How big will tonight's "wave election" be and what will we call it tomorrow?
Some are using the term "tsunami." Others are calling it a hurricane.
I think that there is a strong case to be made that this is less a wave election like 1994 and more like the political equivalent of an asteroid strike.
Mark Blumenthal spells out some concerns all of us are having with the data nowadays as landline use continues its decline. Excluding mobile only households seems to have a 3 point impact on the data. A May 20, 2010 report sums this up best:
Weighted estimates from the landline sample tend to slightly underestimate support for Democratic candidates when compared with estimates from dual frame landline and cell samples in polling for the midterm congressional elections this year. The same result was seen in Pew Research Center polls throughout the 2008 presidential election. In the landline sample, Republican candidates have a 47%-to-41% margin over Democratic candidates on the 2010 generic horserace, but in the combined sample voters are evenly divided in their candidate preferences for this November (44% for each party).
As Mark points out, some of the polling we are getting could be artificially magnifying a wave that is already large in its own right. No matter the outcome, we need to get our heads around this challenge, as it will only intensify. What happens when no one has a landline phone? Do we go back to door-to- door interviewing? Do we run prediction markets across every swing state and House district?
Now, distinguished House election experts like Charlie Cook and Stu Rothenberg are predicting a wave that is between 50 and 60 seats and 55 and 65 seats respectively. These are safe estimates based on (a) the averaged generic ballot data we're seeing and (b) what we have all seen and lived through in other wave elections (notably 1994 at +54 GOP).
But, I think there is a case to be made that tonight will be something beyond these estimates.
The odd couple of Nate Silver and Jay Cost have explored this "uncharted territory."
First, as other's have already pointed out, the final Gallup generic ballot of +15 Republican is firmly within the asteroid strike range, and the Gallup data includes cellular phone respondents. We can discount this number as an outlier, as the well-respected Alan Abramowitz has recently, but what if Gallup is right? If the final Gallup data is accurate, then the 50-60 House seat estimates are rather conservative.
To quote from Gallup:
"Taking Gallup's final survey margin of error into account, the historical model predicts that the Republicans could gain anywhere from 60 seats on up, with gains well beyond that possible. It should be noted, however, that this year's 15 point gap in favor of the Republican candidates among likely voters is unprecedented in Gallup polling and could result in the largest Republican margin in House voting in several generations. This means that seat projections have moved into uncharted territory..."
So what if Gallup is right?
If the Gallup polling is right, we can use a model created by Alan Abramowitz to estimate the impact.
The result?
Abramowitz's report only estimated the results based on the outer edge of a Republican +12 in the final Gallup generic. But, at GOP +12 in the generic, the model has Republicans gaining 68 seats for a total of 247 seats. This means that a +15 generic ballot should equate to 74 or more seats.
This just sounds crazy, and that leads to my second point. People like Charlie Cook or Stu Rothenberg or anyone in this business don't like sounding crazy. A pick up of 70+ seats sounds, at first blush, like a person buying the Acela ticket to Crazytown. We all remember 1994 and this sets the bounds for what we consider normal or possible. Even then it felt a little surreal. I remember 1994 as a young campaign manager in Ohio, and that election caught us all a little off guard. So, it is far more preferable to make good faith estimates underestimating the GOP wave and sounding reasonable than to take the data to their logical conclusion and end up promoting an estimate beyond the expected range.
So, with this in mind, here is the case for an electoral result that looks more like a GOP asteroid strike and less like a wave:
1. The Republican advantage on the national generic ballot is much higher than it was in 1994.
2. The generic ballot momentum appears to be spiking at just the right time for Republicans.
One look at the trend on Pollster shows a fairly sharp rise over the last few weeks.
3. Unemployment is much higher today (9.6%) than it was in 1994 (5.8%).
4. 49. That is the number of Congressional Democrats occupying a House seat won by McCain
in 2008. When you consider this in tandem with the GOP generic ballot advantages we're
witnessing, a massive House seat shift doesn't seem as strange. The political tides washed many of these Democratic members and it may just wash them out tonight.
5. Some of the recent polling in these House districts has some Democratic stalwarts (like Jim
Oberstar) struggling against their opponents. These stalwarts will probably win, but if they're
struggling, then those in more vulnerable seats may truly be imperiled.
We'll know by tomorrow morning, but the proponents of the asteroid strike theory (like Jay Cost) have some valid arguments.