Electile disfunction

There's lots more stuff on "the election mess," as Keith Olbermann calls it.

Conspiracy theory: Daily Kos has a pretty good discussion, which tries to dampen "conspiracy theories." (By the way, can we please have a moratorium on references to aluminum chapeaus? That famous tin foil hat is old hat.) We see here a theme that recurs on other sites: The oddities found in small Florida counties, where many registered Dems voted for Bush, find a large measure of explanation in historical voting patterns. I've been acknowledging that point for some time now. I've also been voicing my opinion that if fraud took place in Florida, we should look at the GEMS central tabulator -- an all-too-hackable standard PC using Microsoft Access -- and not at the results of any particular county.

The Kos article, by one maddrailin, also makes this argument vis-a-vis the Cuyahoga county results mentioned earlier.

To a casual observer, Cuyahoga County seemed to have more votes than voters in some areas on Election Day - a 90,000-vote disparity being hyped in cyberspace under headlines such as "Stolen Election" and "Ohio Fraud."

The problem stems from an election canvassing report posted on the county's Web site, where the number of "ballots cast" in several suburbs exceeds the number of registered voters.

For example, the documents suggest that 13,939 ballots were cast in Beachwood, though only 9,943 voters are registered there.

But county officials say that's not how the data should be read. The "ballots cast" is not a reflection of the votes within a city's borders; the numbers also include votes in the congressional and legislative districts that overlap with the cities.

We shall see. I may have to concede some territory here. (See Salon's take on this situation, below.) But the county was Kerry-land anyways -- and as mentioned, I now believe that our main focus should properly be trained on the GEMS system, not on the oddities that crop up in specific counties. (Although some of those oddities remain very odd indeed.) A "red" precinct may prove to be a red herring, diverting us from the main issue.

A few of the comments posted in response to this Daily Kos story repeat themes that are starting to annoy me.

1. Once again, we hear that devoting attention to this issue is "bad politics." Well...screw that.

This is the post-election season, and nobody looking into this matter is running for office. I happen to think that a gutless Gus engages in "bad politics" when he frets over the ramifications of every syllable he utters. Maybe the best politics is for all of us to speak our minds about the facts as we see them, and to change our minds as new facts come in.

Is the c-word really so damaging, politically speaking? Throughout the 1990s, the right-wing devoted itself to conspiracy theories. Really bad, really stupid conspiracy theories: Whitewater, black helicopters, Soviet troops massing at the Mexican border (in 1994!), Vince Foster, the Illuminati, you name it. Were the rightists harmed politically by the dire tales they loved to spread? Or were they strengthened?

In the present case, the evidence (detailed below and elsewhere) gives us legitimate reason to ask questions. Those questions should not be classified alongside the nonsense theories listed above. We will not be harmed politically if we demand reasonable answers. We will be harmed if we let fear or embarrassment stop us from speaking up.

2. The exit polls. How many times am I going to have to say this? The question of whether exit polls are accurate should not concern us nearly so much as the pattern of inaccuracy. Even if those polls are flagrantly wrong (and they shouldn't be; see our final offering in this column), logic tells us that they should score as many "false positives" for the Democrats as for the Republicans. Yet they consistently undervalue Republican votes, in state after state, election after election. That's suspicious, dammit!

Nobody has offered any persuasive methodological explanation as to why this situation should occur. Nobody has explained why exit polls work better in other countries than they do in ours. Nobody has explained why exits did not favor one party until more recent elections. Nobody has explained why we should ignore the pattern of disparity in this country, even though international observers in other countries consider such a pattern an indicator of vote fraud.

Slate debunks. Josh Levin treads much the same ground as the Daily Kos. Alas, he keeps his eye on Florida's smaller counties, a diversionary issue which we have dealt with already. The larger problem remains unaddressed.

Salon debunks. Much the same can be said of the new piece by Farhad Manjoo (who wrote quite sensibly about bulge-gate). Manjoo makes some good points about Kathy Dopp's analysis, and he may well be correct when he discusses the provisional ballots in Ohio. He gives us an explanation of Cuyahoga county that makes a bit more sense than the one given by the afore-cited Kos-ian:

According to Kimberly Bartlett, a spokeswoman for the county, the reporting software the county uses to display the unofficial summary of election results on its Web site is simply buggy. For some reason, the software combines absentee ballots from several voting precincts into one precinct, and therefore makes it appear as if there were more votes cast in a particular area than there were registered voters there. But this bug does not affect the final election results, because the more detailed "canvass" of all the votes cast in the county shows the correct count, Bartlett told Salon.

But these matters are diversions. Of the exit poll disparities, Manjoo has as little clue as anyone else. And he repeats a point others have made, without taking it far enough:

Another frequent complaint: Very often, voters would attempt to select one candidate on a voting machine and for some mysterious, as yet undetermined reason, the candidate's opponent will have been selected.

