Clues about Voter Turnout
Nate Silver at fivethirtyeight.com wrote an article today about turnout in Oregon. If you don’t know him Nate is excellent at statistics, but he is an unabashed Obama supporter. Usually, he doesn’t let his political leanings get in the way of his statistical analysis, but this time, he has missed it.
From Nate’s blog:
Uniquely among the 50 states, Oregon conducts the entirety of its voting by mail, which among other things can provide early clues as to what the electoral battlefield might look like elsewhere in the country.
Surprisingly, perhaps, turnout is down this year in Oregon. According to statistics compiled by the state’s Elections Division, the state has received 522,188 ballots through the first seven days of its early voting window. This compares with 645,473 ballots received during the first seven days of 2004 — a 19 percent decline.
He goes on to say:
More importantly, however, the counties most culpable for the depressed turnout appear to be those that voted substantially for George W. Bush in 2004.
This is true. However, Oregon is won, not unlike several other Blue states, in the highly populated urban areas. In fact, in 2004, John Kerry won Oregon 52-48% by winning only 8 of 36 counties in the state. It just so happens that 3 of these counties (Multnomah, Washington, and Lane) are among the 4 most populous in the state. In Multnomah County alone, Kerry won by 160,000 votes, more than twice the margin in the state as a whole.
So, turnout is down in the small Republican counties by 25-30% on average, and only down 15% or so in the large Democratic counties. That doesn’t sound good for the GOP, right? Well, if you take the math one step further and see what that means in terms of total voter turnout in the state, it is almost a wash.
Using the 2004 Bush/Kerry split and the 2008 turnout (based on Nate’s data), I calculate that the total vote swing towards the Democratic candidate is only 0.6%. This is certainly not good for McCain, but 0.6% is nothing like the 6-10% “turnout swing” that the Obama campaign is hoping for. The decreased turnout in Multnomah County alone hurts the Democratic candidate by as many votes as the decreased turnout in the seven most turnout-depressed GOP counties combined.
Certainly, this is only one state and there are many reasons why the turnout in Oregon is not necessarily representative of the country as a whole, but, based on these numbers, there is no reason to think that the turnout models of the last couple of elections are flawed. If they are not, then many of these polls showing Obama ahead could be off drastically in his favor.
Again I say: McCain still has a decent chance of winning this thing!