With the polls appearing to drift somewhat in Obama’s direction over the last week, one of Romney’s best data points came from two national subsamples of early voters. Pew and Gallup both found Romney ahead among early voters, which represented a sizable drop-off from 2008. Given that state polls routinely find Obama performing better among early voters than election day voters, these results were both surprising and good news for Team Romney.
But part of why Romney performs well in national polls of early voters is the composition of the states with substantial early voting. Many of the large, Democratic-leaning states in the northeast and Midwest, like New York, New Jersey, Massachusetts, Pennsylvania, Minnesota, and Michigan don’t permit in-person early voting. Illinois now allows in-person early voting, but the state still prefers to vote on Election Day, as just 10 percent voted early in 2008. The only blue states with more than 10 electoral votes casting an above average share of their votes in early voting were Washington and California. In contrast, the three largest red states of Georgia, Texas, and Tennessee each cast an above average share of their ballots early. So far in 2012, each of those states has reported higher early voting turnout than California or Washington, where millions of ballots will be cast by mail. As a result, “red states” are beating “blue states” in early voting by an 8.5 million to 4.9 million vote, according to figures taken yesterday from Michael McDonald’s indispensible site on early voting.
Now, it's important to emphasize that this doesn’t mean that all is well for the Obama campaign. Republicans are doing better than they did four years ago, including in the battleground states. And the fact that the same polls showed Obama ahead by such a wide margin in 2008 suggests that the composition of the early voting states isn’t the only issue at place. But the Republican-lean of the states with high levels of early voting is a reason to focus on the state-level polling or registration data, rather than national poll numbers. Over the next few days, we'll focus on those more relevant data and take a look at the size and potential meaning of GOP early voting advances, but keep in mind that it's hard to draw many conclusions from early voting.