Drop the other shoe, my friend. If this phenomenon is accidental, then the errors should run in both directions. All of the anecdotal reports I have seen (and I've been looking) tell the same story -- an intended vote for Kerry would pop up as a vote for Bush. This article mentions one thousand such complaints across the country. Are there any reports of the contrary scenario?

If not -- at what point do we have enough anecdotal evidence to start worrying? With a thousand complaints, aren't we somewhat past the "anecdotal" stage?

Olbermann. You'll want to read his latest, which talks about the messy results in North Carolina. Alas, he links to a Charlotte Observer story that I can't reach. Olbermann says that North Carolina may have to stage a revote. (For more, see here.)

The Boston Globe tells us that "none of the vote-counting problems and anomalies that have emerged are sufficiently widespread to have affected the election's ultimate result." That's like saying no single car can clog up a freeway. What counts is the cumulative impact.

The Globe refuses to address the question of why the "accidents" all favor Bush. I've said it a hundred times, and I won't mind saying it another thousand times: What matters is not the inaccuracy, it's the pattern of inaccuracy.

Backdoor. (I wonder what sort of remark dear Ana Marie would insert here...?) This older article from New Zealand linked to the files providing the backdoor to the GEMS system. Maybe you'd like to rig the next election...?

The Russian view. Pravda accuses us of being insufficiently democratic. Cute, huh?

A GREAT article. This terrific piece by Paul Rosenberg makes some points I've been trying to hammer home:

Exit polls on Election Day gave a clear indication of a Kerry victory. While sampling error could have explained them individually, something more was going on, as veteran pollster Mark Blumenthal explained, "Kerry's performance on the partial exit polls surpassed his ultimate performance nationally and in 15 of 16 states. So whatever was happening, it was not just the random variation due to sampling error. If you don't believe me, try flipping a coin and see how often you can get heads to come up 16 of 17 times."

Most dramatically, in New Hampshire, Kerry's 17-point margin in the exit poll shrank to a bare one percent in the final tally.

The standard response has been to say that Kerry supporters must have been more eager to talk to pollsters -- but such a phenomena in exit polls has never been seen before. Vote tampering is another explanation, but mainstream analysts dismiss it out of hand, despite repeated warnings from computer specialists about inadequate security measures.

Rosenberg also relays details previously unfamiliar to me:

Voter intimidation was another widespread problem. A disturbing number of violations involved campaign posters or literature in church buildings—evidence, in some cases, of IRS as well as election law violations. A report from the Summit Church of God in Weathersfield read, "Sign at polling place saying "VOTE BUSH" - when voter said sign should come down the minister said 'If you don't like it, vote somewhere else.'"

And:

People in the low-income black neighborhood around Hillman Elementary School had their water turned off that morning if their bill was unpaid, and were told by the water department to stay home until the matter was resolved to let someone in -- which reportedly never happened.

Homeland Security. Heretofore, I've tended to ignore the story of Warren County, Ohio, where the officials locked down the administration building on election night, citing potential terrorism as the reason for removing observers. But this story is odd. An FBI agent supposedly told officials that the place rated a "10" on the terrorism scale.

Apparently, Osama has a bug up his ass about Warren County. And here, I thought his next target would be the Sears Tower...!

An end to exit polls. Ed Gillespie has called for an end to exit polls, which will soon be the sole remaining means of double-checking the accuracy of the electronic vote. (Odd how the right liked exit polls in Venezuela...) This paragraph is rich:

He also said he was personally affected by the early reports, discouraged by what he was seeing. "But I've been through this before," he said. "In 2000 the exit data was wrong on Election Day. In 2002, the exit returns were wrong on Election Day. And in 2004, the exit data were wrong on Election Day -- all three times, by the way, in a way that skewed against Republicans and had a dispiriting effect on Republican voters across the country."

Why is it that the RNC head can see the pattern which so many of our Democratic brethren refuse to acknowledge? Of course, that pattern actually indicates a fraudulent final tally, not fraudulent exit polling. Since the corporate media heads want to stay in the G.O.P.'s good graces, expect Ed to get his evil way on this.

Exit polls vs. predictive polls. Buzzflash quotes an expert whose words deserve widespread repetition:

There's a huge difference between polling what WILL happen and polling something that has already happened. The reliability of polling something that has already happened is highly reliable vs. predictive polls, like Gallup or Zogby, which is very risky. The reliability can be, not plus or minus 4 percent as we see with predictive polls, but rather a much more reliable plus or minus one half or one tenth of one percent with exit polls, because those are based on asking people who already voted. I would even say that if the exit polling were done in the key precincts of Florida and Ohio, which it was, then these results should be practically "bullet proof.

Finally: I may have an interview coming up which could shed further light on this...