I really don't mean to make this into a personal crusade, but Harry Siegel, on the front page of the Politico, has a piece raising (yet again) the possibility that Barack Obama will win the popular vote and lose the electoral college. Siegel manages to find and quote two political consultants who demonstrate a profound ignorance of probability: Hank Sheinkopf says the likelihood of this happening is "50–50," while Lloyd Green puts it at 20 percent. These are the kind of people you want to try to make bets with.
As I've written and blogged before, this is very unlikely to happen. (Nate's current figure is 0.76 percent.) And, coincidentally, here's even more evidence, courtesy of Chris Bowers. Right now, Pollster.com's average of national polls has Obama leading John McCain by 5.6 points, or an 8.1-point improvement over John Kerry's 2004 performance. If you take the 2004 results and assume an 8.1 percent swing in each state, here's what the map would look like:
Look familiar? As Chris notes:
With the exception of Nevada, this is precisely what current state polling projects. The two closest states in the above map would be Virginia (within 0.2%) and Missouri (within 0.9%). Remarkably, all polling conducted since Obama clinched the nomination actually shows Virginia and Missouri to be the two closest states in the nation, and all other states, except Nevada, falling into their above projected categories.
The point is not that Obama will necessarily end up winning these states--just that, given a particular margin in the national popular vote, it's possible to predict with a fair degree of accuracy how the electoral map is likely to look, without resorting to micro-level explanations of the dynamics in each state. The electorate is quite nationalized, and states don't leapfrog each other on the red–blue continuum very easily. Or, put differently: Enough bellyaching about the Electoral College, already.