Mostly Good Reviews for Electronic Voting
By JOHN SCHWARTZ  New York Times

Published: November 12, 2004

 Numerous problems with electronic voting machines were reported around the country on Election Day and immediately afterward, but most election officials and experts say the great majority of the machines functioned as expected.

Critics, however, continue to argue that the election did not prove that all is well. In fact, they say, few of the hardware and software weaknesses that make touchscreen machines unreliable and unprotected from hacking have been corrected. And the lack of a paper confirmation for each vote, they say, remains a major flaw in the machines used in most states.

In fact, the lack of paper trails fueled accusations of voting fraud when problems with the machines arose, even when those problems were corrected before the final vote tallies.

In a few states, including Florida, some voters reported that their ion of Senator John Kerry on touchscreens turned into a vote for President Bush, forcing them to restart the process so that their true votes could be properly recorded.

Other problems emerged, including the machine in the suburbs of Columbus, Ohio, that gave nearly 4,000 votes to Mr. Bush in a precinct that had fewer than 800 voters. That miscount, apparently caused by a malfunctioning device that reads voting machine cartridges, will be corrected in the state's final tally, said Matthew M. Damschroder, the director of the Board of Elections in Franklin County, where the mistake occurred.

Another machine malfunction wiped out some 4,500 votes in local races in Carteret County, N.C. - apparently because the machine had been programmed to accept a lower number of votes than were actually cast on Election Day, according to local officials.

Still, a report released on Tuesday by the Caltech/MIT Voting Technology Project, a group that has been working on ways to improve voting systems, found no correlation between the use of high-tech machines and any kind of skewed voting patterns. "We conclude that there is no evidence, based on exit polls, that electronic voting machines were used to steal the election for President Bush,'' the report said.

Ever since the contested 2000 election in Florida, high-tech voting machines have been promoted as a way to do away with the infamous "hanging chad'' problems that occur when voters do not completely push through a punch-card ballot. But critics say that this generation of machines was designed without enough attention to computer security.

Still, election officials who bought millions of dollars' worth of high-tech voting machines and whose reputations hinged on their performance said that the relatively smooth voting process had vindicated them.

"The voting system worked wonderfully," said Linda H. Lamone, the administrator of elections in Maryland, where voting rights groups had vigorously campaigned against the state's 16,000 Diebold AccuVote machines. "Everything went very smoothly."

The skeptics, however, are not convinced. "I guess if you used to applaud Evel Knievel on his successful motorcycle jumps, then you'd applaud the voting machine companies," said John Pescatore, a technology industry analyst for the Gartner research group.

Mr. Pescatore characterized the attitude as, "Hooray! They didn't crash and burn this time!" But that standard, he said, is inappropriate because voting "should be so trustworthy that we expect everything to go right. That is not the case yet with electronic voting machines."

While the 2004 election may not put the debate over electronic voting to rest, it is certain that the use of touchscreen voting will grow as states and counties, with money from the federal Help America Vote Act, upgrade from older voting technologies like lever machines and punch-card ballot systems.

In response, critics are demanding that the machines, at the very least, produce voter-verified paper trails. Nevada has bought machines that provide paper trails, and California requires that all its voting machines provide one by 2006.

In the weeks before the election, lawsuits seeking to force states to offer voters the assurance of some form of paper verification - either by upgrading paperless machines or giving voters paper ballots as an option - were filed in Florida, Maryland and New Jersey. In each case, the courts refused to order changes but left some questions hanging.

Some election officials say they want more accountability and reliability in elections, but do not want their poll workers to have to deal with finicky printers.

"We need to look at a way that's more reliable than simply printing out a piece of paper," Ms. Lamone, the Maryland elections official, said. Her state is working with a voting technology company, VoteHere, to come up with another verification system.

David Bear, a spokesman for Diebold, said, "We're going to meet whatever needs our customers or regulators call for."

Mr. Bear said Diebold had already made progress on a prototype of a voting machine that creates a running paper record of each vote. Its machines have built-in printers that now produce only end-of-day tallies.

The company said that problems reported with its machines during the election were minimal.

The lack of bigger problems, said David Dill, a computer science professor at Stanford University, is "just a matter of luck." Professor Dill said: "It would be the height of folly for us to decide everything was cool going into the next election. We've been granted a reprieve here, and we need to take advantage of it."

Aviel D. Rubin, a professor of computer science at Johns Hopkins University who conducted the first in-depth study of the security flaws in Diebold software, said the situation stood to get worse. "If we don't fix our voting technology situation,'' he said, "we will have a serious and justifiable erosion of public confidence."

The way to reduce distrust of high-tech voting methods, Professor Dill said, is to provide a means of verification. "It's not enough for elections to be accurate," he said. "We have to know they're accurate."