A "Glitch" in Democracy

Coverage of computer voting problems too little, too late?

By Karen Charman in Fair   14 August 2004

If mainstream media outlets had devoted as much ink and airtime to electronic voting machines as they had to O.J. Simpson, Monicagate or even Janet Jackson's breast, the outcome of our next presidential election might not depend on machines that can be programmed to favor one candidate over another without anyone ever knowing. As it is, nearly one-third of the American electorate will cast their votes on one of the more than 150,000 electronic voting machines whose integrity is in doubt.

The manufacturers of touch-screen computerized voting machines—specifically, direct recording electronic (DRE) voting machines—claim to be able to "do the impossible," says Stanford University computer science professor David Dill. "They are using conventional hardware and software, and given that's the case, they cannot make these machines sufficiently reliable or secure to be used in an election."

The biggest problem, according to Dill, is that there are so many ways to rig the machines by ing hidden code. They can, say, lose one out of every 100 votes for a certain candidate, or switch votes from one candidate to another. The machines could easily be programmed to display one thing but record another, so the voter would be no wiser. By writing code to start and stop programming changes at specific times, "you could hide this so that nobody could see it in the program or catch it during testing," he told Extra!. Without a paper ballot that voters can verify, there is nothing to check against the accuracy of the machine tally. And recounts are no longer possible, because the machines can only restate their results.

Diebold, Inc., a major DRE manufacturer, has strong Republican party ties. Last year, the Ohio-based company's CEO Walden O'Dell wrote in a fundraising letter that he was "committed to helping deliver Ohio's electoral votes to the president." As a Bush "pioneer," O'Dell has raised at least $100,000 for the Bush campaign. The fact that a good proportion of the electorate does not understand the threat to our democracy posed by these machines demonstrates how badly the U.S. press fails in its role as democratic watchdog, says media critic Mark Crispin Miller, who has maintained for the last couple of years that this is the single most important issue facing the country. "This should be on the front pages every single day, instead of having people like me talk about it in countercultural venues," he says.

"Far below minimal standards"

Press coverage of electronic voting machines began increasing last July when Avi Rubin and other researchers at the Information Security Institute at Johns Hopkins University analyzed computer code for one of the most commonly used DREs, Diebold's AccuVote, and found it "far below even the most minimal security standards applicable in other contexts." This was the first time that computer experts were able to examine electronic voting software, since the companies who make the machines claim it is "proprietary information"; Diebold had accidentally posted the code on the company's website.

Still, this revelation of the vulnerability of the democratic process did not make it to the front page of the country's leading papers. The New York Times (7/24/03) placed its story on the Johns Hopkins report on page A16. The Washington Post (7/25/03) put it on the front page of its Metro section. The Chicago Tribune (7/25/03) ran a 368-word Associated Press story on page C16. USA Today didn't bother to cover it at all, and in fact a Nexis search turned up only six stories with the terms "electronic voting," "touch-screen voting" or "computerized voting" in USA Today since the 2002 elections.

None of the major broadcast television network news shows—where most people still turn for their news—deemed the report worthy of coverage; this blackout is in keeping with the scant attention TV news has given to electronic voting machines. CNN has done the most, with more than 20 different reports between July 2003 and March 2004. In the same time period, according to Nexis transcripts, CBS News has run six different reports mentioning e-voting machines, ABC and Fox four, and NBC three.

NBC—which broadcasts the nation's most popular network news program—has been perhaps the most lackadaisical in reporting how vulnerable these machines are to tampering or mistakes. Two "In Depth" reports by NBC Nightly News, in a special election series titled "Making Your Vote Count," contained little depth and did nothing to convey that anyone's vote would count. The day after the Super Tuesday presidential primaries (3/3/04), the first of these two reports noted problems with the machines in Maryland and California and had Avi Rubin commenting that a malicious programmer could rig an election. But then viewers were reassured that "the companies that make the devices say they've been through stringent security reviews and are safe and reliable. And state election officials say they're a dream come true."

Six days later (3/9/04), the second report in this series gave more weight to the security concerns about the machines, but suggested that it was futile to expect a solution. After quoting four voters-three of whom expressed doubts about the integrity of the upcoming election-the report ended by telling viewers: "According to our exit polls today, a quarter of the people voting are expecting for their vote not to be fully tabulated this fall. And people, it turns out, are still leery of computers. As much as we talk about the digital age, they like to feel as if their vote goes through, either with the clunk of a voting machine or, heaven forbid, the sight of a chad falling through a piece of cardboard."

Inaccurate characterizations

Widespread problems with e-voting machines in recent elections in Florida, Maryland and California have increased media scrutiny in the last couple of months. However, inaccurate descriptions of the machines persist and send confusing messages. The Los Angeles Times (3/2/04) described electronic voting systems as "straightforward" and cast the issue as "a national experiment in new technology that pits the hope of fewer errors against the fear of election-night computer hacking."

Experiences in California's primary election that day indicated the touch-screen machines are anything but "straightforward," and that their complexity results in many more opportunities for error than the current lower tech alternatives—paper and punchcard ballots, lever machines and optical scanners. Subsequent L.A. Times stories (3/3/04, 3/4/04, 3/6/04) reported that 7,000 voters using touch-screen machines in Orange County were given incorrect access codes and ended up casting useless ballots for the wrong precincts. The San Diego Union-Tribune (3/11/04) reported that 36 percent of the touch-screen machines in San Diego County failed to boot up when polls opened because of faulty batteries. Voter complaints about being given incorrect ballots and mismatches between ballot totals and the number of people who signed up to vote also surfaced in San Diego.

Media reports commonly refer to these problems as "glitches," which downplays their significance, says Rebecca Mercuri, a computer scientist and research fellow at Harvard University. "This isn't a glitch. It's a massive malfunction," she says. "If 36 percent of the cars did not start up at 7 a.m. on a Tuesday morning, do you think we'd call it a glitch?"

Mercuri, who has focused on electronic voting systems since 1989, says DREs have been around since the mid-'80s, and the inability to independently verify their results has been known since then too. "There was nothing new about the machines in 2000 that wasn't already known, and there's nothing new now," she says. In fact, Ronnie Dugger clearly laid out the problems in an eye-opening 22,000-word New Yorker piece in 1988 (11/7/88). But the issue faded until the 2000 election fiasco.

Faced with the inability of the world's dominant democracy to conduct the most central of its functions—a presidential election—the press bought and sold the myth that technology can solve all our problems. Older technology—especially punchcard machines with chads "dimpled," "pregnant" and "hanging"—was, and continues to be, thoroughly vilified. Com-puter systems, the public was told, would straighten out the unsightly mess of the Florida recount debacle.

Dugger's account in the New Yorker, and previous articles in the New York Times (7/29/85, 7/30/85, 8/21/85, 9/24/85), detailed problems with notoriously error-prone punchcard machines, which are counted by computer. But, as Dugger pointed out in an op-ed in the Los Angeles Times early in the election debacle (11/19/00), at least these systems have paper ballots that can be checked against the computerized tallies. Quoting Peter Neumann, one of the leading experts on computerized vote counting, Dugger reminded readers that "hand-counting is substantially more accurate [than computers] in reporting the true intent of the voters."

Editorial boards at newspapers around the country have questioned the reliability of DRE voting machines, and in the last couple of months, mainstream media have begun to produce more probing reports. E-voting machines' vulnerability to tampering is starting to be presented more as a fact than an allegation. But it remains to be seen whether the corporate media will reveal the full story as to how millions of dollars in public money came to be allocated and spent for such fraud-prone equipment—and more importantly, with the presidential election just months away, whether or not it's too late to do anything about it.