We followed the pre-election polls religiously as part of our project to capture all polls with data on early voting [HERE]. One of the things we noticed right at the end is that all the outliers seemed to regress to the mean at the end. Pollsters who consistently had the national race very tight, at 3-4 points (e.g., Investors Business Daily), suddenly went up to a 6-7 point margin, while others who had forecast double-digit blowouts (e.g. Pew, NYT) likewise came down to the 6-8 point range.
The poll averages used by Pollster and RealClearPolitics, among others, also worked out pretty well, particularly for the states. The final RCP average nationally was 7.3% and Obama ended up at 6%. (Pollster was even closer.)
Likewise, Georgia was close, but at the end it leaned to McCain and that's the way it went; North Carolina was close, but it leaned to Obama and it looks like that's the way it will end up. Virginia got closer, but then moved back toward Obama at the end, and it should wind up with Obama winning by about 5 points. Missouri and Indiana, still undecided, were basically dead heats in the polls.
There were no surprises--nothing like a South Carolina (sorry dad) tilting over to Obama due to some huge turnout of African-Americans.
Nor was there any so-called "Bradley effect"--a term we hope will die after this election.
Nor, to the dismay of Fox News pundits and GOP hangers on, was there some kind of hidden effect for McCain, such as the oft-made suggestion that pollsters were missing white blue-collar workers who refuse to answer their phones.
Fortunately, jerks like Dick Morris, who kept saying it was a lot closer than people thought, or that all those undecideds would break for McCain, were wrong. Can we banish these people from the airwaves now? After all, a good pundit is one who turns out to be right, at least generally.
Meantime, without all those polls to pore over every day, the Curmudgeon will have find something productive to do!
No comments:
Post a Comment