« Listen To The Generals | Main | Today's The Day »

November 05, 2012

Somebody's Wrong

On the left they're incredulous that anyone could possibly believe Obama will lose tomorrow.  Josh Marshall says, "This could get weird."

Listening to a story from a friend this evening. Guy in a social setting talking to a group of Wall Street heavyweights. Every single one in the room certain Romney wins. Has Ohio locked. Has the whole thing tied up. No doubt.

Up to now Marshall wouldn't give Romney a snowball's chance in hell of taking Ohio, but maybe there's a little hint of doubt creeping in. 

One big thing: the impact of intense labor fights in Wisconsin and especially Ohio and the clear stand against labor taken by Mitt Romney. In Ohio at least, there’s little doubt he’s paying a big price. Here’s the story.

If Marshal has doubts, they are not shared by Mark Kleiman.  The certainty of Marshall's "Wall Street heavyweights" is just more evidence for Kleiman of how wildly out of touch those Wall Street guys really are.

Hoping that your guy will win even though he’s down? Reasonable. Figuring out a way he might actually win, despite the evidence? Not reasonable, maybe, but normal.

But being subjectively certain of a very unlikely outcome? Scary, in people who get to play with billions of dollars’ worth of other people’s money.

Kleiman blogs at "The Reality Based-Community" where "Everyone is entitled to his own opinion, but not his own facts."  Interestingly, the facts upon which Kleiman bases his astonishment are opinion poll results.  Let me say that again.  The facts upon which Kleiman bases his astonishment are opinion—.  How about if I break it off right there?

OK, I'll admit it.  Opinion polls results are facts.  It is a fact that a poll comes out the way it does, but the implied prediction of an opinion poll is not fact.  It's an educated guess, but in some cases, like this one, it may not be all that well educated. 

What could be wrong?  For one thing, pollsters have a way of using election models.  According to Gallup, the 2012 electorate looks like 2008, the best election turnout for Democrats in years.  So now you have pollsters like CNN who shows the race tied, but only because it sampled 11% more Democrats than Republicans.  They justfy such a wide disparity by saying the composition of thev2012 electorate will be the same as it was in 2008 instead of the way it was in 2004.  Maybe they feel they can ignore the more recent 2010 elections, where Republicans dominated, because it was not a presidential contest.

For another thing, there are certain quarters where opposition to Obama is automatically blamed on racism.  Chris Matthews is one who makes that charge all the time.  It's a purposeful attack meant to intimidate, but there is a potential backlash to this.  Intelligent people can disagree with Obama's policies, and they don't appreciate being called racist.  They can be intimidated only to the extent that they'll just keep their mouths shut.  The intimidation stops once they get into the voting booth.  It's another flavor of the Bradley Effect.

Finally, there are the 91% who refuse to cooperate with pollsters.  Are they representative of the population at large?  I don't think so.  I think the ones who won't answer the phone, or won't say who they support are more likely to be Tea Party voters than moderates or liberals.  If that's true, and I think it is, the pollsters are going to be way off.

For these reasons and in spite of Obama's recent bump in the polls, I'm sticking with my earlier predictions of a Romney landslide.  My argument gets additional support from the recent example of the Wisconsin recall where Scott Walker held a very slim advantage going in, according to opinion polls, but won comfortably when the actual votes were counted. 

Marshall and Kleiman are certainly entitled to think I'm off my rocker, as I think they are off theirs.  Somebody's wrong, and this is why I think they are.

Posted by Tom Bowler at 11:15 PM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Somebody's Wrong: