by Doctor Science
There was a lot of coverage of the Arkansas and Kentucky Democratic primary results yesterday, because Obama was officially unopposed but only got about 60% of the vote. In following up various discussions, I ended up looking at voter turnout in 2008 as compared to 2004. To my surprise, 30% of the states had lower turnout in 2008 than in 2004 -- including Arkansas and Kentucky. Except for Arkansas, all the states where turnout declined are much whiter than the country as a whole.
R-squared for this relationship is .43, which is fairly significant by social science standards. For comparison, I regressed "change in turnout" to "percent of vote that went to Obama", to address the possibility that Obama voters in general were more enthusiastic and willing to turn out. That R-squared is .13 -- or what we technically call "pffffffft".
I am willing -- nay, eager, nay, begging -- for someone with more data visualization skillz to make this into an infographic map, or at least to figure out how to label the data points with the state abbreviation.
I got the idea to do this from a discussion in the comments at Ta-Nehisi Coates' post on the continued salience of racism as an explanation:
nonotford:[brackets are my addition] Here is the map they're talking about:
Why the distinction between the Upland South/Appalachia and The White South as a whole? The areas of the South in which Obama did worse than Kerry (http://www.nytimes.com/interac...), just happen to be the areas with fewer black people (http://www.censusscope.org/us/.... ). Those blue areas in the first map are essentially the same as the red/purple areas in the second. If it weren't for southern blacks, the first map would've shown the entire South as red.
The implication here is that the increased turnout among the black community in the deep south swamps out the inherent racism of [white] southern Democratic Party supporters (in other words, that [white] southern Democrats are just as racist as their Appalachian counterparts, but you can't see it in the data because those [white] Democrats live near actual black people, who canceled their votes out).
The evidence I've collected suggests that nonotford and Craig may be largely right: white racism (seen as Kerry voters who didn't vote for Obama) was a factor not only in Appalachia but in the South as a whole, but the effects were masked in the Deep South by increased turnout of a large black population.
What I'm also seeing, to my surprise, is that there may have been a drop in white turnout nationwide, not just in the South. I hadn't expected that there would be so many states (as in, *any*) with lower turnout in 2008 than in 2004. Frankly, it looks to me as though there was a really substantial number of white Democrats who'd voted for Kerry but who couldn't make themselves vote for Obama. They weren't going to vote Republican, so they just didn't cast a vote for President.
I was also surprised by the lower turnout in Oregon, with its mail-in balloting. Someone who knows more about the data may want to investigate further.
These data really make it clear why voter suppression is surging: increased turnout of black voters was a huge factor in 2008.
This scene depicts Tory and Whig agents, both attempting to bribe a farmer to vote for them. The crowd outside the tavern is visible in the background. In a reference to the antisemitism of the crowd behind, a Jewish peddler is being employed by another agent who is offering jewels and ribbons to the wives of voters.
On the margins of the composition a soldier (left) and two old sailors (right) represent uncorrupted patriotism. The soldier peeps out from behind a now-impotently decorative figurehead depicting the British lion devouring the French fleur-de-lis. A woman sits on it looking at her bribes. The sailors on the right are re-enacting a naval victory using pieces of broken clay pipe.