Early voting will be wrapping up across the battleground states over the next few days and spin wars are already underway to declare one side the winner. But as a general rule, folks should be more circumspect about early voting.
The main issue is the absence of an adequate baseline. There has never been a close national election in the era of 35-40 percent early voting and we just don’t know what constitutes a good or bad showing for Democrats or Republicans in many of these states.
This problem is most acute in a few places: states where Obama won by a wide margin, but where Republicans are clearly outperforming Democrats and the Democratic vote is down; and states that were close in '08, but where Democrats are performing at or near ’08 levels even as Republicans are turning out more of their voters. Is it fair to argue that Democrats are doing worse in early voting if they’re turning out nearly as many or as many voters as they did four years ago? And given that Republicans weren’t exactly enthused four years ago, how much credit do they receive for large increases in early voting? If Republicans are doing better in early voting in a state like Colorado, how much better do they need to perform to overcome Obama's 9 point edge from '08. There just isn't a good answer.
Early voting is easier to interpret in states where the early vote looks much like 2008 and the '08 result was clear, or in states where the early vote probably needs to look just like '08 for Obama to win, and it doesn't. Only North Carolina and Nevada arguably fit into these categories. In Nevada, substantial GOP gains were necessary but Democrats carried the early vote by 6 percent, while any Republican improvement in North Carolina early voting could be interpreted as a sign that Romney should be considered a favorite in a state that voted for Obama by 14,000 votes.
Another issue is cannibalization—the trendy word for moving a definite Election Day voter to early voting. Obviously, it doesn’t make any difference whether a voter participates early or on November 6th, so early votes are most meaningful if the campaigns are banking tough low-frequency voters. Unfortunately, few metrics are available for judging whether the campaigns are succeeding in this regard. With that in mind, perhaps the best available metric is the total Democratic early vote. Everyone expects a strong Republican turnout, but Democratic turnout more questionable. A Democratic repeat among early voters can be taken as a sign of strong Democratic turnout, even if it's hardly assurance. If Democratic turnout declines, it is important to consider whether changes in early voting dates potentially invalidate a direct comparison with 2008.
Altogether, there’s more spin then reality. With two possible exceptions at either end of the spectrum—North Carolina and Nevada—it’s hard to make an argument that the data precludes a victory by either side. For the most part, the data just tells us the obvious: that Romney is doing better than McCain. But it’s impossible to say whether he’s doing “good enough” without an adequate baseline.