When wave enthusiasm cuts against the base population partisan identification, special elections favor the wave-associated candidate, provided the race is somewhat close to begin with.
After the successful 2006 Tester Senate race, I was hungry for more. A couple weeks drinking champagne, catching up on the third season of The Wire, and asserting dominance over my fantasy football league, I then got in the car and headed for Texas.
In TX-23, incumbent Republican Henry Bonilla was facing special election against former Congressman Ciro Rodriguez. Every single poll showed Bonilla comfortably ahead, including the Election Eve SurveyUSA poll that showed Bonilla up 51-47. The SurveyUSA poll a week before the election put Bonilla up 53-46.
Rodriguez won by 9 points, 54-45. And, with big DCCC investment and hundreds of workers pouring into the effort from around the country, the effort made the difference.
Consider that 2006 was a wave year, like 2008. Consider that turnout in the special election was 70,473, whereas a few weeks earlier the turnout on November 7 had been 123,799.
Or, take another example from this year. Don Cazayoux, who won a special election in May to take the Louisiana 6th district House seat on 101,017 turnout. On November 4, Cazayoux lost, with at least 312,416 having voted in the race.
Special elections simply have lower turnout, and thus it falls to base enthusiasm and getting out the vote. Organization becomes the key. How many Democratic organizers and volunteers in Georgia and around the country are not yet slaked of their thirst for beating Republicans? These people exist. How many will get in their cars and drive to Georgia to work? How many Republicans will be motivated to arrive from around the country to do the same work?
Already, we know that many among Obama's Ohio staff are flooding to Georgia. When we were down there just before Election Day, we got emails from Obama organizers telling us they'd already bought the tickets assuming there would be a runoff.
This doesn't guarantee a win -- by any stretch. When we arrive in ten days or so, we may find a sense among some Democratic voters that Obama's win was the golden prize, and the foot may come off the accelerator a bit. Or we might find steely-nerved motivation.
But the bottom line is that special elections don't tend to draw the iffy voters. It's about voter education and organization and turnout. From what we saw in the state when we were there the weekend before Election Day, Democrats had the edge. Keep an eye on the organizing stories from Georgia. They may be determinative.
[UPDATE] Many commenters seem to be missing the forest for the trees. Runoff, special election... a distinction without a difference. These are races not on the main election day when turnout is large. These are elections unto themselves on different days that attract fewer overall votes. Votes aren't fewer because it's a runoff or because it's a special election. Votes are fewer because it's not when the Big Election Day takes place that's in everyone's mind. Even primaries attract many fewer voters than general election days. Special elections, runoffs, these take special effort. Many more people will not take the effort to show up. The question is, will more Chambliss voters drop off or will more Martin voters drop off? That's the unknown. In wave years I'd tend to bet with the wave party, but I'm nowhere near ready to conclude Martin will win.