Thursday, September 02, 2004
Voting without a net, or paper trail
Electronic voting machines have been hailed across the nation as a solution to problems with the old-fashioned paper and punch card systems that created controversy in the 2000 presidential elections in Florida.
They have also been hailed locally as a solution to the problems with the 1930s-era lever machines that lost an undetermined number of votes for candidate Alice Hincker through a malfunction during May’s Roanoke mayoral race.
But concerns over potential software glitches, vote miscounts and ballot security have prompted Virginia and other states to take a second look at the "new era in voting." As a result, the Virginia General Assembly created a commission charged with studying the security and reliability of electronic voting machines. The first meeting was in August.
Unfortunately, recommendations from the commission and law changes from the General Assembly are not expected to emerge until 2006 or 2007, yet electronic voting machines are already in use in Roanoke City, Roanoke County, Botetourt, Floyd and are coming soon to Montgomery County.
As America gears up for the 2004 presidential election, it is estimated that more than 50 million citizens will use electronic voting machines in November to register their votes.
Most electronic voting machines resemble laptop computers with touch screens. They store votes electronically, and most of them don’t have a paper backup. The Virginia State Board of Elections has approved seven different models of electronic voting machines that localities can use.
The problem is — as any of us who use computers know — computers can crash. If a voting machine crashes, votes can be lost.
The machines can also produce bad data as a result of software bugs. How many times a year does Microsoft — the biggest software company in the world — issue patches to fix software problems they discover after they’ve already sold their products to the public? Several manufacturers of electronic voting machines have released software updates and "fixes" for their equipment. Does that mean that before the "fix," the machine counted votes incorrectly?
Despite this, it seems that many people, including many election officials across the nation, are so dazzled by new technology that they believe these machines are close to infallible.
To the contrary, Verifiedvoting.org is one of many organizations calling for some kind of paper backup for the machines, and it has a 51-page document showing case after case of electronic miscounts and malfunctions in recent elections across the country. Touch screen failures have occurred in elections this year in Georgia, Maryland, California and other states.
One Virginia case occurred in Fairfax County’s November 2003 school board election. After questions were raised about electronic machines that were acting strangely, county officials tested some of the machines and discovered that they subtracted a vote in about "one out of a hundred tries" for one particular candidate. The WINVOTE machines were the same brand of machines currently used in Roanoke County, Botetourt, Floyd and soon in Montgomery County.
In another case of machine failure during the much-reported Florida 2000 election debacle, one story that was not as prominent was about a precinct in Volusia County where only 412 people voted, but a Diebold Election Systems electronic machine actually gave Al Gore a negative 16,022 votes and George Bush a positive 2,813 votes. (I can’t really complain about that one too much.)
In a Virginia-related story, the state has one of the much-discredited Diebold Election Systems electronic machines on its approved list.
Diebold is one of the largest manufacturers of electronic voting machines. The reason that its machines are discredited is that during a web search in 2003, voter activist Bev Harris discovered an unprotected Diebold Web site (open to the public) where she could download the program that ran the company’s AccuVote machine.
Ms. Harris contacted Aviel Rubin, a computer scientist at Johns Hopkins University, director of the university's Information Security Institute and a panelist for the National Science Foundation’s e-voting feasibility study.
In 2003, Rubin and two graduate students analyzed the downloaded Diebold code and found that the company's programmers had written the key for unscrambling the system's encryption directly into the program. This meant that anyone who downloaded the program code from Diebold’s Web site, like Ms. Harris did, would have the key. The same key unlocked the data on every Diebold voting machine of the same model in the country.
Rubin’s report led to congressional hearings and a flood of media coverage. The hearings uncovered testimony that the same problem was pointed out to the company six years earlier by a computer scientist from Iowa, but it wasn’t corrected. Company spokespeople say they finally fixed the problem last September.
Virginia still has Diebold machines on its approved list. Michael Brown, chairman of the Virginia State Board of Elections, said in 2003 that a consultant the state was working with assured him the Rubin report was flawed, although many computer scientists and members of various states’ electronic voting commissions disagree.
Besides these and other documented machine failures, there’s another issue that hurts the credibility of electronic machines. Independent labs that test the machines are paid by the manufacturers, and they are under contract to not talk publicly about any problems they may encounter. Often the testing facilities are shrouded in secrecy. In addition, only three labs are approved to test the machines, and in some cases, they are a year behind in doing it.
