With his graphics wizardry, Prof. Charles Franklin has posted a piece at pollster.com looking at the percentage of African-American voters, the percentage of white voters who voted for Obama, and how they correlate.
Click for a larger view.
Prof. Franklin notes that below 20% of A-A participation, there's not much correlation between white vote for Obama and A-A turnout voting pool. However, once you reach 20% A-A voting, (for whatever reason) the white vote tends to drop precipitously. To help illustrate that, I've added a red line at the 20% range to see it visually. Look over to the right of the red line. The falling curve represents fewer and fewer white voters for Obama.
While it's arbitrary to pick a number to represent what an 'expected' chunk of voters should look like, the exit polls tell us that nationally there were 74% self-identified white voters, 43% of whom voted for Obama (.74 x .43 = .32). I therefore set a purple line at around that number (30%), below which is a "below average" "white voter for Obama" number. [You can set it at 40 to correspond to the % of white Obama voters with no adjustment, and see the same right lower quadrant. I arbitrarily picked a lower number.]
If you combine them, you get the Deep South over in the right lower quadrant.
I have no intention of doing any South bashing, and I'm not making any assumptions about the apparent correlation. I simply note that that's where the Democratic Party still has work to do, along with OK, AR and TX. I acknowledge all those who are fighting the good fight there (hey, we can't all be HI and VT; they have it easy.)
But for now, that right lower quadrant is the GOP's regional base; by 2012, TX will be a swing state.
And the GOP? Beyond that right lower quadrant, they've got more work to do than we do, if they ever want to be a national party again.