A reader writes in, “I’m interested in what the scatterplot quants have to say about this article (and related scatterplots) in the Atlantic:
http://www.theatlantic.com/national/archive/2011/05/the-geography-of-hate/238708/
So, what say you to the claim that “America’s racist groups concentrate in certain regions — and their presence correlates with religion, McCain votes, and poverty”?
5 Comments
I find the various analyses prima facie plausible but I’m a bit concerned about the dependent variable. SPLC’s definition of a “hate group” goes well beyond violent extremists who commit hate crimes to include relatively mainstream social conservative organizations who attempt to influence policy through regular politics. For instance, they include the American Family Association as a hate group. I know a fair amount about this organization (I’ve published book chapters about it) and it’s not exactly commensurable with Aryan Nation.
Furthermore SPLC is not exactly impartial on this as (just like many of the organizations they criticize) they have an economic interest in the politics of outrage as creating hysteria about threats to your donor’s values is an excellent way to about fund-raising. (Hunter has a good chapter about this in Culture wars). Not surprisingly SPLC has shown a continuous rise in hate groups every year for the last decade. They almost have to as how else are they going to come up with stuff to put in the all caps underlined typewritten direct mail fundraising such organizations rely on?
If there was a facebookesque like button in WordPress, I would “like” this comment.
Voting for McCain wouldn’t necessarily have been a good indicator of racism as a primary cause.
Voting for McCain/Palin is rather a different story, as a question of both tone and substance.
As for the last, “Keep ‘em poor, keep ‘em dirty, and keep ‘em too busy to know they’re getting the sh*tty end of the stick” has been a functioning political doctrine since long before now.
Which leaves religion. I’m suspicious.
This is an interesting and provocative article. And the author does qualify that correlations do not imply causation. But I’m surprised that he didn’t have his mapping friend do a spatial analysis — or to use the fancy term “Exploratory Spatial Data Analysis” (ESDA). Such an analysis would have better revealed clustering patterns. Also I think counties would have been a more appropriate level of analysis.
What I noticed is that the two very highest states on hate groups — Montana and Mississippi — were not on any of the regression lines and have much higher scores than would be predicted by any of these factors. In fact, for most predictor variables, one or the other of the states is relatively low on the independent variable in question and the other isn’t exceptionally high. So the most extreme levels for “hate groups” are not explained by these factors.
Otherwise, what we are seeing is that the exclusionary groups SPLC calls hate groups are very highly correlated with partisan politics, religion, and class — which we already knew.
% urban would also correlate pretty well.
One Trackback
[...] ask a scatterbrain: geography of hate « scatterplot [...]