One piece of info I start paying some attention to a week out is the matter of which candidate people think will win. That candidate, recent history shows, usually does win.
But I follow this for another reason. It's my view that at least 1 percent of the electorate, and maybe more, votes for the guy who looks like he's going to win because they want to have voted for the winner. This is inconceivable to those of us with political convictions, but it is easily explicable via my Super Bowl party theory of elections, which I unveiled to great fanfare on Sirius XM's excellent Press Pool radio show the other day, hosted by the elan-rich Julie Mason.
You're invited to a Super Bowl party. You're not a football fan, and you don't really care about the outcome, you're there for the nachos. You ask someone, so who's playing? They say, Houston and Atlanta, or whatever. You don't care much about either city. You size up the uniforms--no strong preference. You ask someone about the teams. Nothing he says excites you. Well, who's gonna win? You ask. Guy says, Houston. Okay, you say, I'll be for Houston. You want to cheer for the team that's going to win.
People want to back a winner. If they don't care very much about politics but feel some vague civic duty to cast a vote, there's a good chunk of that portion of the electorate that just wants to be able to say, "I voted for the one what won." So, who do people think is going to win?
Here: Reuters has it 53-29 Obama, and Gallup says 54-32 Obama. In the 1996, 2000, 2004 and 2008 elections--the last four, that is to say--Americans have accurately predicted the popular vote winner. And that's all I need to say about that, I think.