Over at the FiveThirtyEight blog, which is far and away the best prediction/polling analysis resource in existence, Nate Silver (co-author of the "1 in 60 million" study (.pdf) referenced yesterday) calculates what he calls the "tipping point" state, which is the state that pushes the winner over 270 electoral votes if you start counting with the overall winner's most lopsided state win and keep adding in closer and closer races. Silver runs thousands of simulations every day, every time ranking the states from biggest wins to smallest, finding the tipping point each time. He projects that Ohio has a 49% chance of being the crucial tipping point. This is due to Ohio's position as a relatively close, relatively elector-rich state. If Romney loses Ohio, he's probably screwed. If he wins it, he's probably our next president. Since most of my readership is in Ohio, seemed important to mention.
FYI, based on all kind of data including polls, past elections and economic indicators like unemployment, Silver currently gives Obama a 76% chance of winning Ohio.