[I'm surprised nobody's doing this yet, so I emailed MoveOn/ACT - what do you think of the idea?]
Hi,
Florida has some counties in which Democratic registration numbers and Kerry votes appeared to be reversed (I assume you guys have heard about this; I wouldn't be surprised if it took place elsewhere as well, like Ohio). For example, a county with 4,000 registered voters, a 3:1 Democratic-registration edge, and 3,000 voters on election day would probably be expected to produce about 2,250 Kerry votes and about 750 Bush votes. Maybe Bush gets a few more if it's rural Democrats in the reddening South, maybe Kerry gets a few more if Democrats are angrier at the incumbent. But there are more than a few counties in which Bush seems to have made astonishingly effective inroads into the Democratic base - and in light of the volume of screwy votes and discrepancies between exit polls and vote tallies, especially on electronic machines/servers, this begs for some investigation, and I think you guys are uniquely positioned to do it.
Why? You were doing GOTV work. You guys have lists of voters. You guys should have an idea of who voted and where and who they voted for. And I think you should pick a county or two or five with a relatively small number of voters, a distinct registration disparity and a distinctly different vote disparity, and call around to see if you can actually find the number of people who allegedly voted for Bush this time. Hell, do the same for Kerry, so you can tell how many voters you're finding, and how accurate your lists and expectations are.
If you find out Bush actually got those people to vote for him, fine. But if you can't find the number of Bush voters, it'd mean something pretty serious and bad.
Just a thought.
Thanks,
Chris
cc: MoveOn, ACT, mydd