Ohio's Odd Numbers
Are the stories of vote suppression and rigged machines to be believed? Here is "non-wacko" evidence that something went seriously awry in the Buckeye State on Election Day 2004
By CHRISTOPHER HITCHENS

If it were not for Kenyon College, I might have missed, or skipped, the whole controversy. The place is a visiting lecturer's dream, or the ideal of a campus-movie director in search of a setting. It is situated in wooded Ohio hills, in the small town of Gambier, about an hour's drive from Columbus. Its literary magazine, The Kenyon Review, was founded by John Crowe Ransom in 1939. Its alumni include Paul Newman, E. L. Doctorow, Jonathan Winters, Robert Lowell, Chief Justice William Rehnquist, and President Rutherford B. Hayes. The college's origins are Episcopalian, its students well mannered and well off and predominantly white, but it is by no means Bush-Cheney territory. Arriving to speak there a few days after the presidential election, I found that the place was still buzzing. Here's what happened in Gambier, Ohio, on decision day 2004.

The polls opened at 6:30 a.m. There were only two voting machines (push-button direct-recording electronic systems) for the entire town of 2,200 (with students). The mayor, Kirk Emmert, had called the Board of Elections 10 days earlier, saying that the number of registered voters would require more than that. (He knew, as did many others, that hundreds of students had asked to register in Ohio because it was a critical "swing" state.) The mayor's request was denied. Indeed, instead of there being extra capacity on Election Day, one of the only two machines chose to break down before lunchtime.

By the time the polls officially closed, at 7:30 that evening, the line of those waiting to vote was still way outside the Community Center and well into the parking lot. A federal judge thereupon ordered Knox County, in which Gambier is located, to comply with Ohio law, which grants the right to vote to those who have shown up in time. "Authority to Vote" cards were kindly distributed to those on line (voting is a right, not a privilege), but those on line needed more than that. By the time the 1,175 voters in the precinct had all cast their ballots, it was almost four in the morning, and many had had to wait for up to 11 hours. In the spirit of democratic carnival, pizzas and canned drinks and guitarists were on hand to improve the shining moment. TV crews showed up, and the young Americans all acted as if they had been cast by Frank Capra: cheerful and good-humored, letting older voters get to the front, catching up on laptop essays, many voting for the first time and all convinced that a long and cold wait was a small price to pay. Typical was Pippa White, who said that "even after eight hours and 15 minutes I still had energy. It lets you know how worth it this is." Heartwarming, until you think about it.

The students of Kenyon had one advantage, and they made one mistake. Their advantage was that their president, S. Georgia Nugent, told them that they could be excused from class for voting. Their mistake was to reject the paper ballots that were offered to them late in the evening, after attorneys from the Ohio Democratic Party had filed suit to speed up the voting process in this way. The ballots were being handed out (later to be counted by machine under the supervision of Knox County's Democratic and Republican chairs) when someone yelled through the window of the Community Center, "Don't use the paper ballots! The Republicans are going to appeal it and it won't count!" After that, the majority chose to stick with the machines.

Across the rest of Ohio, the Capra theme was not so noticeable. Reporters and eyewitnesses told of voters who had given up after humiliating or frustrating waits, and who often cited the unwillingness of their employers to accept voting as an excuse for lateness or absence. In some way or another, these bottlenecks had a tendency to occur in working-class and, shall we just say, nonwhite precincts. So did many disputes about "provisional" ballots, the sort that are handed out when a voter can prove his or her identity but not his or her registration at that polling place. These glitches might all be attributable to inefficiency or incompetence (though Gambier had higher turnouts and much shorter lines in 1992 and 1996). Inefficiency and incompetence could also explain the other oddities of the Ohio process?from machines that redirected votes from one column to the other to machines that recorded amazing tallies for unknown fringe candidates, to machines that apparently showed that voters who waited for a long time still somehow failed to register a vote at the top of the ticket for any candidate for the presidency of these United States.

However, for any of that last category of anomaly to be explained, one would need either a voter-verified paper trail of ballots that could be tested against the performance of the machines or a court order that would allow inspection of the machines themselves. The first of these does not exist, and the second has not yet been granted.

I don't know who it was who shouted idiotically to voters not to trust the paper ballots in Gambier, but I do know a lot of people who are convinced that there was dirty work at the crossroads in the Ohio vote. Some of these people are known to me as nutbags and paranoids of the first water, people whose grassy-knoll minds can simply cancel or deny any objective reasons for a high Republican turnout. (Here's how I know some of these people: In November 1999, I wrote a column calling for international observers to monitor the then upcoming presidential election. I was concerned about restrictive ballot-access laws, illegal slush funds, denial of access to media for independents, and abuse of the state laws that banned "felons" from voting. At the end, I managed to mention the official disenfranchisement of voters in my hometown of Washington, D.C., and the questionable "reliability or integrity" of the new voting-machine technology. I've had all these wacko friends ever since.) But here are some of the non-wacko reasons to revisit the Ohio election.

