Election 2004: Stolen or Lost
Stay up to date with the latest headlines via email.
Many of us fear that the Ohio election was stolen because people – like talk-show sleuths, blogger number-crunchers, forensic attorneys, crusading professors and partisan activists – keep telling us so. We don't even know most of these people, yet we gladly forward their e-mails and web links, their pronouncements, analyses, essays and statistical exercises. While their credentials may not be that impressive, we listen to their conspiracy theories because – frightened by the direction our country has taken – we want to believe them.
As an old-style investigative reporter, I, too, was alarmed by charges that outright fraud might have changed the outcome of the most important presidential election in recent times. So I recently traveled to Ohio – where I connected with a group of attorneys who were fighting to have the Ohio presidential results overturned, and the state – and, by extension, the presidency – awarded to Kerry. In legal pleadings known collectively as the "Contest" these attorneys are not shy about using the F-word: "While a variety of methods were used to perpetrate the election fraud of which there is clear and convincing evidence in the form of the exit polls, ... it is likely that traditional easily detectable means were one of the principal methods of the election fraud."
Strong words indeed. Among the evidence supporting them:
- Specific instances in which strange or troubling things happened when people voted or while votes were being counted.
- The discrepancy between exit polls and the final result.
This week, Rep. John Conyers, D-Mich., released a report that catalogues widespread problems in the Ohio vote. The report concludes that the "massive and unprecedented" voting irregularities in Ohio were in many cases caused by "intentional misconduct and illegal behavior." Sounds like fraud to me.
Conyers' report is considerably tamer and more cautious than earlier pronouncements out of his office, and certainly more so than many of the allegations being circulated on the internet. Much of his report, however, is based on charges emerging from the Contest. Let's see how such charges hold up under close scrutiny.
Charge: Misallocation of voting machines
Intentional? Probably not
The Contest petition lists specific counties where voting irregularities occurred, including Franklin and Trumbull: "In Franklin County there was a discriminatory assignment of more voting machines per registered voter to precincts with more white voters than African-American voters. ..."
William Anthony is the chairman of the Franklin County board of elections. As an African American and a Democrat himself (in fact, he is the county chairman and works as a union representative) Anthony resents the suggestion that Franklin County authorities somehow worked to help Bush. "I worked my ass off in those precincts," he says of African-American areas of the county.
A precinct-by-precinct historical comparison of registered and actual voters, and of voting machine assignments, does show that some precincts with a large African-American population ended up with fewer machines per person than some mostly white precincts. But Anthony points out that Franklin County faced a number of challenges. For one thing, it was using very old electronic voting machines that under new state law will be defunct by the next presidential election, when every county will be required to have a paper trail for recounts. Given the short lifespan of the machines, it didn't make economic sense to buy more of them. So it was a matter of allocating a scarce resource. That resource was stretched thinner by an increasing population. Franklin County had a spurt of growth in outlying areas, with blocks of apartments sprouting recently where cornfields had been. Suddenly, authorities had 29 additional precincts to conside – requiring approximately 200 more machines.
Also, although incoming voter registration figures showed surges in certain areas, that didn't mean the newly registered would necessarily vote. And certainly not in greater numbers than in many established precincts where a high percentage of registered voters typically went to the polls.
When the county elections director recently explained the machine assignment process as "a little bit art, a little bit science," he was ridiculed by the critics. But in fact, what he meant was that a whole multiplicity of factors had to be considered – it wasn't a simple formula.
Significantly, the people making these decisions aren't necessarily Bush partisans. Every county in Ohio, by law, divides its elections personnel evenly between the Democrats and Republicans. This means that where the chief administrator of elections is, say, a Republican, the chairperson of the elections board is a Democrat. In the case of Franklin County, two individuals shared the task of allocating machines – and one was a Democrat.
Charge: Miscounting of absentee votes
A Contest attorney who asked that his name not be used told me that he considered irregularities in Trumbull County perhaps the most damning of all. Here are the specifics: Dr. Werner Lange, a Trumbull County resident, examined poll books in county offices, looked at 106 precincts and calculated that, in all, "580 absentee votes were cast for which there was no notation of absentee voting in the poll books." Extrapolated statewide, this pattern – if it existed – could translate into 62,513 fraudulent votes, or, more than half of Bush's advantage.
Lange, who – according to his affidavit – holds a Ph.D. in political science and is an ordained Minister of Word and Sacrament, told me that he had suspicions because the area was heavily Democratic, but that Bush had done surprisingly well.
