... I wrote up a post on statistics and the assault weapons ban
. Someone else with a bit of an education has applied statistical reasoning to the list of car dealerships being shut down
. The theory is that the Obama administration is shutting down dealerships (in connection with the GM and Chrysler bankruptcies and bailouts) that are political opponents (ie, Republicans or who contribute to Republicans). It turns out that there's nothing to most of the conspiracy theories...
... but Clinton donors tend not to get shut down, at a p value of .125. That's a 12.5% chance that the correlation with Clinton donors was attributable to chance, compared to p values of about 50% with other variables. As the person who did the research points out, a 5% chance is the "gold standard" for this sort of work, but 10% is sometimes used, and 12.5% is unusual in the context of the others. It's not proof, but it's food for thought.
I should mention also it's not clear what other variables were being measured. Looking at the geographic data, that could perhaps explain the correlation -- a lot of political attention from Clinton during the primary campaign plus a decision making criteria emphasizing "one dealer per region". There's nothing solid here -- yet. But these initial results do suggest that there is something going on that should be explained, even if that explanation is innocuous.
UPDATE: Lott doesn't think there's anything to this claim
as it applies to Republicans, but doesn't say anything about the possible Clinton donor angle.