When I give presentations about the election across the country or abroad, I usually talk about which types of exit poll or Election Day information that, were I to have it in advance, would probably tell me the winner of the presidential race. Can Barack Obama get 65 percent or more of the Hispanic vote? Does John McCain carry white women? Who wins Catholics? And so on.
Some analysts are doing this on a state-winner basis, and there is an emerging consensus around Colorado as the tipping point state. Stu Rothenberg, who has joined us here at Salon before to discuss his analysis of U.S. House races, makes the case for Colorado in his latest column. Patrick Ottenhoff of Electoral-vote.com seems to agree, though he tosses in New Mexico as well. Nate Silver of FiveThirtyEight.com ranks Colorado fourth, however, behind Pennsylvania, Ohio and Virginia, based on results from his multiple-iteration computer simulations of most-common tipping point states. (See down the right side of his home page for details.)
As Silver correctly explains it, knowing that Colorado flipped likely tells you Obama also won the more competitive state of New Mexico; or if McCain wins Pennsylvania he probably holds more Republican-favorable Ohio as well. There is some new buzz about Obama taking North Carolina, which, were he to do so, probably means he will be the first Democrat since 1964 to carry Virginia, too. You get the idea.
But these are one-way results. Knowing Obama, not McCain, won Pennsylvania is less informative about who won Ohio.* That said, if I had to ask for a single state result, with the possibility of that result coming in either way, I’d still say I want to know the outcome in Ohio. (*If knowing the answer to the generic question “Who won swing state X?” seems redundant with knowing that a specific candidate in fact won that state -- it is not -- consider this: If I told you McCain won Vermont instead of just saying I'll tell you the winner of Vermont, the first provides more revealing information because learning that Obama carried Vermont provides no real insight into the overall outcome.)
Why Ohio? Because knowing that Obama won Colorado does not preclude scenarios where he wins there and loses the election because, say, McCain flipped Michigan and New Hampshire to red and held Ohio. And knowing that McCain won Virginia** may also be inconclusive if Obama can flip Colorado/New Mexico/Iowa, or just Ohio, he doesn’t need Virginia. (**The Washington Post today has a story about a new poll showing Obama now leading McCain in the Commonwealth.)
Are there scenarios where Obama wins without Ohio? Sure: The Colorado/New Mexico/Iowa is the most obvious one. But Ohio’s winner, either way, still seems like the one piece of future info I'd like to have today to predict who is going to win in six weeks.