Monday, November 03, 2008

End game, poll-wise

Today's Poll Smack


I pulled this from Volokh
Polls:

Today's polls released so far show everything from a 47-45% Obama lead (well within the margin of error), with 8% undecided, in the TIPP poll, to a 52-43-5 result in the "Gallup expanded." The CBS and ABC News polls, so far the most favorable to Obama, aren't out for today. So the polls are predicting everything from an Obama blowout to a tossup, perhaps leaning McCain (if, as most observers seem to expect, the undecideds break at least 2-1 in McCain's favor).

Overall, the picture is of a comfortable victory for Obama, but if the pollsters are wrong in their assumptions about turnout (will young people turn out now that the Iraq War has faded from the scene? Will Sarah Palin bring out a higher turnout of conservatives than expected? Will African Americans vote in unprecedented numbers, as expected?) and voter party i.d. (Rasmussen, for example, who has Obama up 51-46, is assuming 39% to 33% favoring the Democrats), it could turn out to be either a huge Obama victory or a close race. One interesting datum: McCain seems to have made serious inroads in Pennsylvania, reducing a double-digit gap of a week ago into the mid-single digits. Very likely too little, too late, but something of a rebuke to the "experts" who claimed that McCain was foolishly wasting resources there.

UPDATE: FWIW, a reader points out that Clinton did three points better in the Pennsylvania primary that the Real Clear Politics average of polls had suggested. Most likely, the undecideds broke overwhelmingly for Clinton.

I have two serious quibbles with this otherwise sober analysis.
  1. Bradley Effect.  - While everyone talks about it, no poll takes it into consideration, nor can they.
  2. Internals v. Media - We had here a documented instance where Obama's internal polling (PA) showed him below 2%.  This at a time when media polls were showing him 11-15% ahead.  Rasmussen, Marist, et al, are  more thoroughly objective than Obama's own campaign polling? Nuh-uh.   A week later a" nervous" Gov. Rendell seemed to confirm it.  What effect did any of this have on media polls?  None. So why place stock in them at all?
On the negative side for McCain is his puny, if it exists at all, get out the vote machine. He should have paid Rove $20M - if that's what it took - to get that into shape.  If moderate Republican voters do vote, however, and they have to be more brain dead than usual not to have at least one soul searing vision of impending Obama apocalypse not to be so motivated, then McCain wins. Period.

Oh. One more intangible.  God hates communists even worse than I.

3 comments:

Anonymous said...

Rodge,
I don't understand why no one adjusts for the Bradley effect. Obama routinely lost 5-10 points between the final poll and the vote in the primaries. But they not only won't adjust for it, they won't even talk about it. The pollsters will have serious egg on their face - and maybe blood on their hands - if the One loses.

Rodger the Real King of France said...

How the hell would they quantify it? It's there, but nobody will admit to it, and nobody can see it.

Billll said...

Tom Bradley was a hard-over left winger who lost to a more conservative white candidate. The obvious reason for the leftists loss is racism. Right?

Post a Comment

Just type your name and post as anonymous if you don't have a Blogger profile.