First, the county-by-county and precinct-by-precinct discrepancies. In Butler County, for example, a Democrat running for the State Supreme Court chief justice received 61,559 votes. The Kerry-Edwards ticket drew about 5,000 fewer votes, at 56,243. This contrasts rather markedly with the behavior of the Republican electorate in that county, who cast about 40,000 fewer votes for their judicial nominee than they did for Bush and Cheney. (The latter pattern, with vote totals tapering down from the top of the ticket, is by far the more general?and probable?one nationwide and statewide.)

In 11 other counties, the same Democratic judicial nominee, C. Ellen Connally, managed to outpoll the Democratic presidential and vice-presidential nominees by hundreds and sometimes thousands of votes. So maybe we have a barn-burning, charismatic future candidate on our hands, and Ms. Connally is a force to be reckoned with on a national scale. Or is it perhaps a trick of the Ohio atmosphere? There do seem to be a lot of eccentrics in the state. In Cuyahoga County, which includes the city of Cleveland, two largely black precincts on the East Side voted like this. In Precinct 4F: Kerry, 290; Bush, 21; Peroutka, 215. In Precinct 4N: Kerry, 318; Bush, 11; Badnarik, 163. Mr. Peroutka and Mr. Badnarik are, respectively, the presidential candidates of the Constitution and Libertarian Parties. In addition to this eminence, they also possess distinctive (but not particularly African-American-sounding) names. In 2000, Ralph Nader's best year, the total vote received in Precinct 4F by all third-party candidates combined was eight.

In Montgomery County, two precincts recorded a combined undervote of almost 6,000. This is to say that that many people waited to vote but, when their turn came, had no opinion on who should be the president, voting only for lesser offices. In these two precincts alone, that number represents an undervote of 25 percent, in a county where undervoting averages out at just 2 percent. Democratic precincts had 75 percent more undervotes than Republican ones.

In Precinct 1B of Gahanna, in Franklin County, a computerized voting machine recorded a total of 4,258 votes for Bush and 260 votes for Kerry. In that precinct, however, there are only 800 registered voters, of whom 638 showed up. Once the "glitch" had been identified, the president had to be content with 3,893 fewer votes than the computer had awarded him.

In Miami County, a Saddam Hussein?type turnout was recorded in the Concord Southwest and Concord South precincts, which boasted 98.5 percent and 94.27 percent turnouts, respectively, both of them registering overwhelming majorities for Bush. Miami County also managed to report 19,000 additional votes for Bush after 100 percent of the precincts had reported on Election Day.

In Mahoning County, Washington Post reporters found that many people had been victims of "vote hopping," which is to say that voting machines highlighted a choice of one candidate after the voter had recorded a preference for another. Some specialists in election software diagnose this as a "calibration issue."

Machines are fallible and so are humans, and shit happens, to be sure, and no doubt many Ohio voters were able to record their choices promptly and without grotesque anomalies. But what strikes my eye is this: in practically every case where lines were too long or machines too few the foul-up was in a Democratic county or precinct, and in practically every case where machines produced impossible or improbable outcomes it was the challenger who suffered and the actual or potential Democratic voters who were shortchanged, discouraged, or held up to ridicule as chronic undervoters or as sudden converts to fringe-party losers.

This might argue in itself against any conspiracy or organized rigging, since surely anyone clever enough to pre-fix a vote would make sure, just for the look of the thing, that the discrepancies and obstructions were more evenly distributed. I called all my smartest conservative friends to ask them about this. Back came their answer: Look at what happened in Warren County.

On Election Night, citing unspecified concerns about terrorism and homeland security, officials "locked down" the Warren County administration building and prevented any reporters from monitoring the vote count. It was announced, using who knows what "scale," that on a scale of 1 to 10 the terrorist threat was a 10. It was also claimed that the information came from an F.B.I. agent, even though the F.B.I. denies that.

Warren County is certainly a part of Republican territory in Ohio: it went only 28 percent for Gore last time and 28 percent for Kerry this time. On the face of it, therefore, not a county where the G.O.P. would have felt the need to engage in any voter "suppression." A point for the anti-conspiracy side, then. Yet even those exact-same voting totals have their odd aspect. In 2000, Gore stopped running television commercials in Ohio some weeks before the election. He also faced a Nader challenge. Kerry put huge resources into Ohio, did not face any Nader competition, and yet got exactly the same proportion of the Warren County votes.

