I will grant you that McCain will win ND, AR, SD and WV.
Obama has a fighting chance in MT and AZ, partly because those are non-Confederate states and racism isn’t so strong there. But let’s assume that McCain wins those too. MT is too insignificant and McCain has the edge in all polls in AZ.
Obama will win NM without a question, a few recent polls have shown him with a double-digit lead there.
Your hypothesis boils down to this. Obama starts with 243 solid electoral votes and McCain starts with 151. There are ten “swing states” in play. They are CO (9), FL (27), NV (5), PA (21), VA (13), OH (20), MO (11), NC (12), GA (15), IN (11). To lose, Obama needs to get less than 26 votes from this group.
Now, Mason-Dixon poll is the only recent poll that shows Obama with less than 50% in CO or NV. That’s 14 points right there. More importantly, most polls also show more than 50% for Obama in PA. A win in PA would push Obama over 269 electoral candidates, even if each undecided voter in every one of the remaining 7 states decides to vote for McCain.
Will they? Just because they are white and undecided, does not mean that they are going to vote for McCain. There’s no good reason to think that they would vote anything other than 50:50 – if they show up at the polls at all.
The most interesting thing that might happen is a combination of Bradley effect and unprecedented Democratic turnout. In which case all these polls are doubly unreliable. (First, because they assume that everyone will vote the way they say they will vote, second, because they assume certain demographic characteristics of voters when they construct the sample) Depending on which factor dominates, we can see all kinds of entertaining outcomes. Bradley factor could give New Hampshire and New Jersey to McCain. Turnout bias could give West Virginia and Louisiana to Obama. If those kinds of results start showing up on election night, it will be a long wait to find out who’s the winner.
