In today’s Morning Examiner, I linked to FiveThirtyEight’s Nate Silver for his analysis of the early voting trends, based on whose partisans have come out to vote or cast their absentee ballots so far. I did this because it seems like everyone has their own spin on early voting. Silver has a good record of looking at numbers dispassionately and distilling them, whereas most partisans do not. (Remember the high-flying expectations of a Kerry victory? The talk on the Right in 2008 about how the polls were all just wrong?)
Since his post, which suggested that Republicans are getting the upper hand, Silver has come in for a mild drubbing by some lefty bloggers — some of whose thinking I have to admit I agreed with at first blush. For example, why is Silver focusing on party registration instead of past early voting trends? Wouldn’t the latter be a better comparison?
Silver has responded with this post, which I think addresses the question quite well. Early voting is a relatively young phenomenon, and extrapolation from such a small number of prior data points is bound to create misleading results.
Or would these lefty bloggers prefer the supposedly apples-to-apples comparison of the 2008 share of the early vote with this year’s share so far? That measure gives Republicans a 14-point swing. But this probably isn’t correct either. Ideally, you’d compare to 2006 instead — the last midterm — but Democrats did not emphasize early voting as much that year. As early voting becomes more popular and reaches an equilibrium point, we’ll know more by looking at the partisan breakdowns in each state.
The broader point is what Silver’s critics are missing. Democrats are trying to tell a story right now about how they have expanded the voter pool, and they are bringing those first-time voters of 2008 back to the polls. This is a circle that parties have long tried to square: If you can get all of the same voters from the high-turnout presidential year to show up and vote in the midterm, you can blow the other guy away, right?
Somehow, this trick never works, and the early vote tallies for this year offer no indication that this has changed. Democrats are underperforming their registration shares in most states, and underperforming their 2008 early vote performance in every state where there’s comprehensive data available.
Then there’s the long-forgotten fact that turnout isn’t everything. Just look at the national partisan turnout numbers from the 2004 exit polls, when Republicans gained a handful of House seats, and the 2006 exit polls, when they lost 31 seats:
Just looking at these numbers, you wouldn’t guess that much had happened, would you? But in 2004, Republicans lost the independent vote by only three points. In 2006, they lost independents by 18 points. The lesson, contrary to what seems to be the Rove-Plouffe-energize-the-base theory of politics, is that independent voters still matter, sometimes decisively.
UPDATE: And note that in both of the new Gallup generic ballot models released today, Republicans lead among independent voters by 23 points.