Fun With Polls
Less than two weeks out from the election, there isn’t much to do. The narratives are set, the debates are over, and everything there is to know about the candidates (except Barack Obama’s relationship with Bill Ayers) has been revealed and overanalyzed. There’s nothing to do now except look at the polls.
Actually, maybe we shouldn’t look at the polls, because for a McCain supporter, they are pretty depressing. The RealClearPolitics poll average puts Obama up nationwide by 7.5 points, a pretty comfortable lead.
McCain can take comfort in the fact that these polls have a really wide range—Obama’s lead is one to thirteen points, depending on the poll. So if the polls with lower Obama leads are correct, McCain is still within striking distance. If the polls with higher Obama leads are correct, McCain probably shouldn’t count on becoming president.
The question is: which polls are right? The outcome of a poll depends, of course, largely on methodology—pollsters don’t just call up folks and start asking for their opinions. They have to weigh the answer they receive by race, party, sex, age, and any other demographic data—after all, a poll that oversamples, say, white people will be skewed towards the candidate with more support from whites.
We more or less know what proportion of the different races and sexes will vote—there probably won’t be any major surprises there. Party identification is more complicated—most polls seem to assume that Democrats hold an approximately six point party ID lead on Republicans. The party ID gap has not been greater than three points in any election since 1976 (this includes the 2006 elections). This suggests that either Democrats have an unprecedented advantage in party ID, or pollsters are oversampling Democrats.
Age is difficult too. It’s usually easy to figure out who most young voters will vote for—nobody, since young people don’t vote in significant numbers. But Obama has taken pains to attract the youth vote, which means that either there will be a lot of first time voters, or Obama really wasted his time there.
We can’t be sure, but I’m going with the second option. I’m a young voter myself, and I can’t say I’ve detected any great amount of enthusiasm for Obama among my peers. Most young people seem to support Obama, but not really in a particularly passionate way. Few seem to be really fervent Obama supporters—most support him just because it’s, you know, the accepted thing to do.
Maybe my experiences are atypical, but I don’t think so, at least not to any large degree. If there are any news stories of a massive youth grassroots movement supporting Obama, I’ve missed them. (And showing up at Obama’s rallies for an afternoon doesn’t count as a grassroots movement). Young people are always expected to show up in force and vote, and they never do. I wouldn’t, and this is an uninformed opinion, expect them to show up this year. If my uninformed belief is correct, then pollsters are oversampling another overwhelmingly pro-Obama demographic.
The polls that show Obama up typically give him about fifty percent of the vote, while McCain gets around forty-three. The polls showing the race essentially even give both candidates support at around the forty-three percent mark. This could be a result of pollster methodology. Pollsters who push voters for their decision (Rasmussen does this) seem to show higher Obama leads. Other pollsters, who might not be as insistent in their questions, show the race closer. This could mean that much of Obama’s poll support comes from voters who are essentially undecided, and simply pick the better known and more charismatic candidate when pushed for an answer.
Moral of the story? It could be that I am totally ignorant about polls, and McCain is finished. Or maybe it’s that I actually am a bit of a poll expert and my conclusions prove that McCain is right in this thing. I don’t know—we will find out on Election Day. But every person reading this should try to prove the polls wrong—go out and vote for McCain.