As many as 6 million votes went uncounted in the 2000 presidential election because of deficiencies in ballots, equipment and voter registration records, a study released Monday showed.
"They found the problem more serious than we imagined . . . with perhaps 3 million to 6 million voters disenfranchised by the complications and uncertainties inherent in present technology," Caltech President David Baltimore said. "The voting process has simply not been taken seriously as a central element of the democratic process."
Nonetheless, researchers at Caltech and the Massachusetts Institute of Technology said that, by the time the 2004 presidential election comes around, reforms to the nation's creaky voting system could pare by one-half the number of "lost votes"--instances in which a person was not able to vote or their ballot was not counted.
Those reforms would include eliminating punch-card balloting, the installation at polling places across the nation of machines that scan paper ballots, and equipping poll workers with comprehensive voter lists.
"We want every American's vote to count in our elections," MIT President Charles Vest said during a news conference staged simultaneously in Pasadena and Cambridge, Mass. The study was financed with $250,000 from the Carnegie Corp.
Researchers found that optical scanning machines are the most reliable method of counting votes, particularly when used at the precinct level. Along with eliminating punch cards and their now infamous chads, lever-operated machines also should be replaced, the report said.
Researchers warned, however, that technology is not a panacea to vote-counting problems.
The brain trust of 10 political scientists, electronic engineers and computer scientists found that some electronic voting machines, such as those that use buttons, were more prone to error than optical scanners. It also said that touch-screen voting remains "unproven" and that Internet voting is far from secure. Internet voting, the authors said, is a decade away because of the potential for fraud and hackers.
Researchers estimate that 1.5 million of the 2 million presidential ballots that were unmarked, spoiled or ambiguous in November were the result of faulty equipment. Researchers base that estimate on studies of equipment failure rates.
Using answers to census survey questions from people who did vote, researchers estimated that 1.5 million to 3 million more people were disenfranchised because of "registration problems."
Census data also led the researchers to estimate that 500,000 to 1.2 million potential votes never were cast because of problems at the polling stations, such as lines, operating hours and inconvenient locations.
"The lower numbers are conservative estimates," Caltech associate professor of political science R. Michael Alvarez said. "Even those numbers are a lot more than we've considered in the past."
Many of the voter registration problems stem from inadequate election information for poll workers and the absence of provisional ballots in some states, the 92-page report said.
A provisional ballot is one cast by a voter when poll workers cannot verify eligibility. Such ballots aren't counted until it is verified that the person is an eligible voter and did not vote elsewhere. The report said that provisional votes should be legal nationwide. Alvarez said that, unlike California, 19 states currently don't allow provisional ballots.
Poll workers, the study concluded, need to be equipped with paper lists or, preferably, laptop computers with complete county or state registration lists. Such lists, plus provisional ballots, could allow 2 million more votes to be cast, the study estimated.
"When I voted in the recent school board election, a neighbor of mine was turned away because he wasn't listed for that polling place," Alvarez said. "He was sent to a school to vote where there weren't ballot boxes. That wouldn't happen if every polling station had the entire registration list."
The reforms, the study said, would come at a price: about $400 million, with half of that going for new equipment. "We think the price of these reforms is a small price to pay for insurance against a reprise of November 2000," said Thomas Palfrey, a Caltech economics and political science professor.
While saying they want to increase voting, the study's authors suggest that absentee voting should be replaced by early voting, which would allow people to cast ballots days or weeks before an election.
"Absentee ballots are fundamentally not secret ballots," said Charles Stewart, an MIT political science professor. Researchers concede that any such move would face opposition from voters, politicians and the two major political parties.
The electoral system also needs a long-term overhaul, the study said. Researchers said that all 3,066 counties nationwide cannot afford the overhaul and would need federal grants to make the improvements.
A research program also is needed to test new voting equipment and develop improved ballot designs, they said.
Researchers had a proposal for double-checking that a ballot is properly cast. They said that, after voters are identified at polling places, they would be handed an electronic card or an electronic recording device. which the study called a FROG.
The voter would insert the FROG device in one machine and make their ballot selections. Then the voter would insert the FROG recorder in another machine and cast their actual ballot. During the news conference, a researcher produced a toy frog from his pocket to explain the concept.
Caltech's Palfrey, however, said that whatever evolves at the ballot box, there are limitations. "It's true a lot of these errors are due to human interface with the technology." In other words: human error.