There is absolutely no question that the Republicans have a problem with voting. Despite their rhetoric as being the most “American” of the two major parties in the US, they clearly have little respect for that which is vital to the democratic process. The Republicans have attempted to use the law to effectively disenfranchise sectors of the voting population unfriendly to their goals. Happily, most of these attempts fail, but sadly, the damage is usually done anyway.
Mickey Duniho, a former NSA computer programmer and voting activist, claims that he has found evidence that might indicate that “vote flipping” is occurring on a wide scale, both in his home state of Arizona and nationwide. His methodology comes from an analysis done by Francois Choquette and James Johnson who attempted to show that the Republican primary elections had been artificially swung in favor of Mit Romney.
I downloaded the voting data for the 2008 elections in Michigan and compared the voting shares of Obama and McCain, ordered by precinct size. The graphic is up on the left, you can see that the larger the precinct, the more likely it is to swing toward McCain. Duniho’s claim is that the relationship would be flat if there were no vote flipping.
Of course, this defies intuition, particularly here in Michigan. We would expect that larger precincts would actually swing the other way. Urban areas are far more Democratic than Republican, though anything is possible. The suburbs of Detroit could very well house a lot of Republicans.
To be clear, I’m not sure this demonstrates vote fixing at all. The point here is that large precincts are likely quite different from small precincts, and these differences could be graded by size. Duniho claims that he ran the analyses controlling for things like income, age and gender distribution and that the relationships did not change. How he did this is fairly obscure.
It’s also fairly suspect when Duniho finds that these trends in percentages in small, local elections are flat. In all cases the number of precincts are very small.
Duniho’s aim was to demonstrate potential vote flipping in his home county and he may be able to show this, given the proper tools. Certainly, that the discussion is open should be an indicator. He certainly has the luxury of opening up the ballots and hand counting them. His methodology, though, and how this is supposed to definitively demonstrate vote flipping in the absence of what could be important information, is fairly unclear to me.
I like that he tried though I admit I am not convinced. If there is evidence for electronic vote flipping, then it needs to be exposed. I fear, however, that successfully bringing this to light could be a very hard road.
The graphic below is somewhat more interesting, however: