Many thanks to Diane for the kind introduction of me and my wife, Sabrina (who might be joining in from time to time). It’s great to be part of the conversation here at Ricochet!
I’ve been told that I’m bad at small-talk, so let’s get right to the point . . . is Obama really going to win a second term?
Let’s face it, this is the big question. The one that keeps supporters and opponents of the President up at night. It’s why each new poll shift and fundraising total sends the politically-obsessed on both sides of the aisle swinging toward relief or depression.
For Romney supporters/Obama opponents, it seems preposterous to think the President could win with the sputtering economy, a fizzled “stimulus,” Obamacare, and all the trouble in the world.
And yet . . . the polls show a tight election, with Romney typically close but trailing. The election forecast models mostly predict an Obama victory. These models look primarily at economic indicators and presidential approval from recent polling. The economic indicators, at least those used in the models, are bad for Obama but not so bad that they tank his prospects. Same goes for the President’s standing in the polls.
But the polls don’t tell us who will make up the actual electorate this November. Polling does a pretty good job of telling us a few things; a) the opinion, right now, of all adults, b) the opinion of all those who say they are registered voters (sometimes actual registered), and c) the opinion of “likely voters,” which means a pollster’s best guess at the kinds of people who will actually turn out.
I come at all of this from an academic political behavior background, not from straight polling/public opinion. What I find most interesting are what message experiments, like our PocketTrials, reveal about how certain kinds of voters move in response to political “treatments” like campaign ads, news clips or articles with new information. These kinds of experiments identify causation and add to knowledge in a way that traditional polls simply can’t.
But we have to start somewhere for a baseline. So what’s the best baseline? The polls, or are they hopelessly biased or inadequate? Which models, or are they hopelessly simplified for an election in a time like this?
I have some thoughts and some really interesting, big, but mostly overlooked academic datasets that I think help ground us, but I wanted to see what everyone thought about this at the outset.
Is it President Obama’s election to lose, or Governor Romney’s?