Two audits of South Carolina’s November 2010 general election found scores of human errors that led to incorrect vote counts and other problems. None of these errors were large enough to have changed the outcome of a election or referendum, but they were significant enough to prompt the State Election Commission to make several procedural and policy changes. The problems also emboldened the chorus of critics questioning the accuracy, reliability and accountability of the state’s iVotronic voting machines.
And they could prompt the Legislature to lengthen the time period between Election Day and when counties meet to certify the results. That added time would give counties extra time to audit their data before formalizing their tallies. State Sen. Chip Campsen, R-Isle of Palms, has chaired a Senate Judiciary Subcommittee looking at elections and has reviewed the audits’ results. “The problem is these problems were uncovered after the election was certified,” he said. “Once an election is certified, it can’t be undone.”
Barbara Zia, co-president of South Carolina’s League of Women Voters, said the scrutiny of the state’s election system was triggered in part by the June 2010 Senate Democratic primary in which an unknown candidate who didn’t campaign won handily with 60 percent of the vote. The league’s recent audit — which requested information from all 46 counties under the state’s Freedom of Information Act — was an outgrowth of that.
“We’re calling on the state and the state Legislature to ensure election integrity by enabling state and local election officials to conduct vote audit after each election to ensure the accuracy of the vote count,” she said.
Duncan Buell, a University of South Carolina computer science professor, helped spearhead the League’s audit. “By my count, there were eight counties for which there were significant problems in terms of being able to get an independent audit to determine if the numbers were correct.”
He noted Williamsburg, Orangeburg and Lancaster counties had no electronic date available, while Oconee and Horry counties were unable to produce usable audit files. In Richland County, more than 1,000 votes in two different precincts didn’t get counted. Colleton certified incorrect totals because of human error, and Charleston County was unable to account for 35,000 votes, or about 25 percent of the total, in the audit.
Taken together, Buell said, “we have a serious problem in having an election where we can go back and get results and be sure we have the right answers.
I think what we have a system that’s not acceptable because there are too many errors happening.”
Buell noted only 15 of 46 counties provided all five files requested. “That’
s not a very good record,” he said. “For some of those counties, there were a few that produced only tiny fractions of the right data. I think we deserve better than that.”