Will Obama Experience a Reverse Bradley Effect?

I had been planning to run this post some time in the next week when I didn't have anything more timely to print. Then this morning I saw Electoral-vote.com refer to a University of Washington study going over basically the same thing. Oh well, I don't get credit for breaking the concept, but here is my take anyway:

A month or so ago, on a cable news show, I heard an interview with a voter from a Rust Belt state talking about the voting tendencies of his co-workers. He said that none of them publicly said they were voting for Barack Obama, but privately several had confided in him that was what their plan.

That got me thinking. Was the Bradley Effect just describing white voters reluctance to tell pollsters they were not voting for the black candidate because they felt that was not the politically correct thing to say? Or, is it really more general - a person believing that it is unacceptable in their local social circle to announce they are supporting a particular candidate? If the later is true, and given that the Bradley race in California in 1982 and a few others that may have been associated with the Bradley effect (Wilder, etc.) were 2 decades ago, is there a more likely outcome of this effect in the 2008 Presidential Election?

What if, in several circles, voters, especially white males, plan to vote for Barack Obama, but aren't sure how their co-workers, neighbors, bowling and softball league members, etc. would respond to it? What are the politically correct things to say in those social circles and might they be very different than in California, a fairly progressive state throughout?

I believe this effect will have some role in the outcome, but it will be very difficult to determine the level. There will be 2 other forces moving in the same direction that will be difficult to separate out. First, I still believe that most pollsters are underestimating the youth vote. The demographic between 18-30 is substantially less likely to not have a land line home telephone which is the predominant method of polling. While pollsters are aware of this fact I do not believe they have sufficiently accounted for it. Secondly, the Obama Ground Game and the passion of Obama voters, in particular young and/or African American voters will probably produce a turnout that is being understated by pollsters 'likely voter' totals.

As an aside, I am not sure I agree with the depth of the findings of the UW study. I believe some of what they experienced was a greater turnout among voters who favored Barack Obama. I think this will also play out in the general election as mentioned above, but although real, I don't think the Reverse Bradley effect was as large as they claim during the primaries.