The white Democratic vote in the South

I know the South is racially polarized and hostile to Democrats but I was still shocked that Obama only received 10-15 percent of the white vote in Alabama, Mississippi, and Louisiana:

Maybe I'm naïve, but I would have said 25-30 percent.

Comments

[Ben]

any interesting correlation in the converse? That is, were African-Americans more likely to vote for Obama in states with bigger Black populations?

[Jinchi]

Eliminate Louisiana, Mississippi and Alabama (and possibly Georgia) and your trend goes away.

This looks much more like a regional trait than a purely racial one.

[Andrew Biggs]

What kind of R-squared are you getting from the data above? The relationship seems to hold in extreme cases, but there's a lot of variation in the white vote for Obama even in states where the black share of the vote is around the same.

[Brendan Nyhan]

Per the above two comments, the graphic is from Charles Franklin (see link above), not me, so I'm not putting particular stock in it (nor do I know R-squared etc.). It looks like he just fit a locally-weighted regression to the data.

[ikl]

Kerry didn't get 25%-30% of the white vote in those states. He wasn't even close in AL and MS. I'm not sure why you would expect Obama to do so without even running a campaign there.

The Louisiana result surprised me a bit (Kerry got 24% of the white vote there). MS and AL not at all. However, given demographic changes in LA since 2004, I'm not sure how much faith one should put in the exit poll numbers there . . .

And, of course, as Jinchi points out, the graph is highly misleading - whites in the deep South don't like Obama. Percentage of blacks in the state isn't driving the voting behavior - if this were causal, Obama would have done much worse in MD, DE, MI, IL, and NY.

[sarang]

As Tom Schaller pointed out in an old NYT editorial, MS would be blue, and AL/GA/SC competitive, if 25-30% of whites there voted Democratic.