The wonkosphere has turned it's attention to the role of the undecideds in the US Presidential election. With the polls showing a stable trend, with perhaps a little tightening, and with McCain forced to run robocalls in his own state and dodgy adverts on C&W stations in Virginia, the picture is in Obama's favour.
However, the one remaining mystery area is the undecideds. In each poll there is around 7% of the little devils. If they all break for McCain (which wouldn't surprise me) Obama is going to have great difficulty getting to 270 electoral votes.
In the Salon yesterday, Bill Greener argued, based on the supposed "Bradley effect", that most of the undecideds will go for McCain, leaving Obama on shaky ground.
There are a number of difficulties with Greener's argument.
1. It is arguable that the Bradley effect didn't actually exist in the first place. The polls on the actual election day showed Bradley heading for defeat anyway.
2. All Greener's examples of Bradleyesque effects can be offset, I suspect, by other examples of where African Americans have been elected according to the polls.
3. If the Bradley effect were, for the sake of argument, a 3 percentage points phenomenum, then it could well be offset by a 2-3 point "mobile phone" effect going the other way. It has been pointed out that pollsters are consistently going for landline voters while ignoring mobile phone owners, who overwhelming go for Obama.
4. Obama isn't black. He is of mixed race. He straddles two cultures. His upbringing in Hawaii in the care of his (white) grandparents does, I suspect, offset many Bradleyesque effects in people's minds.
5. It is 26 years since the Tom Bradley election. US voters have moved on.
6. Greener's theory is based on the sweeping and unsubstantiated supposition that the remaining undecideds have already seen enough of Obama and don't like him. However, today I have seen some ultra-wonkish figures which demonstrate, reasonably authoritatively, that the undecideds will split about 54-46 in Obama's favour.
To read more about this and actually drown in wonkishness, read Mark Blumenthal in the National Journal Online here.