Man oh man, I've been looking forward to this moment.

One of my dirty little secrets: I hate the dang polls.

A few years ago I was reading Jean Edward Smith's biography of Franklin Roosevelt, whose political career began before the age of polling but ended after professional polling began (you can actually check out FDR's job approval numbers here). Early in his career - when he ran for Vice-President in 1920 - there were no polls, so Smith gives a sense of the expectations in the Roosevelt camp based on oddsmakers in New York City. When FDR becomes president and the age of polling begins, Smith switches to using the polls. As I read the book and realized that FDR and the political world made that transition in the early 1930s, I actually felt a little sad. If I had a choice to live in a world with polls or without them, I'd choose without. I much prefer analyzing votes to polls. Polls are variable and usually come from the mainstream media, which I do not trust. Votes are concrete, reliable and come from the people.

[Professional polling has also ruined political science, if you ask me. Nobody writes books like V.O. Key's Southern Politics in State and Nation anymore. It's too much work, I guess, when polling data is free to download from scores of public sites and tenure committees demand ever more articles on Bayesian updating in public opinion.]

So, it's with great pleasure that I can announce that this will be my last review of the polls this midterm cycle. Four quick poll checks:

(1) Gallup has updated its presidential job approval numbers: 44% approve to 48% disapprove. Bill Clinton's pre-election job approval in 1994? 45% approve to 46% disapprove.

(2) RealClearPolitics has put in place its final generic ballot average. The result: 50.7% Republican to 41.3% Democratic. While some analysts who have leaned heavily on the Gallup poll for years are now suddenly treating it as suspect, the only real point of disagreement with the other polls is the percentage allocated to Republicans (55% as opposed to 50%). It is right there with the other polls in terms of the Democratic share (41.5%). That latter number is interesting, isn't it? Here we are at the end of the election cycle, something like a billion dollars spent (more by Democrats), and the Democrats can't get above 42% of the vote.

As a point of comparison, John McCain's final poll average in 2008 was 44.5%.

If we suppose the remaining undecided voters this year split evenly, we get 54.7% to 45.3%. If they split proportionally, it would be 55.1% to 44.9%. Both numbers are right at the 10-point victory that I think the Republicans are going to get today.

(3) As for the RealClearPolitics job approval average for President Obama, it stands at 45.6% approve to 49.4% disapprove. This is right in line with where the average has been for the entire fall. It suggests again that the net effect of this campaign has been very little, which makes sense. Midterm voters know everything that has happened since Obama became president. No amount of campaigning will change what they know, and probably can alter only marginally how they feel about the state of the nation. This midterm has been over for months, if you ask me.

(4) Oh, and about that enthusiasm gap. It's still ginormous. But then again, it's only Gallup. And we're to ignore Gallup unless and until it starts giving Democrats better news. Otherwise, it must remain in the outer darkness with Rasmussen and Fox -- y'know, those "biased" pollsters who nailed the 2008 presidential campaign!

Next Page