I think you may be right about the Bradley effect. People who would feel uncomfortable about admitting they intend to vote for a black candidate have plenty of other canards they could use to justify their lack of support for Obama (e.g., the crypto-Muslim angle).

Another theory I've read is that the Bradley effect (which hasn't been observed since 1996, BTW) may be caused by people legitimately intending to vote for the first black candidate to run for Senate, or the first black candidate to run for governor of their state but at the last minute admitting they like the other guy's policies and experience better. The author of this article said that if this is the case, we may be seeing a Bradley effect again _because_ it's the presidency. But I don't think many people are going to suddenly decide they like McCain's policies at the last minute. And I suspect that anyone who freaks out over Obama's inexperience will see Sarah Palin's name there under McCain's and get freaked right back toward Obama.

Personally, I think the economic climate will push the undecideds toward Obama, and on top of that, Obama's lead in organization will push the undecided voters his way at the time of the election. Then there's the edge in ad spending. Which isn't nearly as big as some pundits say, because the RNC has a lot more money to spend than the DNC, and their support for McCain could bring it equal to or slightly better than Obama's funds, if they go all in. But then, the RNC also has other candidates to support, so they're not going to put all of it behind McCain when they have a lot of close congressional campaigns that they have a better chance of winning. And more importantly, O. doesn't need to massively outspend M. to win the voters who can be swayed by advertising--even if he were just to tie, he would still be massively ahead of Kerry and Gore. And Obama is budgeting his ad money a lot more carefully than McCain, who seems to just now be realizing that this isn't a sprint.

So a lot of factors favor a reverse Bradley effect for Obama, not the least of which is that he's running against a markedly inferior opponent whose ideologies over the last 25 years have been proven wrong over the last 2 weeks. But the biggest of these (in my estimation) will be the ground game--the drives to get new voters registered and to get them to the voting booths. Not because the votes he gets this way will be more than the votes he gets by _showing_ the American people that he is the better choice, but because the votes he gets in this way are the ones that are most likely to be completely overlooked in all the polls.

National polls are meaningless.  Only state polls matter...

Overly dismissive, IMO.

Yes, state polls are much better for making accurate estimates of the EV vote.  Trouble is, they lag the sentiment of the electorate much more than the national polls do.  So, you've got your choice of imperfections.

There is a correlation between national and state polls.  If the national polls trend one way, most state polls are likely to follow -- when the pollsters get around to conducting them.  The correlation isn't perfect, of course, but it's good enough for the national polls to be of significant value.

He broke 50. I can't wait for the next liberal to come on the thread and ask why the hell he can't break 60.