Florida

And then we have Florida. Let's try to sift through what we know about what happened in Florida on November 2. A lot has been made about heavily Democratic counties voting overwhelmingly in favor of Bush. But in some of those cases, those districts might be heavily Democratic -- but they're Dixiecrat counties, counties which tend to be incredibly conservative despite heavy Democratic affiliation. Yet that doesn't explain all the irregularities. Far from it.

So let's go through some of what we know and what's a little off.

1) Opti-Scan Counties vs. Touchscreen Counties
When this was first sent my way, I was alarmed by how much larger the actual Republican vote totals were for Florida counties than what would be expected if everyone voted for the candidate of their party registration. We're talking increases of 450% in some cases. In some counties, particularly those in the conservative Panhandle and near the Georgia border, it's very plausible that a county that's extremely Democrat would vote for George Bush -- it's the Dixiecrat effect, where many Southern Democrats would find the "Massachusetts liberal" an unsavory choice.

But not all of the counties which experienced enormous surges in Republican vote over party registrations were Dixiecrat counties. More troubling: look at the comparison between the counties that used Touchscreen and those that used Optiscan. That's right: in Touchscreen counties, there was never more than 50% of an increase in Republican votes from the votes expected due to party registration. So then why would there be that small an increase in those counties, yet consistently larger increases -- of 200%, 300%, even 450% -- in counties that used the Opti-Scan systems? Good question.

You can claim that those Opti-Scan counties were Dixiecrat counties, full of conservative Democrats, and that explains the big surge. But then why wasn't there a huge surge in Republican votes in more conservative Touchscreen counties like Nassau, on the northeast coast (Bush had less than 50% increase over the vote expected from party registrations), or Lee in the Southwest -- a friend of mine who grew up there, in Ft. Myers, described it as very redneck. (Bush only had a 24.2% increase). (Here's a helpful guide to Florida counties.)

A Penn professor wrote on a listserv I'm on:

Right now, it appears one of the most glaring statistical inconsistencies nationwide is the behavior of voters in Florida in terms of the voting machine used. In counties which used Direct Record Electronic machines, Dem/Rep ratios in final vote counts were consistent with Dem/Rep registration ratios. In Optical Scan counties of comparable size, there was a dramatic increase in turnout for Republicans.

Again: a reminder about Opti-Scan systems. They do leave a paper trail, but Diebold and Sequoia and similar companies manufacture the machines that read the Opti-Scan cards -- the paper ballots marked with ink. And those machines -- and the counting computers that had no protective software beyond simple Microsoft Windows -- could potentially have been hacked. (You'd be amazed how easily a hacking could occur -- Thom Hartmann explains it here.)

If someone could explain to me this discrepancy, I'd appreciate it. Why was there such a surge in Republican votes in counties using Opti-Scan voter machines that didn't exist in those counties using Touchscreen technology?

From what I hear, there are statisticians right now trying to answer that very question.

Thom Hartmann wrote a helpful piece. Check it out.

2. Something fishy in Broward County

However, as alarming as the discrepancies in the rural Florida counties might be, what happened in Broward County seems even more disturbing.

How to explain this: Collier and Broward Counties are in the same area of Florida, and have exactly the same increase in Republican turnout, while there's a vast difference between them in Democrat turn out. Someone on Mark Crispin Miller's listserv pointed out that "this can't be a Dixiecrat effect -- since Collier county, the more rural, actually saw a 46% increase in Democratic votes. broward, a bastion of Democrats, showed only a 16% increase in Democratic votes." This when Collier saw an increase of 38.2% in Republican votes while Broward saw a 37.1% gain. That's a bit curious. Especially when, as the emailer pointed out, "If Broward had the same vote swings as the demographically and spatially similar Collier county, Kerry would have 140,000 more votes. Such a figure alone swamps the rural vote."

The turnout in Broward of registered voters was 66.7% in 2000 and 66.7% in 2004. Somewhat unlikely given how fired up people were. I'm just saying, I'd like an explanation beyond, "Maybe the GOP GOTV effort was better in that county."

Broward was also where election workers discovered that the shiny new Touchscreen machines were counting backwards. Officials discovered that the software used in Broward could only handle 32,000 votes per precinct. After that, the system started counting backward.

3. Registrations

The Dixecrat theory, again, might explain how counties with large Democratic registration went for Bush. But it doesn't explain these statistics from more traditionally liberal, urban counties in Florida:

Broward (Gore2000, Kerry2004) E-Touch
The number of registered Republicans increased 20735
The number of registered Democrats increased 90551
The number of votes for Bush increased 58,892
The number of votes for Dem candidate increased 54,030

Hillsborough (Bush2000, Bush 2004) E-Touch
The number of registered Republicans increased 38787
The number of registered Democrats increased 44128
The number of votes for Bush increased 60,870
The number of votes for Dem candidate increased 41,335

Palm Beach (Gore2000, Kerry2004) E-Touch
The number of registered Republicans increased 5064
The number of registered Democrats increased 40856
The number of votes for Bush increased 21,282
The number of votes for Dem candidate increased 5,298

Pinellas (Gore2000, Bush2004) E-Touch
The number of registered Republicans DECREASED 8578
The number of registered Democrats increased 12,668
The number of votes for Bush increased 37,805
The number of votes for the Dem candidate increased 21,473

Hernando (Gore2000, Bush 2004) Op-scan
The number of registered Republicans increased 4953
The number of registered Democrats increased 4977
The number of votes for Bush increased 9,491
The number of votes for the Dem candidate increased 2,362

Orange (Gore2000, Kerry2004) Op-scan
The number of registered Republicans increased 30557
The number of registered Democrats increased 57818
The number of votes for Bush increased 56,872
The number of votes for the Dem candidate increased 51,810

Osceola (Gore2000, Bush 2004) Op-scan
The number of registered Republicans increased 10320
The number of registered Democrats increased 14748
The number of votes for Bush increased 6,600
The number of votes for the Dem candidate increased 2,114