Many proponents of electronic voting will say that the punch-card machines, optical scan machines and mechanical lever machines used in the past had flaws, too. They are absolutely correct. Those machines broke down and miscounted votes as well. The problem is that if we are spending billions of taxpayer dollars nationwide (and millions in Virginia) on upgrading our voting equipment, let’s make sure the technology works and that it’s truly an upgrade over what we already have. Let’s open the veil of secrecy that surrounds the companies that produce these machines and the labs that test them.
The common saying many politicians seem to share is that "voters must have confidence in the integrity of their elections." It is frightening that in the country that espouses the highest principles of democracy and even helps start democracies around the world, we have a problem where the citizenry is skeptical that their votes are counted.
The solution to these problems seems to be (1) a more transparent testing procedure where electronic voting machine testing can be done in the open and face public scrutiny, rather than behind closed doors, and (2) voter-verified paper ballots for machines. The overwhelming recommendation from voting activitists, including the highly respected League of Women Voters, is that every machine should produce a paper printout of the voter’s selection that the voter could see but not touch. That printout would verify that the machine recorded the votes correctly before the receipt is deposited into a secure ballot box. This would address the issue of software malfunctions. It would also provide a paper trail for potential recounts. Without a printout, there is no way a recount can be done accurately.
Current proposed legislation in Congress, H.R.2239 and S.1980 (House and Senate bills), for example, has language that addresses many of these issues.
Many of the voting machine manufacturers oppose paper-verified voting, have lobbied the federal and state governments hard against it, and have tried to smear and discredit paper trail advocates. The question is "Why?" Do they have something to hide?
Many state and local elections officials around the country and in Virginia are resistant to paper verification, too. In a recent Wired News report, wired.com reported that many officials oppose the idea because it will cause more problems for poll workers if the printers break down or run out of paper. It will cause longer lines at the polls as people take time to verify their ballots, and it will increase the call for more recounts.
Well, when we go to the grocery store and the cash register runs out of paper, we wait the whole one minute for the 16-year old cashier to replace it. I don’t think that replacing the paper roll once or twice on a voting machine will cause a poll worker to go into fits.
And frankly, when the future of the free world is at stake, I don’t mind waiting in line at the voting booth for 20-30 minutes for one day out of every couple of years while citizens verify they voted for the right people.
Regarding the potential for increased recount calls, these challenges show that democracy works. If an election is close or voting is questioned, recounts can reassure the public that the election was fair. Again, when the future of the free world is at stake, I’m really not that worried that a recount is going to ruin some election official’s vacation he planned to take after the election.
I think there is some cultish-like sacredness about elections to many (not all) election officials, to where they will almost never admit when there is a problem. It seems more important to some of them to make sure that the integrity of the electoral process is never questioned, even if that means sweeping the questions under the rug to pretend they don’t exist.
I saw this firsthand in Roanoke’s mayoral race last May. When we detected an obvious voting machine malfunction (it was on the old lever machines – and they have no paper backup), it seemed to me (the campaign manager), the candidate (Hincker), and the press that on several occasions the electoral board wanted to come up with a quick answer to explain away the problem, rather than investigating to see whether the problem existed in other machines. We were continually told that malfunctions in other machines were "impossible," because the board would have detected them if they had happened. But we were never offered convincing details to erase the doubt — and neither were the voters. We were just expected to believe what we were told. In the end, the lost votes were just thrown out, because there was no paper trail from which to recover them.
I even had one election official write an e-mail response to my questions about the malfunction incident. In it, he said that I could write him back all the e-mails I wanted, but that he would not respond to any more of them because he said all he was going to say about the incident, and in his opinion, there was nothing more to be discussed.
Nevada is the only state that will have paper verification statewide for the upcoming November elections (California and Missouri will have it for some precincts). A lot will ride on how successfully that process goes. At least 20 other states are considering legislation to require a paper record of every vote cast.
Whether paper-verified ballots will be added to Virginia’s electronic voting machines depends on whether Congress or the Virginia General Assembly passes a law after the commissioned studies are concluded. Either way, sadly it looks like paper verification is at least a couple of years (and elections) away.