You're biased in this election.
You might wonder how I know that. Well, for one thing, you're reading Wonkblog. That means you're really, really interested in American politics. And that means you're probably operating with a bias. Not a bias toward Democrats or Republicans -- though you may have that, too -- but a bias toward underdogs, and toward the idea that campaigns are unpredictable, even late in the game.
Toward the end of their book "The Timeline of Presidential Elections," Robert Erikson and Christopher Wlezien hypothesize that "the most informed and unbiased views are those of the gamblers who bet on elections to make money." And so they set off to compare the predictions from these betting markets with actual election outcomes.
What they find is sobering. "In recent elections for which we have data from electronic stock markets, prices have persistently overvalued underdog candidates," they write.
That bias is probably much worse for journalists. Unlike participants on InTrade, we don't lose any money if we bet on the underdog and bet wrong. But we lose readers if we fail to make the campaign interesting, and we lose reputation if we say the front-runner is going to win, and then he loses. Every professional incentive we have is pushing us to exaggerate the uncertainty that exists in elections, to emphasize that things could still change. And that means most Americans who pay attention to elections are getting a view that's distorted by the media's underdog bias.
Take this election. Right now, there are three obvious sources of uncertainty.
First, it's possible that President Obama's convention bump could fade more sharply than anyone expects.
Second, Obama and his allies have been outspending Mitt Romney and his allies on the air. But Team Romney has been saving its cash for a massive air assault over the next month. Perhaps that will move the needle.
Third, while debates don't typically make much of a difference in elections, that doesn't mean they never make a difference. If Romney's debate performance persuades 2 percent of the electorate to abandon Obama and support Romney, that could make the difference.
There's also, to paraphrase Donald Rumseld, "unknown uncertainty." A good example of this would be Romney's comments on the events in Egypt and Libya, which voters don't seem to have liked very much.
The campaign beat is, in a sense, the election uncertainty beat, and my guess is that will prove almost as true at Wonkblog as it does elsewhere. There's not much I can do about that, and there's not even much I want to do about that. But you should know that we're biased.
Related: If you look at the numbers, the Romney campaign is in trouble.