ELECTION SCIENCE
December 2, 2004

Spencer Michels reports on some efforts to research voting irregularities in this year's election. Terence Smith follows up with Doug Chapin, director of electionline.org, about how the voting process can be improved. The NewsHour Science Unit is funded, in part, by a grant from the National Science Foundation.

 

Online NewsHour Special Report: NewsHour Science Reports Dec. 2, 2004: How the voting process can be improved May 5, 2004: Spencer Michels looks at the controversy in California over electronic voting methods. Browse the NewsHour's coverage of science and politics.   News for Students: Science and Technology reports     Federal Election Commission U.S. Election Assistance Commission VoteWatch	SPENCER MICHELS: The presidential election may have been decided weeks ago, but both activists and scientists are still looking closely at how the vote worked. In the battleground state of Ohio, which President Bush carried by 136,000 votes, civil rights leader Jesse Jackson, speaking at a church, called for a congressional investigation of the Ohio vote. Other Ohio activists have demanded an inquiry into possible fraud. And two third-party presidential candidates filed a lawsuit asking for a recount -- a suit the Democrats said they would join. SPOKESPERSON: We are going to get to the bottom of this. Thank you. (Applause) SPENCER MICHELS: At hearings held by local activists, many complained that Ohio had 92,000 punch cards where no vote for president was registered. They point out that in Cuyahoga County, some areas showed more votes counted than actual voters; a machine near Columbus spit out more Bush voters than there were, but the glitch was caught. Some people charge that long lines, and waits as long as 22 hours in some minority neighborhoods, discouraged voting, and that some election officials barred the media from observing vote counting. The complaints heard in Ohio are typical of those being posted on the Internet from other states. Activists charge that in North Carolina, 4,500 early ballots were lost because of outdated software; in Utah, 33,000 ballots were initially uncounted because of a programming glitch. Researchers say in Florida, President Bush may have benefited from unexplained discrepancies in votes cast by electronic machines. And in other states, they say, votes cast for Senator Kerry were recorded as being for the president. But despite long lines and machine malfunctions in some areas, the final result was not in question, according to DeForest Soaries, head of the Election Assistance Commission, which was created by the 2002 Help America Vote Act. However, he acknowledged some problems did surface.
	Finding the right voting data
	DeFOREST SOARIES: We just didn't have the kinds of problems that many people anticipated. The truth, however, is that we did have glitches. We had problems. The reason we're not talking about those problems today as we were four years ago is because no one perceives that any particular problem impacted the outcome of the election. SPOKESPERSON: Do we subtract it from this? SPENCER MICHELS: Soaries does say, however, that lack of public confidence, as evidenced in the Ohio hearings, and the potential for incorrect results makes it imperative that elections be studied impartially and scientifically. SPOKESPERSON: Voter turnout is awesome. SPOKESPERSON: We need data that would help us understand how many people voted. We can begin to look at issues like equipment, like reasons for over-votes and under-votes. And that kind of data just hasn't been collected. SPENCER MICHELS: Now, a few fledgling efforts are underway, including early steps by Soaries' agency to contact states for voting information the federal government has not previously collected. In addition, the commission will look into machine malfunctions and hold hearings. Until now, scientific studies of election procedures, including registration, have been rare. For the 2004 election, the non-profit, nonpartisan organization VoteWatch, began a nationwide study it hopes will answer some questions about the reliability of the voting process. SPOKESMAN: How confident are you that your vote is going to be counted accurately? SPOKESMAN: I am very confident. SPENCER MICHELS: Steve Hertzberg, an aeronautical engineer and founder of VoteWatch, says that data must be gathered as a first step in solving potentially serious voting snafus. STEVE HERTZBERG: No one has done this in a large, systematic way before, and really this is what's necessary in this country to understand how our election system works. SPENCER MICHELS: On Election Day in Ohio and New Mexico, VoteWatch used volunteers at polling places to gather what they considered scientifically accurate data that could be used to examine the election. Susannah Goodman, a lobbyist for Common Cause, organized the effort in Cleveland. SUSANNAH GOODMAN: We are asking people what their voting experience was. We're asking them, were the voting machines hard to use? Were -- did they get assistance at the polls? Was it hard to find their names on the register? We're trying to document what the trends that we see anecdotally, and we're trying to do that through a scientific way. SPENCER MICHELS: Among the questioners was Charles Izzo, a social scientist researcher from Cornell University. CHARLES IZZO: I would say it's very scientific. A couple of things that really make me say that: If you're getting 70 percent response rate or more, that's extremely good. And we're getting that. And also, this was a very carefully and well-done survey. And they asked all those questions in a very nonbiased way. SPENCER MICHELS: In New Mexico, Fritz Scheuren, president elect of the American Statistical Association, worked for VoteWatch.
	Investigating voting irregularities
	FRITZ SCHEUREN: We're looking for irregularities, okay? We're looking for the kind of thing that would alert us to something wasn't done the way it should be. SPENCER MICHELS: Scheuren says statisticians, not just lawyers and politicians, have to work on elections, treating them as what he calls "measuring" problems. The results of the VoteWatch exit polls in both Ohio and New Mexico were tabulated in the field, gathered in a motel room and then sent back via fax to the organization's headquarters near San Francisco. There, further tabulation took place over the next few days so the results could be released publicly. VoteWatch founder Hertzberg says the results so far are very preliminary, but some things are clear. STEVEN HERTZBERG: The longest wait times that we saw across the country were in Ohio in excess of an hour long on average. We also saw approximately 9 percent of the people in New Mexico not be confident that their vote would count or count accurately. We saw approximately 7 percent have problems with the equipment that they used, where the action they took did not register the result that they wanted. SPENCER MICHELS: With the Election Assistance Commission's Soaries saying the results of this year's presidential race are seemingly clear, the pressure to investigate flaws isn't as great as it was four years ago. But, he says, that shouldn't matter. DeFOREST SOARIES: From an integrity standpoint, the standards of democracy require that the practice of voting be as flawless as possible. SPOKESPERSON: The polls are now officially closed for the 2004 election. SPOKESPERSON: Did you leave blank the presidential candidate? SPOKESPERSON: No. SPENCER MICHELS: VoteWatch and Soaries say Congress can use the data they collect to begin to fix a system whose integrity is in question. In the meantime, activists continue to press for a recount in Ohio.

The NewsHour Science Unit is funded by a grant from: The National Science Foundation. Reports are produced solely by the NewsHour and do not necessarily reflect the views of the NSF.