Share Judging Wave Sizes: Election Day Ruminations with your friends.


E-mail It

Social Web

November 01, 2010

Judging Wave Sizes: Election Day Ruminations

Posted in: Uncategorized

The final pre-election polls have all indicated a huge advantage to the GOP, and expectations of a Republican pickup of 50 seats or more in the House is commonplace. But it wasn’t so long ago that predictions were wildly off in a similarly odd election year, 1994.

Simply put, every prognosticator making a prediction in that election found themselves expecting a 25-seat gain for Republicans in the first midterm of Bill Clinton’s Presidency. In reality, the Republicans took 54 seats. The idea that wave elections are hard to measure would be the moral of the story.

But are they? We have arguably more sophisticated instrumentation these days – more polls, the ability to read local media from everywhere, more statistical analysis available. But I wonder whether the culmination of that is a group of pundit future-tellers who know a bit too much. As Nate Silver points out, it’s not entirely possible to predict the size of the Republican wave until we see some hard vote counts. There’s no doubt that the trend line for Democrats is sharply negative, without much resistance in the polling data. But it’s hard to see precisely how that translates. Could it be the difference between a House of Representatives with Democrats in charge and one with Republicans in charge? That’s not so likely. But Nate also noted today that the difference between polls with cell phones included and landline-only polls is about four points. That’s statistically significant – and could mean quite a bit on Election Day.

Does that suggest some kind of Democratic poll-defying triumph tomorrow? I don’t believe so. The losses were baked into the cake in 2009, particularly by the closely-watched process on the health care bill, which validated conservative claims and which happened while unemployment shot up. That was basically the moment Democrats lost touch with independents. But it does mean that there’s substantial pull one way or another in those numbers. And furthermore, the numbers might play on one side of that wave in one part of the country, and another side in another. For instance, California does not look particularly susceptible to the wave, but some Northeast states, in addition to the South, do.

I’m going to be spending tomorrow looking at individual House and Senate races, along with the all-important Attorney General contests, particularly in the Midwest.

Return to: Judging Wave Sizes: Election Day Ruminations