Anti-Climacus

WELL: So I've been thinking a little lately about the theory that's been getting more and more play, which says that voting blocs are very fluid at the moment, and it's hard to say where people's loyalties lie, so, for example, the poor job-recovery part of the economy will hurt Bush in Ohio (or Missouri, or Tennessee, etc), because the sort of blue-collar voters who normally turn out for the Republicans either won't turn out or will vote Democratic. I even had a conversation with OGIW where I floated this thesis, and she emphatically agreed with me. It was then that I knew something had to be wrong with it.

Then I read this Matthew Yglesias post on polls (ignore his conclusions, they're not particularly interesting), which goes to demonstrate one of the central theses of realist democratic theory: people often have no idea what they're talking about*, so look at what they do, not at what they say. Which means your best reliable predictor for how people are going to vote is how they voted last time, not how they say they're going to vote this time. Think about the states that are typically designated as in play**--they're virtually all states the Dems won in 2000. So it doesn't matter if Bush doesn't get Macomb County Repubs to turn out--he's probably not expecting to win in Michigan anyway, and if those MCR's return to type and vote for Bush, he might end up winning unexpectedly. In other words, until you start seeing polls where Kerry is running above 55% nationally, assume everything is going on like normal.

*in the sense that most people don't follow politics that closely, so their opinions are liable to sway wildly over time, even if their voting behavior never changes

**note to OGIW: Florida is not in play--if the Dems couldn't get it done in 2002, they're not going to get in done in 2004, especially with that open Senate seat.