A non-partisan coalition monitoring problems at polling sites has reported failures of electronic voting machines around the United States – some of which recorded touch screen votes for candidates voters had not selected. While errors were resolved in the cases brought to the attention of poll watchers, many voters remain uncertain whether their proper vote was cast in a bitterly contested election in which President George Bush has claimed victory.
“A number of people who thought they were voting for Kerry, when the screen came up it showed they were voting for Bush,” said Cindy Cohn, legal director for the Electronic Frontier Foundation (EFF) which is a member of the coalition. “We’ve seen it across several voting systems, not just one machine.”
As of November 3rd, members of the Election Protection Coalition received 4,459 reports of ballot problems, 2,867 calls about polling place irregularities, and 7,152 complaints regarding voter registration glitches. Another 1,074 people phoned in to say that they had witnessed voter intimidation. Unlike the voting problems that occurred during the 2000 elections, the voting problems reported during the 2004 election were spread across the entire country and aggravated the long lines voters endured at polling places. The group Common Cause also reported 50,000 calls reporting voting problems to its election hot-line, although not all the complaints were related to e-voting problems.
Election observers with EFF and the Verified Voting Foundation (VVF) said the reported e-voting failures may be relatively common with touch-screen machines. The groups noted that incorrectly recorded votes made up about 20 percent of the e-voting errors reported through the Election Incident Reporting System (EIRS), an on-line database used by the Election Protection Coalition to record and track voting problems. The EIRS tracked 16,906 total voting incidents on election day.
According to coalition members, the most severe e-voting failures occurred in Pennsylvania and New Orleans where poll workers turned voters away because they did not have enough paper ballots on hand after machines failed. Some voters, perhaps believing claims by e-voting companies that paper ballots were somehow less reliable, or worried about the counting of provisional ballots, refused to fill out paper ballots. Other voters found that their ballot had been pre-marked.
In New Orleans, e-voting machines manufactured by Sequoia Voting Systems failed to boot up on election day and local election officials had no back up plan. According to Cohn, EFF attorneys filed a complaint in Civil District Court attempting to force election officials in the Parish of New Orleans to keep polls open late. The National Association for the Advancement of Colored People (NAACP) also filed a complaint urging polls to remain open late to accommodate disenfranchised voters.
The machines that failed in New Orleans were older Sequoia AVC Edge machines and 80 incidents of failure were recorded across a number of precincts. Some precincts never opened. EFF lawyers also negotiated with election officials in Ohio, Mississippi and Pennsylvania in an effort to keep polls open longer in areas where voting problems had been reported. The coalition said voting machines made by a number of manufacturers malfunctioned including those produced by Diebold Election Systems, Election Systems & Software, Hart InterCivic and Sequoia. Problems with Danaher voting machines were also reported in Pennsylvania.
Will Doherty, VVF Executive Director, said the largest overall problem was a total breakdown of e-voting machines causing some voters to stand in long lines for four to five hours. Some poll workers reported calibration problems with machines. Voters in some precincts waited for an election judge to decide whether to offer paper ballots prompting some to leave before casting their vote. “It’s hard to see how voters in these locations were offered a fair chance to vote,” said Doherty.
In some reported cases, candidates were missing entirely from the ballots. Doherty said there were more reports of Democrats gone missing than Republicans. “Overall we are seeing systemic problems with these machines across all manufacturers,” said Doherty.
Doherty said the problems made clear the need for voter verified paper trails which can be used to document potential tampering of machines or complete failure of the equipment. He added that as of mid-afternoon PST on election day, there were also 405 identification related problems reported involving poll officials demanding that voters prove their identity something they are not required to do.
The Verified Voting Foundation and Computer Professionals For Socials Responsibility created the Election Incident Reporting System (EIRS) to report, track and respond to election concerns. The groups set up a website with an incident report form, and toll-free telephone hot-line 1-866-OUR-VOTE to gather information on voting irregularities.
The web site http://verifiedvoting.org features interactive maps which illustrate voting incidents by county, state and the nation as a whole. The EIRS system is being used by the 60 member organizations of the Election Protection Coalition to build a database of voting information for use in post-election litigation and legislation. The coalition will also use the information to push for new voting regulations while lobbying public officials and e-voting vendors for improved voting processes and procedures.
