Sunday, November 05, 2006

Incautiously Optimistic

Dean Barnett:

My homeys at the Weekly Standard have published their election predictions, and they range from dire to calamitous. Meanwhile, I’m merrily skipping about Soxblog Manor (in a completely manly way, mind you) whistling a happy tune. I expect Tuesday to be an extremely successful day for the Republican Party.

First of all, let’s concede the obvious – if the polls are accurate, then we’re in medium to deep shinola. There’s no question about that. It reminds me a bit of Election Day 2004 when we had to reach no other conclusion than if the exit polls were accurate, we were looking at the horror of President Kerry.

At the time, it was unthinkable to serious followers of politics that the exit polls would be completely worthless. They had never been wrong before. But something fundamental had changed in the way the public was responding to exit polls or the way the pollsters were gathering information to make their work-product obsolete. By 9:00 pm., it had become obvious to everyone except a hysterical and incredulous Susan Estrich that the exit polls were wrong. Why it happened that way is still something of a mystery.

What is even more of a mystery is why they all erred in the same direction, dramatically favoring the Democrat. I’ve offered theories in the past why this is so, and I’ll briefly summarize them for those with better things to do this fine Saturday than thumb through my archives: Liberals are bigmouths who can’t wait to share their opinions with strangers. Conservatives have lives.

To take a more serious look at the matter, what the polls measure right now are people passionate or bored enough to spend a half hour talking to a stranger or, worse still, punching buttons on their telephone when prompted to by a recording. Democrats are more likely to tolerate this exercise, just as they were more likely to tolerate the inquisition of an eager grad student as they left the polls on Election Day ’04.

The pollsters have also yet to devise a way of predicting who’s actually going to show up on Election Day. One poll I read talked to some 1200 registered voters and deemed 1000 of them “likely voters.” Since even a presidential year generates only 60% turnout max, the pollster’s conclusion that over 80% of the people he spoke with are “likely voters” is the professional equivalent of him throwing his hands in the air and saying, “How the hell should I know who’s actually going to vote?”

Of course, the pollsters can’t say that because if they confess weaknesses in their methods, they’re less likely to have customers queued up to purchase their services. So instead they must perform a charade in which they profess omniscience.

SO WHAT ARE THE POLLSTERS MISSING? Well, first they’re missing the fact that a disproportionate amount of Republicans are likely to tell them to take a hike. Next, they’re overestimating the enthusiasm on the left. For all the cacophonous din that emanates from the left, it’s critical to note that even their greatest hero, Ned Lamont, underperformed the polls on primary day. By several points.

Lastly, and most importantly, they’re missing a historic Republican Get Out The Vote (GOTV) effort. Republicans are going to turn out like it’s a presidential year. Independents and Democrats will turn out like it’s an important midterm. The Republican turnout will be worth between a few and several points in every race where there’s an effective Republican machine. And that includes every battleground state...

No comments: