In response to my piece on election conspiracy theories, many readers have asked about supposed irregularities in Florida counties that use optical-scan voting machines. In Baker County, where 69 percent of registered voters identify as Democrats, 77 percent of voters went for Bush. Dixie County: 78 percent registered Democrats, 69 percent for Bush. Franklin County: 77 percent registered Dems, 59 percent for Bush. Holmes County: 73 percent registered Dems, 77 percent for Bush. (USTogether.org has more charts and graphs of this data here and here.)
These are just the most egregious examples. If we trust the exit polls, about 90 percent of voters nationwide vote based on party affiliation. That means that, according to USTogether’s calculations, if Floridians voted according to party affiliation in the state’s 52 optical-scan counties, the GOP would lose about 600,000 votes—perhaps not enough to swing the election, but a huge number nonetheless. The theory goes that since all eyes were on the counties with new touch-screen machines, it would be far easier to futz with the numbers in the less-scrutinized optical-scan counties. Optical-scan machines are used to count paper ballots, and the results are then fed into a computer. If someone could hack into the computer, they could surreptitiously change the vote count.
The big problem with this theory is that this year’s results match those from 2000. (And with the exception of Dixie, which used punch cards in 2000, all of these counties used optical-scan machines four years ago.) In 2000, Baker County had 83 percent registered Democrats, and 69 percent of the county’s voters went for Bush. Dixie County had 86 percent registered Democrats, and 58 percent went for Bush. Franklin County: 81 percent registered Dems, 53 percent for Bush. Holmes County: 83 percent registered Dems, 68 percent for Bush. (For complete 2000 results in Florida by county, look here. For 2000 results as compared to voter registrations, look here.)
While each of these counties had a lower percentage of Bush voters in 2000 than in 2004, the 2000 election was much closer than this one. Each of these counties also appears to be moving toward the GOP. In all four, there is a lower percentage of registered Democrats and a higher percentage of registered Republicans in 2004 than in 2000.
Of course, you could argue that this similarity to the 2000 results just points to an even bigger conspiracy. But that’s a bit, well, crazy. Take a look at the geography: All four counties are either in Florida’s panhandle—known by some as the Redneck Riviera—or the northern part of the state. Like their neighbors in Georgia and Alabama, northern Florida voters tend to be very conservative. Baker, Dixie, Franklin, and Holmes counties are represented in the House by two Republicans and a Blue Dog Democrat who lists his No. 1 issue as “Second Amendment rights.” Democratic registrations here are more an artifact of history than evidence of massive fraud.
Update, Nov. 10, 2004: For a thorough debunking of the Florida op-scan conspiracy based on an examination of 28 counties in the 1996, 2000, and 2004 presidential elections, check out Yevgeny Vilensky’s analysis here. Also, it seems there’s not much to another alleged scandal that readers have been asking about— Keith Olbermann’s report that there were 93,000 more ballots cast than registered voters in Cuyahoga County, Ohio. You can find an explanation for the supposed disparity here.