Speaking from a witness stand rather than a Silsby Hall lecturn, government professor Michael Herron testified on Dec. 20 in defense of an electronic voting company being sued by Florida congressional candidate Christine Jennings. Jennings, the Democratic candidate for Florida's 13th-District House of Representatives seat, sued Election Systems & Software for access to the programming codes from the touch-screen voting machines that were used in November's election -- codes she believes could prove machine errors deleted votes.
In Jennings' race against Republican candidate Vern Buchanan, which Buchanan won by 369 votes, approximately 18,000 Sarasota County ballots showed no choice in the Jennings-Buchanan contest. Jennings, contending that such an unusually large undervote might be evidence of a machine malfunction, has sued ES&S and is calling for a re-vote. The company, however, has refused to reveal the programming codes, calling them "trade secrets."
Miguel DeGrandy, the lawyer representing ES&S, contacted Herron after the professor and three colleagues posted a paper online about the election.
"When we originally wrote it we just thought it was an interesting exercise," Herron said, calling the experience "an interesting opportunity to get data on voting machines."
DeGrandy's office had been exploring a ballot format explanation for the undervote when they found Herron's paper, which supported their theory that the results were not due to machine error.
In the work, Herron, Laurin Frisna of the University of Rochester and James Honaker and Jeffrey Lewis of University of California Los Angeles looked at voting data from several counties and used it to argue that the dramatic undervote recorded in Sarasota County resulted from voter error from a "priming effect."
"Voters appear to be conditioned by what happens on the first page [of the ballot]," he explained.
In the Sarasota County ballot, the first page presented voters with only one race -- the contest for U.S. Senate. However, on the second page of the ballot, both the Jennings-Buchanan race and the gubernatorial race were displayed.
Herron said that, based on research and evidence of other ballots formatted like this one, this change in format between the first and second pages led voters to expect only one race per page, meaning voters accidentally did not vote in the Jennings-Buchanan race.
During the trial, Herron also pointed to a voting trend in Florida's Miami-Dade County as further evidence for this explanation. On the first page of the ballot presented to voters in this area, some voters saw two races, those for House and Senate, while others saw only a Senate race because there were uncontested elections. All voters faced three elections on the second page of the ballot.
When Herron examined the voting patterns of ballots from this area, he found that among those voters who had encountered only one race on the first page, 89 percent of those voters who skipped the second race on the second page of the ballot also skipped the third race on this page -- suggesting that these voters were "primed" to only expect one race.
In contrast, among those voters who faced two races on the first page of their ballot, only 73 percent of those voters who did not vote in the second race also skipped the third, indicating a less strong priming effect.
"When you first start looking at these things you say 'Boy, it's a mystery,' but once you look at ballot format, it seems it's not a mystery anymore," Herron said.
Herron said that his research did not lead him to believe that machine malfunction was involved in the Sarasota County undervote.
Circuit Judge William Gary ruled with the defense on Dec. 29.
DeGrandy said that he believes that the fact that Herron had already done research on this issue previously to connecting with the case was likely significant to Judge Gary.
"What was very strong about [Herron's part in the case] was that we didn't hire him to render an opinion," said DeGrandy, "He already had done his research and already had an opinion."
In his decision, Judge Gary called Jennings' evidence of machine malfunction "speculation and conjecture."
Jennings' office has said they will ask for an immediate appeal.
"We were disappointed [with Gary's decision] and also frankly just shocked that the state of Florida is continuing to put the interests of a private company in front of the public," said Kathy Vermazen, Jennings' press secretary.
Vermazen noted that the Jennings campaign has agreed to a protective order that they feel would aid privacy concerns.
As a potential counter to Herron's theory, Vermazen pointed to accounts from Florida citizens who have said that they encountered machine problems while attempting to vote.
"Just walking around town in Sarasota," Vermazen said, "[Jennings] gets stopped nearly every day by people talking about the problems they had."
And some of these voters have joined Jennings' Florida lawsuit.
"To discount hundreds of voters without even taking a look at the machines is pretty disturbing," Vermazen said.
Jennings has also asked Congress to conduct an investigation into the election. Salley Collins, a spokesperson for the House Administration Committee, the body responsible for federal election oversight, explained this process.
"The committee will conduct an investigation based on the nature of the contest [and] compile a report along with a recommendations to present to the House," Collins said.
Collins said that the committee is likely to wait for the state lawsuit to play out before performing its investigation.
Jennings officials believe that Jennings would win a re-vote in which the 18,000 "suppressed" votes were accounted for. In his report, Herron agreed that Jennings would likely win an election with the missing votes included.
Buchanan will be seated as scheduled in Congress on Jan. 4, though he could be removed if a re-vote finds Jennings as the election's winner.
