This will be my last post on this subject, as, after I delved further into the data, I found no reason to believe that Clarkson's original claim that the correlation between precinct size and republican votes was evidence of some kind of massive Republican electoral fraud across the country. In essence, my earlier thought that this was likely caused by other covariates was confirmed.
I did find some other interesting correlations, but none are really that surprising, such as less densely populated precincts vote more Republican. One of moderate surprise was that precincts with more persons under 18 voted more Republican, but this may be due to gentrification factors (young, childless voting more democrat?)
Anyways, here's a link to my last blog on the subject here.
Next I believe I'm on to Kansas Education outcome data (a much more exciting subject).