in September so that they’d be able to have more money than the Republicans in October (not including independent expenditures, of course), and that means a lot of people will have voted before being exposed to their messaging.
I am still uncertain what to expect on election day. I’ve been taught a lesson about doubting pollsters but, in this crazy environment, with so many factors at play, I do still doubt that they’ve got the electorate right. The one thing I feel confident in predicting is that the polls will continue to move the Democrats’ way. Some, like Rasmussen, have been about setting expectations, so they will move toward getting the elections right to maintain their credibility. The rest will merely pick up a reduction in the enthusiasm gap that is the natural result of Election Day approaching.