Then I checked in at the Trumbull County offices. "Mr. Lange came in here looking for problems and he didn't want to ask us anything," says Rokey Suleman, the deputy director of the Trumbull County Board of Elections. Suleman explains that the poll books Lange looked at had been printed before absentee voting ended – including those who voted in the final days before the election at the Board's offices. The books would – according to practice – be updated to include everyone. Like Anthony in Franklin County, Suleman is a Democrat.
Charge: Tampering with voting machines
Finding: Probably false
There were a number of anecdotal claims that personnel from voting machine companies came into several counties and seemed to do something improper with the machines before the recount began. I had the opportunity to listen to an audio tape of a film crew interviewing an official of Triad, a ballot counting contractor accused on the internet of various indiscretions – in which the man appears to be very patiently and logically explaining the exact role of company personnel in preparing machines for recounts. I asked Contest attorneys if they wanted to listen to the tape, but they were too busy rushing out filings – which included allegations involving Triad.
In a couple of precincts in Cuyahoga County (Cleveland), third party candidates did inexplicably well. In one precinct located in a predominantly African-American area, Kerry got 290 votes, Bush 21 and Michael Peroutka – candidate of the anti-immigrant Constitutional Party – got 215. In another precinct that voted at the same high school, the tally was Kerry 318, Bush 21, and Libertarian Michael Badnarik 163. These are, for the moment, mysteries, but they are not indications of widespread fraud.
"I think incompetence is the most likely explanation in most of these cases," says Mark Griffin, who ran the legal team for Kerry that oversaw the provisional ballot count and recount in Cuyahoga. "If there's fraud, it's in the tabulation. But it wouldn't be in Cuyahoga, where we got a big turnout."
Challengers May Have Good Intentions But Bad Facts
The lawyers on the Contest team are well-meaning, intelligent people. However, like all lawyers, they're about arguing their side, not getting to the bottom of things. Each time I checked in, one of them (always the same person, always insisting that our conversations were "off the record") would introduce a new operative theory of what had happened, of what "evidence" was the most meaningful indicator of the shenanigans that had gone on.
By the day I left, my "source" was telling me that the legal discovery process was not going well, and so they could not, in a timely fashion, get the information needed to show how the "fraud" was perpetrated.
Attorney Don McTigue, a senior official in the Ohio Secretary of State's office when a Democrat held the office, now devotes his entire legal practice to elections. "We don't have evidence of fraud," says McTigue, who represented the Kerry campaign for the recount, an entirely separate proceeding from the Contest, which resulted in almost no change in the vote totals, and left Bush with a hefty 118,000-vote margin.
You wouldn't have much of a case for conspiracy if you didn't have Secretary of State J. Kenneth Blackwell. As Ohio's chief elections official and the Bush state campaign chairman, he was already juggling a couple of hats that should have been nowhere near each other. Furthermore, Bush traveled to Ohio on election day to meet with Blackwell. There's no doubt that Blackwell consistently ruled in a manner that seemed to favor Bush. But the impact of these rulings on the election was probably minimal. For example, Blackwell ruled that Ohio voters could not cast provisional ballots outside their designated precinct. Local officials had hoped to alleviate confusion over polling places by letting people come to "zones" – regional locations where they could cast a ballot when their assigned precinct was in doubt. When Blackwell blocked this option, a number of counties, including Franklin, went ahead and accepted provisional ballots at polling places from people who signed a statement explaining why they were in the wrong place, why they didn't have time to go to their designated precinct, etc.
Blackwell's attempt to disqualify voter registration cards that didn't meet an 80-lb paper test also failed, as did his push to get counties to switch from punch cards to "black box" balloting (using Direct Recording Electronic machines, or DREs) Paradoxically, if he had prevailed on DREs, the election would likely have been closer, because punchcards generate more spoilage. Electronic voting, i.e. "black box" balloting where there is no paper trail, existed only in a handful of counties. Franklin is the largest of these – and Kerry did better than expected there, even with the long lines.
Charge: Voting company fraud
As for Diebold and other vilified companies, in all probability, they didn't, and wouldn't, risk the ignominy and consequences of fixing an election. The primary reason so many people are suspicious of Diebold in the first place is because of the CEO's ill-advised promise, in a GOP fundraising letter, to do everything he could to see Ohio's electors awarded to Bush. That was an outrageous thing to say, but even on its face more likely a sign of cluelessness than of hidden plans to alter the outcome.
Charge: Exit poll results were more accurate than actual ballots
Explanation of Problem: Imperfect nature of polls
Now to the central issue: the claim that exit polls, which never lie, showed Kerry winning. Our understanding of this – and the argumentation in the Contest – is based largely on an analysis by Steven F. Freeman, Ph.D. But Freeman is not an expert in polling. According to his affidavit, he is a visiting scholar in the Graduate Division, School of Arts and Sciences, University of Pennsylvania Center for Organizational Dynamics.
