Researcher teases out truth of election ballots
November 03, 2008
Election forensics, the post-mortem study of elections, was useful in ferreting out the truth of the 2004 presidential race. This time, it might again be brought to the fore in a tight race.
By Lara Zielin
Spero News
The election was in turmoil. The Democratic nominee had won the popular vote, but the electoral vote proved harder to tally. Especially in Florida. An Electoral Commission was appointed to devise a solution. After 16 weeks of heated debate, the Commission put the Republican nominee into the White House, a controversial decision that left a large segment of the United States’ population crying foul. This describes the 1876 presidential election between Democrat Samuel J. Tilden and Republican Rutherford B. Hayes, though the Bush/Gore election in 2000 makes the scenario seem eerily familiar. Presidential elections have never been perfect, but the 2000 election, coupled with problems in 2004 with polling stations in Ohio during the Bush/Kerry election, have renewed concerns about the fallibility and manipulability of the voting system in the United States. As Americans get ready for the polls in November, who is to say their votes will be tabulated correctly, and the true winner will ascend to the White House?
Enter political science Professor Walter R. Mebane, Jr, of the University of Michigan, whose research on elections has led him to coin a new academic niche: election forensics. Like the TV shows where forensic scientists meticulously sift through clues to determine the details of a crime, Mebane collects data surrounding elections and runs that data through complex mathematical formulas to determine where anomalies occurred, and why.
Take, for example, the 2000 election and the much-disputed Florida votes. In Palm Beach County, Mebane and five other researchers looked closely at the butterfly ballot, a voting medium that caused so much confusion that “thousands of voters complained that they had difficulty understanding [it]” (Mebane et al, December, 2001). Many voters in Palm Beach stated they wanted to vote for Gore but, because of the structure of the butterfly ballot, wound up voting for Pat Buchanan instead. The Republican party argued Buchanan had strong support in Palm Beach; the voters argued they’d been duped. Mebane and his colleagues wanted to know for certain one way or the other. Mebane’s solution was to look for anomalies in the data.
“It’s like a fingerprint left at the crime scene,” he says “because it’s like a signature. Only in this case it’s a signature with information instead of human cells.” Mebane tested whether Democratic voters mistakenly voted for Buchanan through multiple methods and concluded that Palm Beach county had an “anomalous excess” of votes for Buchanan — in other words, the heavily Democratic and politically liberal county was filled with people who truly meant to vote for Gore but, because of the ballot structure, voted for Buchanan instead. Mebane says there were about 2,000 of them. Since Bush had a 537-vote victory in the state, Mebane has concluded that Gore should have won, and he penned a paper in 2004 with the definitive title, "The Wrong Man is President!”
So whose fault is that, and what can be done about it? While Mebane believes that the 1876 election between Tilden and Hayes was definitively plagued by widespread, documented fraud, he won’t go as far as to say the 2000 election fits the same bill. Sure, the United States put the wrong guy in the White House, but “people voted for Buchanan by accident,” he says. “They intended to vote for Gore, but they made a mistake, and the quality of the election administration was so poor, voters were never given a chance to correct their ballots, even when they had an idea they were wrong.”
Mebane has considered the role of fraud carefully because in Florida in 2000, voting problems disproportionately affected blacks and Democrats, which raised more than a few eyebrows. Additionally, in 2004, after the voting debacle in Ohio — where voting machines lost votes, insufficient polling equipment created crowded conditions, and long lines deterred voters from casting ballots — the language surrounding the election mayhem had changed. “The word ‘accidents’ was replaced by the word ‘fraud,’” says Mebane. It was an interesting language shift, but it wasn’t one Mebane could address directly since election forensics can’t reveal motives.
“At a crime scene, you can find hair and blood samples that lead you to determine a certain person committed the crime,” says Mebane. “But the hair and blood can’t tell you why. In the same vein, with statistics we can find patterns in the data that suggest a certain outcome is wrong, or that votes were miscalculated, but we can’t tell you if it was deliberate fraud or simply human error.”
In 2007, Mebane presented a paper at the American Association for the Advancement of Science conference in San Francisco, California, and said the goal of developing indisputable election methods is difficult for three reasons. First, because elections for high offices will almost always be closely contested; second, because in close elections, emotions run high and neutral mitigation may be scarce; and third, because voting is largely anonymous, so there’s no way to trace a ballot back to a voter to determine how they meant to vote in the case of a dispute.
But all is not lost. “There are bills before Congress right now that are aimed at getting better results in elections,” says Mebane. “And I’m not the only guy working on this. There are computer scientists, statisticians, other political scientists, lawyers, all working on how to make the election process better.”
Ballot Bumbles
In 2004, according to the federal commission charged with implementing election reforms, as many as one million ballots were spoiled by faulty voting equipment—roughly one for every 100 votes cast. Source: "A Summary of the 2004 Election Day Survey; How We Voted: People, Ballots & Polling Places; A Report to the American People by the United States Election Assistance Commission," September 2005, p. 10.
In 2004, almost half of the 6 million American voters living abroad never received their ballots, or received them too late to vote, after the Pentagon unaccountably shut down a state-of-the-art website used to file overseas registrations. Source: Jennifer Joan Lee, "Pentagon Blocks Site for Voters Outside U.S.," International Herald Tribune, September 20, 2004.
In Ohio alone, at least 357,000 voters — an overwhelming majority of them Democratic — were prevented from casting ballots or did not have their votes counted. Source: “Ohio’s Missing Votes,” Rolling Stone, June 2006. www.rollingstone.com/photos/gallery/10467024/
One in every four Ohio citizens who registered to vote in 2004 showed up at the polls only to discover that they were not listed on the rolls.
Source:(full report in PDF Format)
In 2004, Representative John Conyers (D-Michigan), the ranking Democrat on the House Judiciary Committee, along with the committee’s minority staff, held public hearings in Ohio, looking into more than 50,000 complaints from voters. Source: Preserving Democracy: What Went Wrong in Ohio, Status Report of the House Judiciary Committee
Lara Zielin is Editor of LSAmagazine of the University of Michigan and appears here with permission.
Source: Spero News
