State conducts random vote audit
Toolbox
By Daniel Barlow Vermont Press Bureau - Published: December 3, 2008
MONTPELIER – It was the human eye versus optical scanners Tuesday as state voting officials organized a by-hand recount of election results from four towns as part of a random audit of vote-counting machines.
Vermont Secretary of State Deb Markowitz said the state has held random audits of the vote-counting machines after every general election since 2003 to determine if there are any problems with how the optical scanners record votes.
After the votes were retabulated, the results showed either no or very minor discrepancies.
"This is designed to ensure that we have security and transparency in how we conduct elections here in Vermont," said Markowitz, noting that confidence in elections has been shaken since the presidential recount in Florida eight years ago.
Dozens of volunteers from across the state turned out Tuesday morning for the recount of the governor and lieutenant governor's race results from four towns – Brattleboro, Rockingham, Barnet and Stowe – in an empty school hall at Union Institute in Montpelier.
Volunteers, including many from the Montpelier and Barre area who were pulled in Monday night when officials realized there would not be enough people on-hand for the recount, took turns counting ballots in stacks of 50 Tuesday morning.
Three of the towns were picked at random to audit, but Stowe was chosen because it did experience some problems on Election Day after the machine's vote total was lower than the number of ballots cast.
That was a problem with how the machine was operated, Markowitz said, and not with the optical scanner.
Brattleboro Town Clerk Annette Cappy said she brought all of the ballots from the town's second voting district, a total of about 2,000 ballots, to Montpelier for the recount audit.
Cappy said these tests shore up confidence in Vermont's election system by showing how accurate the optical scanners used in many towns to count votes really are.
"I expect the hand count to be very close to what the optical scanners found on Election Day," she said. "I'd be very surprised if it was different."
The final tally at the end of the four-plus hour counting session came very close to the results from the machines, Markowitz said, with some towns coming in exactly and some others being off by one or two votes.
That difference is expected because sometimes the human eye reads a ballot differently than a machine, she said.
For example, workers deliberated over a Rockingham ballot that contained a filled in oval for a gubernatorial candidate, but also what appeared to be a cross across the oval. Did the person cross-out the vote after changing their mind?
"It was decided that the mark was nothing more than the movement of the pen," Markowitz said. "But that would be the sort of mark that causes a machine to reject a ballot."
Also on hand Tuesday for the audit were members of the organization Vermonters for Voting Integrity. The group, which formed after the 2000 election debacle, was instrumental in pushing through a 2004 law that required all votes cast in Vermont to be backed up by paper ballots.
Jim Hogue of Calais, a member of the group, said he is not as distrustful of the optical scanners used in Vermont to count ballots as he is of computer-based systems used in other states. But he wanted to be present Tuesday to see how the audit was conducted.
Contact Daniel Barlow at Daniel.Barlow@timesargus.com.
11