To get some insight into this issue, I spoke with a source who, in the common parlance, is "familiar with the thinking of" Warren Mitofsky, the "father" of the exit poll.
Asked about Freeman's analysis, my source told me that it is "all wrong." We spent several hours going through Freeman's specific claims, and reviewed how exit polls – and Mitofsky's in particular – work.
Much of the belief that the election was stolen was based on "screen shots" of raw numbers provided by CNN. In exit polling, raw numbers mean almost nothing – since the essence of a successful exit poll is to interview a sampling of voters, and then apply a variety of methods in order to adjust to the most probable accurate assessment. "To say you want the raw data is ludicrous," said the source. "You can't use it until you do something with it. You're talking about a bunch of naïve people that had [only] the first course in statistics."
Bill Leonard, a former CBS News VP who was a polling pioneer, has called exit polls "blunt instruments." The widely circulated notion that they are always right is dead wrong.
The notion that a single definitive number showing Kerry winning ever existed is also wrong. "We never had unadjusted unofficial totals," said the source. "As we get more data, we're always adjusting."
In this case, what most likely happened was that more Bush supporters failed to complete exit poll surveys than Kerry backers. The reason for that can be as trivial as a sampler skipping someone who looks unfriendly or voters not liking the race or demeanor of the sampler.
(For what it is worth, I learned that Mitofsky is a lifelong liberal and apparently holds no brief for Bush. But a job's a job, and a professional is a professional.)
While it's appallingly easy to mess around with a computer, it's a lot more difficult to rig an election.
To have a conspiracy of this magnitude, you'd need more than a bunch of individual mishaps – you need a plan and coordination. And you'd need a large number of collaborators willing to commit felonious and treasonous behavior of the highest order.
None of this is to say that election fraud could not, theoretically, happen – particularly in a truly opaque system that produces no paper trail. Indeed, with 88 counties, a bewildering variety of voting systems, and often conflicting decisions by courts, and state and local officials, it's a wonder that elections work at all.
More allegations of fraud in the 2004 election will be floated. The problem is that those who float these claims don't bother to perform any measure of due diligence. So far, every claim of fraud that I have examined has turned out to have a credible alternate explanation.
Legitimate avenues of inquiry remain. Exploring how particular companies get contracts is one; the appropriate role and behavior of top state elections officials is another. Then there are the kinds of bad administrative decisions made at the county level.
Certainly, there were many instances of small-scale cheating and intimidation – as there probably always are. Limited numbers of voters received phone calls and letters on bogus official letterheads, telling them they could not vote; in one egregious instance, elderly Democrats were "informed" that it might be more comfortable for them to vote on Wednesday, when lines were shorter. But there's no evidence or even likelihood that this was authorized from on high; it's far more like that such amateurish interventions were locally conceived.
Technical and administrative failings were certainly apparent. The inadequate distribution of machines was just one. Another was the fact that poll workers couldn't get through continuously busy phone lines to county officials.
"Overall, Ohio has a good system," said Democratic election lawyer McTigue. "Like any system, if you scrutinize it enough, you're going to find weaknesses."
One conclusion seems obvious: Because of the growing partisan animus and attendant suspicions, everyone connected with the electoral process is going to have to be a whole lot more careful and a whole lot more forthcoming. That goes, especially, for those responsible for creating the system, installing and maintaining it, making the decisions. "Times have changed," says Chris Wilson, Franklin County's election technology administrator, and a "Republican" hire who is well-regarded and insists on his independence. "You can't go in and do maintenance without everybody knowing what's going on. You can't talk gobbledygook anymore."
Sadly, it appears that much of the blame for the Bush victory rests with those who wished it were otherwise. In Cuyahoga County, for example, Republicans worked very aggressively to get out their voters, while Democrats often did not. Mark Griffin, the former Kerry campaign attorney, recalls being told by colleagues that, according to the numbers, a county with 1.4 million people had only a couple of dozen swing voters.
"I don't think the evidence shows a conspiracy," says Griffin. "It doesn't show the Republicans stole the election. It shows we are continuing to have mechanical and systemic failures that, in a closer election, could have flipped the results."
Half-baked conspiracy theories are damaging to the public confidence in democracy. We could use a few less conspiracy theorists, and a few more Griffins. It takes a pretty big person to admit that one's own side screwed up, or was simply bested in a fight (even a nasty one), or to accept, and tackle, the growing alienation of potential voters in America. And the unexciting, labor-intensive process of analyzing and fixing the machinery of the people's will.