Two more for the road

Updated I'm not one for conspiracy theories, I'm really not. But I'm becoming convinced that there were genuine attempts to manipulate the results of the 2004 presidential election. It's not just that there are oddities and anomalies and statistical quirks; you'll find those in every election where enough people take part. It's not just the cockups with machines or the confusions about voting status; those, too, will be a part of every national election. It's rather than over and over again, in case after case, the oddities, anomalies, quirks, and cockups all went one way: They favored George Bush.

Where are the anomalies that favored Kerry? Where were the problems in the heavily Republican districts? Where is the statistical analysis showing an odd Kerry surge somewhere? Are there any? I don't mean anecdotes about single precincts in New Mexico or some such, I mean an actual pattern. Show it to me and ease my mind.

Now, note well that I'm not flatly saying Bush stole the election - unlike 2000 - because I don't know. The things I've seen, as far as I'm aware, would not, it appears, change the final outcome of the election. It's still possible that they plus fraud undiscovered did just that, but I don't know. What I am saying, what I'm coming to believe, is that some things were arranged so that if the election had been even closer (let's not forget that despite the bull about a "sweeping" or "decisive" victory, the election was close), if it had been a real squeaker, the manipulation would have been enough to turn a tiny loss into a tiny victory. That is, don't outright steal the election, that might be too obvious. Just give your guy a little boost that could get lost in the shuffle. Shorter version: They were prepared to steal the election, it just may not have proved necessary.

These are the latest two items pushing me in that direction, both via Buzzflash.

The first, from ComputerWorld for November 18, says that a team at UCal-Berkeley led by Michael Hout, an expert on statistical methods, has

uncovered statistical irregularities associated with electronic voting machines in three Florida counties that may have given President George W. Bush 130,000 or more excess votes. The researchers are now calling on state and federal authorities to look into the problems. ...

According to the study, counties with electronic voting machines were significantly more likely to show increases in support for Bush between 2000 and 2004 compared to counties with paper ballots or optical scan equipment. This change cannot be explained by differences between counties in income, number of voters, change in voter turnout, or size of the Hispanic/Latino population, said Hout. ...

"No matter how many factors and variables we took into consideration, the significant correlation in the votes for President Bush and electronic voting cannot be explained," said Hout. "The study shows that a county's use of electronic voting resulted in a disproportionate increase in votes for President Bush. There is just a trivial probability of evidence like this appearing in a population where the true difference is zero - less than one in a thousand chances."

What the researchers did was to examine the voting patterns of counties that did not use e-voting machines and by accounting for factors such as noted above predicted how Bush would do in the three counties as compared to how he did in 2000.

Buzzflash, in its own report, provided some details of the numbers:

- Broward County: The prediction was for Bush to get 28,000 fewer votes than in 2000. Instead, the machines tallied a 51,000 vote increase, a gain of 79,000 votes over predictions.

- Palm Beach County: A loss of 8,900 votes as compared to 2000 was predicted; instead Bush gained 41,000, a gain of nearly 50,000.

- Miami-Dade County: Bush was predicted to gain 18,000 votes but instead gained 37,000, an extra tally of 19,000.

Total net gain over predictions: 148,000

The actual gain, the team said, was somewhere between 130,000 and 260,000 votes, depending on if these were "phantom" votes just added to Shrub's total or if they were Kerry votes wrongly credited to Bush. Again, this would not have been enough to change the result but if Florida had been closer, it could have been. And that's what's got me worried.

The other item also involves Florida, this time optical scan equipment, about which I posted on November 7. The saga stars another unsung hero of what's left of our democracy: Bev Harris of BlackBoxVoting.org, who may have found another reason for the odd results those machines produced. The entire article, found here is worth a read, especially for the image of a tug-of-war over a garbage bag containing trashed public records. But this is the gist of it:

A "poll tape" is the phrase used to describe a printout from an optical scan voting machine made the evening of an election, after the machine has read all the ballots and crunched the numbers on its internal computer. It shows the total results of the election in that location. The printout is signed by the polling officials present in that precinct/location, and then submitted to the county elections office as the official record of how the people in that particular precinct had voted.

Harris went to Florida's Volusia County Elections Office on Tuesday, November 16 to see the poll tapes for the 100+ optical scanners in that county. However, the printouts she was given - apparently copies of the ones sent by the elections office to the Florida Secretary of State's office and used to tabulate the official count - were dated November 15 and were unsigned.

When she finally secured the original, signed, dated tapes from November 2 - which involved that tug-of-war over a garbage bag - Harris and her associates discovered something disturbing: The original (November 2) tapes and the later (November 15) tapes sent to the state did not match.

"The difference was hundreds of votes in each of the different places we examined," said Bev, "and most of those were in minority areas."

When I asked Bev if the errors they were finding in precinct after precinct were random, as one would expect from technical, clerical, or computer errors, she became uncomfortable.

"You have to understand that we are non-partisan," she said. "We're not trying to change the outcome of an election, just to find out if there was any voting fraud."

That said, Bev added: "The pattern was very clear. The anomalies favored George W. Bush. Every single time."

And that has been the pattern: "Anomalies" favoring Shrub. Every single time. Our future is being stolen from us, anomaly by anomaly.

Footnote: Media Matters for America (MMFA) has an item about media coverage of the Berkeley study about the e-voting machines (there really wasn't any) which also notes that

- at the request of the Oakland (CA) Tribune, an MIT political scientist re-examined the data and replicated the results, i.e., got the same answer, and

- a Princeton University professor of microbiology conducted an independent analysis, using different methods, that produced similar results to the Berkeley study.

Updated to include the Footnote and to correct a misunderstanding on my part as to the source of the range of 130,000-260,000 votes Bush gained.