It ain’t peanuts

The 2002 Georgia elections were odd.

Last-minute polls showed incumbent Gov. Roy Barnes leading by between nine and 11 points. In the U.S. Senate race, polls showed popular incumbent Max Cleland leading Republican challenger Saxby Chambliss by two to five points.

Georgia has a long tradition of electing Democrats to statewide office, so these polls were no surprise. But when the votes were tallied, Barnes lost to Republican Sonny Perdue, 46 percent to 51 percent — a shift of up to 16 percentage points. Cleland lost to Chambliss 46 per cent to 53, an election-day shift of up to 12 points.

Voters do change their minds, and odd results do happen from time to time. But on closer examination, some very curious details came to light, leading some analysts to question both the legality and fairness of the election.

In 2002, Georgia became the first state in the country to install touch-screen voting machines in every precinct, and the result was a shambles. Machines froze up and waiting lines grew as technicians struggled to reboot. In greater Atlanta, 77 memory cards disappeared, delaying certification of the results for 10 days. Some of the missing cards were found in terminals that had supposedly been taken out of service.

How or even if those missing cards were actually tabulated is unknown, because the vote count was not conducted by state election officials. Diebold sold Georgia the voting machines with a strict trade-secrecy contract that made it illegal for the state to touch the equipment or examine the proprietary software. There was no auditable record, so recounts were impossible.

The Independent reported last October that further inquiry by a group of Georgia citizens revealed that in the days leading up to the election, a software patch had been installed on 22,000 machines and the patch had never been certified. According to technicians who administered the patches, they were told to download them directly off the Internet from the Diebold FTP site.

Last November, Mark Radke, director of marketing for Diebold Election Systems, told the Boston Globe: “Each of the jurisdictions, before they deploy this equipment, will put it through … logic and accuracy testing. … It goes through Federal Election Commission certification, using independent testing authorities, before the product is ever sent out.”

But in response to a Freedom of Information Act request by state voters, Georgia officials admitted that they didn’t have any of the certification documents required for the software following installation of the patches.

All electronic voting machine manufacturers routinely issue software patches that are dutifully loaded onto local machines. And computer-security experts say there’s no way to be absolutely sure that these patches do not contain inadvertent or intentional “errors.”

Similarly odd results have occurred in elections across the country where DREs have been used. According to published reports in the Independent, The Nation, The New York Times, the Philadelphia Inquirer, Vanity Fair and Wired News, candidates for different offices have received identical vote totals; in still others, the point spread between candidates in different races were identical; precincts have reported total votes cast dozens of times the number of their registered voters; and one county in Florida actually reported negative 16,000 votes for Al Gore — an electronic mistake that might not have been noticed in a less tumultuous year.

— Cecil Bothwell