Before the vote-counting was done, the e-mails started arriving. The election’s been stolen! Fraud! John Kerry won! In the following days, these charges flew over the Internet. The basic claim was that the early exit polls–which showed Kerry ahead of George W. Bush–were right; the vote tallies were rigged. Could this be? Or have ballot booths with electronic voting machines become the new Grassy Knoll for conspiracy theorists?

Anyone who questioned the integrity of the nation’s voting system–before the election or after–has had good reason to do so. Electronic voting that does not produce an auditable paper trail is worrisome–as is the possibility that the machines can be hacked. The proponents of these systems claim there are sufficient safeguards. But in this election there were numerous reports of e-voting gone bad. Votes cast for one candidate were registered for another. In Broward County, Florida, software subtracted votes rather than added them. In Franklin County, Ohio, an older electronic machine reported an extra 3,893 votes for Bush. Local election officials caught that error. But when I asked Peggy Howell, one of those officials, why the mistake occurred, she replied, “We really don’t know.” Were these errors statistically insignificant glitches that inevitably happen in any large system? “It gives us the uneasy feeling that we’re only seeing the tip of the iceberg,” Cindy Cohn of the Electronic Frontier Foundation, which is part of the Election Protection Coalition, told Reuters. “What has most concerned scientists are problems that are not observable,” David Jefferson, a computer scientist at Lawrence Livermore National Laboratory in California, explained to the Associated Press. “The fact that we had a relatively smooth election…does not change at all the vulnerability these systems have to fraud or bugs.” And the 2000 fiasco in Florida demonstrated that non-electronic voting can also have serious problems, which often disproportionately affect low-income counties.

Then there’s the issue of who is running the show. Only a few companies manufacture electronic voting machines. They are not transparent. They do not use open-source code. Last year, Walden O’Dell, the head of Diebold, a leading manufacturer of touch-screen machines, declared in a fundraising letter for the Ohio Republican Party that he was “committed to helping Ohio deliver its electoral votes to the president next year.” That hardly inspired confidence. And across the country, oversight of voting is conducted by partisan officials. In Ohio, Secretary of State Kenneth Blackwell, a Republican and conservative activist, oversaw the voting. On his watch, the polling place for Kenyon College was equipped with only two voting machines. Yet about 1,100 people–mostly students–wanted to vote there. These voters (and you can guess whom they preferred) had to wait up to nine hours. It doesn’t require much cynicism to suspect that this was no accident.

But did something more foul than minor slip-ups and routine political chicanery occur? Those who say yes–at this point–are relying more on supposition than evidence. They cite the exit polls to claim the vote count was falsified to benefit Bush. The pollsters say they oversampled women, that their survey takers were not allowed to get close enough to the polls and that Kerry supporters may have been more willing to cooperate with the pollsters than Bush backers. Impossible, huffs pollster/consultant Dick Morris: “Exit polls are almost never wrong.” But Morris argues that the faulty exit polls are not a sign the vote count was off but an indication that the pollsters deliberately produced pro-Kerry results “to try to chill the Bush turnout.” (Talk about conspiracy theory.) The screwy exit polls do raise questions, but they are not proof of sabotage. And left-of-center accusers have promoted contradictory theories. Many suggest Diebold and other vendors put in the fix via the paperless touch-screen machines. But other critics–including progressive talk-show host and author Thom Hartmann–also point to a spreadsheet created by an activist named Kathy Dopp that shows what she considers anomalous pro-Bush results in Florida counties that used optical-scan voting, not electronic touch-screen voting. (The optical-scan machines were manufactured by Diebold and the other firms that produce the touch-screen machines.) But Walter Mebane, a Cornell professor, and colleagues at Harvard and Stanford examined this allegation of fraud and concluded that it is “baseless.” They note that the counties in question are mostly in the conservative Florida Panhandle and “have trended strongly Republican over the past twelve years.”

Making a different we-wuz-robbed claim, journalist Greg Palast, in an article bluntly titled “Kerry Won…,” contends the Democrat would have definitely triumphed in Ohio had the final tally included the uncounted ballots–by which he means 92,672 ballots that did not register a vote when run through a counting machine–and the 155,000 provisional ballots. Palast wrongly assumes that an overwhelming majority of these ballots contain votes for Kerry, who lost by 136,000 votes. Not all of the provisional ballots, however, would pass legal muster. (Ohio Democrats estimated less than 90 percent would be valid.) And more important, the 92,672 other ballots, if hand-counted, probably would not have produced a major vote gain for Kerry. After the Florida 2000 mess, I examined almost a third of the 10,500 uncounted votes in Miami-Dade County. Of those, only a few hundred contained a discernible vote. Tallying them produced merely a five-vote edge for Al Gore. It is highly improbable that the pool of uncounted and provisional ballots in Ohio could have yielded Kerry a net gain of more than 136,000 votes.

Clear away the rhetoric, and what’s mainly left are the odd early exit polls (which did show Kerry’s lead in Ohio and Florida declining as Election Day went on and which ended up with the current national Bush-Kerry spread), troubling instances of bad electronic voting, and curious–or possibly curious–trends in Florida. This may be the beginning of a case; it is not a case in itself. Investigative reporter Robert Parry observes, “Theoretically, at least, it is conceivable that sophisticated CIA-style computer hacking–known as ‘cyber-warfare’–could have let George W. Bush’s campaign transform a three-percentage-point defeat, as measured by exit polls, into an official victory of about the same margin. Whether such a scheme is feasible, however, is another matter, since it would require penetration of hundreds of local computer systems across the country, presumably from a single remote location. The known CIA successes in cyber-war have come from targeting a specific bank account or from shutting down an adversary’s computer system, not from altering data simultaneously in a large number of computers.”

The skeptics–correct or not in their claims of fraud–are right to be concerned in general about the vote-counting system. Representatives John Conyers, Jerrold Nadler and Robert Wexler have asked the Government Accountability Office (formerly the General Accounting Office) to investigate the “voting machines and new technologies used in the 2004 election.” Blackboxvoting.org–a group that has long decried electronic voting and now claims that “fraud took place in the 2004 election”–has filed Freedom of Information Act requests to obtain internal computer logs and other documents from 3,000 counties and localities, in an attempt to audit the election. The public does deserve any information that would allow it to evaluate vote-counting. Beyond that, extensive election reform is necessary. Electronic voting ought to produce a paper trail that can be examined. There should be national standards for voting systems and for verifying vote tallies. And vote counters should be nonpartisan public servants, not secretive corporations or party hacks. The system ought to be so solid that no one would have cause even to wonder whether an election has been stolen.