Whichever way you shake it, or hold it to the light, there is something about the Ohio election that refuses to add up. The sheer number of irregularities compelled a formal recount, which was completed in late December and which came out much the same as the original one, with 176 fewer votes for George Bush. But this was a meaningless exercise in reassurance, since there is simply no means of checking, for example, how many "vote hops" the computerized machines might have performed unnoticed.

There are some other, more random factors to be noted. The Ohio secretary of state, Kenneth Blackwell, was a state co-chair of the Bush-Cheney campaign at the same time as he was discharging his responsibilities for an aboveboard election in his home state. Diebold, which manufactures paper-free, touch-screen voting machines, likewise has its corporate headquarters in Ohio. Its chairman, president, and C.E.O., Walden O'Dell, is a prominent Bush supporter and fund-raiser who proclaimed in 2003 that he was "committed to helping Ohio deliver its electoral votes to the president next year." (See "Hack the Vote," by Michael Shnayerson, Vanity Fair, April 2004.) Diebold, together with its competitor, E.S.&S., counts more than half the votes cast in the United States. This not very acute competition is perhaps made still less acute by the fact that a vice president of E.S.&S. and a Diebold director of strategic services are brothers.

I would myself tend to discount most of the above, since an oligarchy bent on stealing an election would probably not announce itself so brashly as to fit into a Michael Moore script. Then, all state secretaries of state are partisan, after all, while in Ohio each of the 88 county election boards contains two Democrats and two Republicans. The chairman of Diebold is entitled to his political opinion just as much as any other citizen.

However, there is one soothing explanation that I don't trust anymore. It was often said, in reply to charges of vote tampering, that it would have had to be "a conspiracy so immense" as to involve a dangerously large number of people. Indeed, some Ohio Democrats themselves laughed off some of the charges, saying that they too would have had to have been part of the plan. The stakes here are very high: one defector or turncoat with hard evidence could send the principals to jail forever and permanently discredit the party that had engaged in fraud.

I had the chance to spend quality time with someone who came to me well recommended, who did not believe that fraud had yet actually been demonstrated, whose background was in the manufacture of the machines, and who wanted to be anonymous. It certainly could be done, she said, and only a very, very few people would have to be "in on it." This is because of the small number of firms engaged in the manufacturing and the even smaller number of people, subject as they are to the hiring practices of these firms, who understand the technology. "Machines were put in place with no sampling to make sure they were 'in control' and no comparison studies," she explained. "The code of the machines is not public knowledge, and none of these machines has since been impounded." In these circumstances, she continued, it's possible to manipulate both the count and the proportions of votes.

In the bad old days of Tammany Hall, she pointed out, you had to break the counter pins on the lever machines, and if there was any vigilance in an investigation, the broken pins would automatically incriminate the machine. With touch-screen technology, the crudeness and predictability of the old ward-heeler racketeers isn't the question anymore. But had there been a biased "setting" on the new machines it could be uncovered?if a few of them could be impounded. The Ohio courts are currently refusing all motions to put the state's voting machines, punch-card or touch-screen, in the public domain. It's not clear to me, or to anyone else, who is tending the machines in the meanwhile ?

I asked her, finally, what would be the logical grounds for deducing that any tampering had in fact occurred. "Well, I understand from what I have read," she said, "that the early exit polls on the day were believed by both parties." That, I was able to tell her from direct experience, was indeed true. But it wasn't quite enough, either. So I asked, "What if all the anomalies and malfunctions, to give them a neutral name, were distributed along one axis of consistency: in other words, that they kept on disadvantaging only one candidate?" My question was hypothetical, as she had made no particular study of Ohio, but she replied at once: "Then that would be quite serious."

I am not any sort of statistician or technologist, and (like many Democrats in private) I did not think that John Kerry should have been president of any country at any time. But I have been reviewing books on history and politics all my life, making notes in the margin when I come across a wrong date, or any other factual blunder, or a missing point in the evidence. No book is ever free from this. But if all the mistakes and omissions occur in such a way as to be consistent, to support or attack only one position, then you give the author a lousy review. The Federal Election Commission, which has been a risible body for far too long, ought to make Ohio its business. The Diebold company, which also manufactures A.T.M.s, should not receive another dime until it can produce a voting system that is similarly reliable. And Americans should cease to be treated like serfs or extras when they present themselves to exercise their franchise.