NurseWeek: Study finds butterfly ballot baffling

Study finds butterfly ballot baffling
Design causes 'systematic errors'

By Nicolle Charbonneau
HealthScout Reporter
December 1, 2000


	You've read the article. Now tell us what you think. What To Do So how does the Electoral College work? Check out the Federal Election Commission for this explanation. When did you decide whom to support for president? The National Election Studies Web site takes a look at when most Americans decide. And with the ever-increasing problem of low voter turnout, this Atlantic Monthly article proposes that Election Day be made a national holiday.

(HealthScout). A team of Canadian and American psychologists have waded into the controversy over Palm Beach County's "butterfly" ballot – and the findings raise even more questions about the validity of the presidential election results.

In the Dec. 7 issue of Nature, a study of a simulated election in Edmonton, Alberta using the butterfly ballot found a higher error rate compared to a single column ballot format. And according to the lead author, University of Alberta professor Robert Sinclair, it casts doubts on the results from Florida.

Sinclair followed the roller coaster ride of election returns during the American elections and on the following morning, he watched as CNN displayed a picture of the now-infamous butterfly ballot used in Palm Beach County. "I looked at this and I said, 'Who designed this? This is absolutely silly.'"

He sat down and designed a ballot that emulated the butterfly ballot, substituting the names of Canadian political party leaders for Texas Governor George W. Bush, Vice-President Al Gore and Pat Buchanan. The Canadian leaders' names were arranged so that Prime Minister Jean Chretien – expected to win the majority of votes – took Gore's place, and Progressive Conservative party leader Joe Clark – expected to win few votes – took Buchanan's position.

The day after the U.S. election, Sinclair presented the butterfly ballot to 161 university students, while another 163 students "voted" using a single column format ballot. The students marked their choices by darkening an adjacent circle, rather than punching a hole. After voting, the students reported whether the ballot they used was confusing and indicated the candidate for whom they had voted.

While none of the students made any errors, they reported that the butterfly ballot was more confusing than the single column format. Believing that students might have an advantage because they are used to taking multiple-choice examinations, Sinclair decided to test the ballot on a group of average citizens.

By Nov. 9, Sinclair was able to reproduce a more accurate version of the butterfly ballot from the CNN Web site, again using Canadian candidates. He and his colleagues set up a mock polling station at the Bonnie Doon Shopping Mall in Edmonton and randomly approached 116 shoppers. Fifty-three were randomly assigned to the butterfly ballot, while the others filled out the single column format, and all were asked about ballot confusion and whom they had voted for.

Error rate 7.55%
Sinclair says his study took advantage of the (at the time) upcoming Nov. 27 Canadian federal election. "I wouldn't have done this study if there hadn't have been an election coming up in Canada," he says. "If you go out and you're doing a political survey in a shopping mall and getting people to vote, it makes perfect sense to them, because there are political surveys going on all the time before elections."

Four participants made errors on the butterfly ballot, an error rate of 7.55 percent, compared to no errors on the single column format.

"What's confusing is that on the butterfly ballot, the second person in the first column is associated with the third punch hole," he says. "The person that was on our ballot that corresponded to Gore on the Palm Beach ballot was erred against so that some of his votes went to the person on our ballot that corresponded to Pat Buchanan."

Sinclair also asked participants whether they were aware of the butterfly ballot controversy in Florida and whether it affected their vote. Only three of the shopping mall respondents were aware, and none made any errors in their voting.

Stephen Ansolabehere, a professor of political studies at the Massachusetts Institute of Technology in Cambridge, Mass., and a member of the Board of Overseers for the National Election Studies Center for Political Studies in Ann Arbor, Mich., still thinks that awareness of the ballot issue could have been a factor. He also says that such a study would have to include far more participants to be statistically significant.

No intent to meddle
"I showed that the Palm Beach County butterfly ballot causes confusion and it leads to systematic error," says Sinclair. He says that prior to submitting it for publication, he sent copies to both the Bush and Gore campaigns. Neither campaign responded.

Laura Garwin, the North American editor for Nature, says that while the publication of the study was rushed to make it timely, "we didn't publish it with any view to becoming involved in legal processes."

Ansolabehere thinks that in the short term, Sinclair's study probably won't have an impact on the ongoing litigation. The court case would need to find both cause and remedy, he says. "While it's pretty clear that there might be concern about a cause – in other words, that the ballot was confusing and therefore might have produced higher errors – it's very unclear what the remedy is."

At this point, the courts could discard all the butterfly ballots, or hold a re-vote, an option that Ansolabehere believes is unlikely.

But in the long term, these findings may convince counties to drop the use of butterfly ballots in future elections.

For his part, Sinclair is surprised that the problem wasn't dealt with earlier. "It cost me $50 Canadian to do this study," says Sinclair. "Through the application of social science methods and theory, it would have been very simple and inexpensive to do this before the election and prevent the fiasco."

The Republican National Committee, the Democratic National Committee and the Palm Beach Country Supervisor of Elections did not respond to interview requests.

This is a News story from HealthScout, a service of Rx Remedy, Inc.