Those faulty exit polls were sabotage in The Hill, or A Tour of the 2004 Exit Poll: What It Says and What It Doesn't, part one and part two on Donkey Rising.
What I haven't been able to find is anything with data to confirm this, one way or the other. CNN has an interactive page allowing checks of specific aspects of exit poll data, but that's no help for analyzing nationwide data, say, by county. And in any case, it seems that CNN changed the online data after the fact, so there's no telling what this means in terms of raw numbers.
Lawrence Lessig gives the answer, in Free the Exit Poll Data: the poll numbers are privately held, not publically available. Lessig calls for the data to be made public, so that it will be possible to find out why the numbers were so misleading compared to the final election tally. You'd think both sides would be interested in knowing what went wrong.
Terrific maps for visualizing the electionMaps and cartograms of the 2004 US presidential election results gives a wonderful set of maps showing "purple states" by county, with the sizes adjusted for population.
Other stories about voting irregularities:Outrage in Ohio: Angry residents storm State House in response to massive voter suppression and corruption (Michigan Independant Media Center): Protests on November 3 in Ohio over all the voting problems the state experienced. Includes lots of anecdotes about voters who experienced problems.
Surprising Pattern of Florida's Election Results (US Together): a comparison of party registration data to reported election results in Florida counties using different types of voting equipment. In counties using touchscreen machines, the percentage vote for Kerry matched the party registrations fairly closely; in counties using optical scan machines, there's a huge shift over to Bush votes, completely uncorrelated with party affiliation. The article includes a data table by county.
Evidence Mounts That The Vote May Have Been Hacked (Common Dreams): a text discussion of the US Together results, their correlation with exit poll results, and some discussion of possible explanations other than foul play (and why those reasons are unlikely to be the actual explanation).
Palm Beach County Logs 88,000 More Votes Than Voters (Washington Dispatch): Palm Beach County's official election results web site showed 542,835 ballots were cast for a presidential candidate while only 454,427 voters turned out for the election. Apparently they've since updated the web site to show numbers that add up. I guess this tells us how far we can trust the "official" numbers on the web site.
Tons of other links on the Op Ed News: Votergate 2004 page.
Bev Harris of Black Box Voting, Ralph Nader and others have teamed up for Help America Recount, a project to buy recounts in Ohio and other states. They're soliciting donations. I'd love to see recounts, but what they don't explain is where the money is going. What's involved in getting a recount, and does it cost money, or is this to pay salaries and expenses of the (volunteer?) people doing the counting, or what? The effort sounds like it might be a little disorganized at the moment.
Kerry Won. . . (Tom Paine.common Sense): Editorial about irregularities in various states. No new data, though.
[ 11:31 Nov 12, 2004 More politics/election04 | permalink to this entry ]