Because of the 2000 election meltdown in Florida, election watchers closely monitored polling places in that state. Matt Zimmerman, an EFF voting attorney in Miami, Florida said there have been multiple reports of voting machine problems in Florida where incorrect candidates had been selected by e-voting machines and voters had problems going back and changing their votes. He said most of these malfunctions occurred with the Sequoia Edge machine in Palm Beach County were voters were presented with preselected choices on the entire electronic ballot which were often skewed away from Democratic candidates.
Zimmerman added that that there were also questions about how well the security of e-voting machines were maintained while they were being repaired. “Election day is not the day to trouble shoot machines,” he said. “It is hard to make sure that preventative measures can be followed on election day and it raises a whole host of security and integrity problems about those kinds of risks that cannot be fully documented at the end of the day.”
According to Zimmerman, Miami Dade County officials sent a directive preventing polling officials from using backup paper ballots unless there was some catastrophic failure. “The combination of long lines and machine breakdown did not trigger that,” he said.
“We find this very disturbing,” said Cohn. “Voters are upset and some had to go back six or seven times to vote.”Cohn pointed out that in very close elections such as the presidential race, small changes make a difference in who wins or losses, “We don’t have any margin of error for voting machines in a close race,” said Cohen. “That’s particularly troubling.”
Companies that produce electronic voting machines could not be reached for comment on the charges of machine failure. But the Information Technology Association of America which represents e-voting machine vendors issued a statement calling the use of electronic voting machines in the 2004 election a success. “Returns suggested nothing but he accurate and secure operation of electronic voting machines,” said ITAA president Harris Miller.
David Dill, VVF Founder and Stanford University Computer Science professor acknowledged that the coalition received only a relatively small number of reports compared to the number of overall votes cast on the 100,000 electronic voting machines used in 29 states. But he said this was likely because because many voters don’t know how to report problems. “What we are seeing may not be statistically meaningful if you divide the total number of machines by the total number of voters,” said Dill.
But Dill said that extending voting hours helps ensure that everyone could vote even if machines fail. He also pointed out that missing races on ballots and recording the wrong candidate may not even be considered evidence of machine inaccuracy since evaluations by voting machine vendors and academics often focus on analyzing the numbers of vote cast. Evaluators focus on undervotes where the machine fails to record a voter or where voters do not make a selection. Or machines record overvotes where voters chose more options than permitted.
According to Dill, recording a vote for the wrong candidate is not something that shows up in machine statistics and officials have no way of knowing whether the machines inaccurately recorded votes. With no paper trail on many e-voting machines, Dill said there is not way to independently verify where the machine recorded what the voter intended.
Ed Felton, a Princeton University computer science professor who served on the coalition, said errors found by analyzing the number of votes cast in certain precincts – and matching them to the numbers of registered voters or known voting patterns – will only be found days after the election. While mainstream press reports insist that the voting process was generally trouble-free at the nation’s hundreds of thousands of voting places, election watchers caution that there is still much data to be gathered and analyzed.
“We don’t know what percentage of voters we are hearing from and we won’t know until a day or two later how big the problem is,” said Dill. “There are clearly problems out there and I would not prejudge and say the election is going smoothly just because we have a small number of incidents reported based on populations.”
Cohn believes that observers will be forced to file requests for information via the Open Records Act and the Freedom of Information Act before they can really determine the extent of the problems with voting machines. She noted that most reports of election problems go to local election officials who are notoriously stingy about making that information available.
Cohn said she would like to see e-voting machines, especially the troublesome Sequoia machines, outfitted with voter verified paper trails as they were in Nevada this year. But since EFF and other coalition members are non-partisan organizations, Cohn notes that they were shut out from actually watching the votes counted a process which only representatives from political parties can observe. The coalition will not be able to evaluate the machine’s counting software or compare the number of votes cast on the machines with the number of voters who arrived at the polls. “Despite trying, we unfortunately won’t have the kind of widespread nationwide statistics that we hoped to have,” said Doherty.
And there is also no way of really knowing how many e-voting machine failures occurred that were not observed or reported or how many erroneously assigned votes to President Bush. “It gives us this uneasy feeling,” said Cohn. “That we are only seeing the tip of the iceberg for these kinds of problems.”
ANN HARRISON is a freelance reporter working in the Bay Area. She can be reached at firstname.lastname@example.org