So, with all the news of the Florida counties having large Republican swings, I decided to do the most scientific study I could to see if it was true.
Here's what I did- I found the percentages that Bush and Gore got in 2000 and Bush and Kerry got in 2004. I found how much Bush won or lost by (i.e. if Bush won a county 52-47, he won by 5 and if he lost 47-52 he "won" by -5) for every county in both elections. I then subtracted the difference in 2004 from the difference in 2000. So these numbers represent the increase in Bush's margin of victory from 2000 to 2004 per county. I then seperated the numbers by what type of machine they used in 2000 found here and 2004 found here. Here are the results . . .
Former Punchcard turned Electronic:
Nassau 7
Sumter 14
Pasco 10
Hillsborough 4
Pinellas 5
Sarasota 3
Lee 3
Collier -2
Miami-Dade -1
Broward 7
Palm Beach 5
Indian River 3
Average 4.83
Optical-Central turned Electronic
Lake 6
Charlotte 4
Average 5
Lever turned Electronic
Martin 3
Average 3
All 15 Counties using Electronic Voting Average: 4.73
Former Punchcard turned Optical-Precint
Wakulla 8
Jefferson -1
Madison 2
Dixie 20
Gilchrist 15
Duval -1
Marion 7
Osceola 9
Hardee 17
DeSoto 5
Highlands 8
Glades 5
Average 7.83
Optical-Central turned Optical-Precint
Jackson 9
Gadsden -6
Liberty 16
Gulf 15
Franklin 10
Taylor 7
Hamilton -1
Lafayette 13
Suwanee 10
Bradford 12
Levy 15
Hendry -1
Okeechobee 9
Average 8.31
Paper/Hand turned Optical-Precint
Union 21
Average 21
Stayed Optical-Precint
Escambia 4
Santa Rosa 8
Okaloosa 6
Walton 11
Holmes 16
Bay 9
Washington 16
Calhoun 13
Leon -3
Columbia 14
Baker 16
Clay 5
Putnam 14
St. Johns 4
Alachua 3
Flagler 8
Volusia 6
Hernando 10
Seminole 4
Orange 2
Brevard 8
Polk 9
Manatee 6
St. Lucie 5
Monroe 0
Citrus 8
Average 7.77
All 52 Counties Using Optical-Precint Voting Average: 8.32
Difference between Optical-Precint Counties and Electronic countes 3.59
Scientifically, this isn't nearly a large enough difference to claim any fraud from this evidence. Especially sense urban counties are more likely to be electronic.
This is assuming that the 2000 election was generally free of fraud. If you believe there was massive fraud in 2000, this study does nothing to disprove that. This survey shows that there was not a statewide change from 2000. Some of the larger individual changes (Especially Dixie-20 and Union-21) do seem rather large, but those two counties only had 7,000 and 5,000 votes cast respectively, so large swings are more reasonable.
Well that's what I've found. Take it as you will.