By Zachary A. Goldfarb
Special to The Washington Post
Wednesday, July 26, 2006; A15
Last year, a report called "Asking the Right Questions About Electronic Voting" took a look at the issues surrounding the move by most of the country's election jurisdictions to electronic voting machines. The report's theoretical approach contrasted with the often bitter dispute about the security of the technology between activists and voting-machine vendors.
The report's authors -- a committee of National Research Council experts, including prominent computer scientists and two former governors -- then turned their attention to this year's elections. What they found, according to a council analysis released yesterday, is not reassuring:
"Some jurisdictions -- and possibly many -- may not be well prepared for the arrival of the November 2006 elections with respect to the deployment and use of electronic voting equipment and related technology, and anxiety about this state of affairs among election officials is evident in a number of jurisdictions."
More than a third of all of the nation's 8,000 voting jurisdictions will use new voting technology for the first time this year, according to Election Data Services.
"This is a moment of truth for electronic voting," said panel co-chairman Richard L. Thornburgh, a former Republican governor of Pennsylvania and U.S. attorney general. "You've got a lot of people who are working for the first time with the new technology. It should impart a greater note of caution than what you might normally attend to a regular election."
Thornburgh said the analysis is a "caution sign, not a stop sign, but not a clean bill of health for a technology that everyone recognizes there may be problems with."
The new voting technology includes optical-scan and touch-screen machines. In 2004, only 10 to 15 percent of jurisdictions had replaced old voting machines. Widespread efforts to replace outdated voting machines came after passage of the 2002 Help America Vote Act, which set new standards and procedures.
Concerns about the new technology -- largely about alleged vulnerabilities to manipulation -- were raised nearly as soon as the machines were rolled out.
So far, in this year's primaries, the problems have been related to the machines breaking down or being used incorrectly by election officials. For example, optical-scan machines used in a May primary in Cuyahoga County in Ohio could not read the ballots because the black lines separating sections were thicker than on ballots elsewhere in the state, and the fill-in ovals were in a different place, a review recently found. The result was a long delay in ballot counting.
Numerous other localities have experienced problems, most notably the delay in results of a March primary in Cook County, Ill.
The National Research Council analysis notes several potentially problematic areas. Some states may be unable to comply with the 2002 law's deadlines for upgrading technology, meaning it is not yet clear whether they will use old or new technology this year. There are questions about whether voters will be able to use the new equipment without confusion, and whether there is enough time to train poll workers.
"When organizations roll out technology, they do it in a small way. They do a lot of testing and prototyping. We're doing it in one fell swoop and that creates certain kinds of risks," said Herbert S. Lin, a senior research scientist who served on the staff of the committee.
Among the report's recommendations is that jurisdictions run tests on Election Day on randomly selected machines.
Dana DeBeauvoir, clerk of Travis County, Tex., home to Austin, is credited with implementing one of the most comprehensive plans for Election Day. She'll do no fewer than three tests on her voting machines to ensure they are giving accurate results.
"You're always looking for the latest threat. That's not paranoid," she said. "That's good scientific method. We're dealing with voting systems that are scientific instruments."