2004 election revisited, part 6: Florida

Monday night’s GOP debate in Tampa ( NBC) may not have produced much warmth or light, except for Ron Paul’s comments on the Strait of Hormuz. But it further highlighted reason to look back at the 2004 campaign. These guys are going to need all the help they can get in the general election. Citizens United notwithstanding, half a billion in paid political ads goes only so far when the the other ticket is Newt Gingrich. He says his message is that of Washington outsider because he was detested in Washington.

Gingrich is campaigning on a claim to be about “changing Washington.” It’s like the fat, intemperate, unreliable old Falstaff yelling, “They hate us youth!”

Shakespeare

Déjà vu all over again

The issue went nowhere in the political press at the time, but there were striking anomalies in the Florida vote count in 2004. The headlines were different from 2000. The problems were less blatant. No tiny white-collar mob of Republican congressional staffers and lobbyists raised fists against the vote counters. But even after all the attention directed to Florida’s electoral process in 2000, there were still problems in 2004.

Take a look at the numbers on the ground, including party registration. Unlike South Carolina, Florida registers voters by party.

Florida counties 2008 election

Of Florida’s 67 counties, in 2004 Republicans constituted more than 50 percent of registered voters in only eight. This fact might surprise readers who get their information from the national political press, which represents Florida as a red state. But eight it was:

• Collier
• Indian River
• Martin, Clay
• Okaloosa (57.2%)
• Santa Rosa
• St. Johns
• Walton (barely)

Population of the eight red counties in 2000:  1,083,846.  Population of Florida:  15,982,378.

In contrast, Florida had 31 counties where Democrats constituted more than 50% of registered voters. In 21 counties, Democrats constituted more than 60% of registered voters. In thirteen of them more than 70% of registered voters were Democrats, and in four of them more than 80% of registered voters were Democrats.

Population in preponderantly Democratic counties:  2,700,000+.

This is a red state?

In 28 counties where neither party registered more than 50%, nineteen had more Republicans and nine had more Democrats. The biggest plurality county was huge Miami-Dade (pop. 2.3 million, 43% Dem). The smallest was Highlands (pop. 87K, 45% GOP).

Population in counties where GOP registration was heaviest, over 50 percent, totaled less than majority-Democratic Broward County alone. Population in counties with a less lopsided Democratic majority totaled 8.4 million.

A bigger anomaly

The biggest divide between Florida counties in the 2004 election was not red and blue but touch-screen and op-scan.

As in voting machines.

Fifteen counties used touch-screen voting machines, produced by ES&S or Sequoia. The other 52 counties used paper ballots, BUT not counted manually. Instead, the paper ballots were processed by optical-scanning equipment similar to that used by supermarkets, manufactured by ES&S, Diebold and (in one county) Sequoia.

Optical scanning in voting has been used for years, generally without the checking that turns up mistakes about 5 percent of the time in supermarket scanners. Mathematician and independent researcher Kathy Dopp tabulated differences between touch-screen counties and op-scan counties.

The difference? A simple and blatant pattern:

• In touch-screen counties, the county’s vote for president went with its majority party almost always.
• In op-scan counties, the county’s vote for president went opposite to its majority party most of the time.

If this sounds like a small difference, it’s not. Whatever problems the touch-screens had, 14 out of 15 counties using touch-screen equipment had an outcome at least in line with registration. Counties with more Republicans went Republican. Counties with more Democrats went Democratic. Plant a tomato, get a tomato.

Of the 52 counties using op-scanned ballots, 21 voted in the direction predicted by their voter registration–fewer than half. The other 31 counties went opposite their own voter registration. The kicker is that almost always, they went to Bush.

In the 21 op-scan counties where the vote ran with party registration, it was often skewed. Somehow Democrats there did not vote Democratic, and Kerry also picked up NO percentage from independents and unaffiliated–in a national election where the independent vote trended toward Kerry.

How it works

If an operative wanted to help a candidate win, in a state like Florida with many counties, the way to do it would be subtly so as not to affect the outcome of any individual county. That way, no local challenges would be provoked; the only way to examine the outcome would be to challenge the entire state. A few hundred or a few thousand votes in a lopsided county would not be missed, or suspect.

This process would be aided by the predominant media focus on red and blue.

Access to county statistics on population, demographics, and voter registration is already in the hopper, remember. Source code does not control turnout, but the political experts could weigh in on that little problem; look at the lines in big touch-screen counties inadequately supplied with voting machines, the problems with provisional ballots and early voting, the misleading flyers and robo-calls, etc. In the op-scan counties, I wouldn’t need help with turnout; I would need primarily to be able to work without scrutiny.

This is not to say that touch-screen machines are off the hook. If as a shady operative I wanted that badly to help my man win, odds are that I would overreach once in a while.

Florida counties 2004

More later