Moreover, the voting machine issues and the confusion they caused among poll workers appear to have compounded the delays faced by untold thousands of voters this fall, a preliminary analysis of 1-800-OUR-VOTE reports by Joseph Lorenzo Hall, a researcher at Princeton University and the University of California, has found.

"If we can do anything to improve the experience of the average voter facing a machine problem, it should be reduce the amount of time they spend in line," Hall wrote this week, adding that voters who had machine problems and got back-up paper ballots often were not confident that their votes would count.

"Another curious feature of the data is the voters' uniformly negative attitudes toward contingency or back-up plans," he said. "Voters are often upset and mistrustful."

Hall's analysis is one of the first assessments to look at electronic voting in the 2008 fall election. Many voting rights groups have said the biggest problems this year were inaccurate voter registration records, not enough early voting sites, and planning that did not accommodate high turnout. Hall's findings suggest that the voting machinery used exacerbated these very issues.

One surprise, Hall said, was that the delays in voting did not just come with checking in voters -- but with voters wanting to run their ballots through vote-count scanners. "We have reports of people waiting in line for 3 hours in New Jersey, 3.5 hours in Georgia, 5 hours in Ohio, 6 hours in Missouri," he said. "In many cases, long lines were exacerbated by voters insisting on feeding their own ballot into an optical scan machine, despite it taking a long time to service or replace the affected equipment."

Hall said he was "very encouraged to see that in most cases, emergency ballots were available," though he noted that in Virginia some precincts ran out of back-up ballots. "What I didn't count on was that voters consider voting via an emergency ballot to be fundamentally suspect; that is, most were worried that their vote wouldn't count if cast via emergency ballot."

Poll workers compounded this issue, he said, by confusing back-up and provisional ballots. The latter are used for unlisted voters and must be verified to count.

"We saw cases in at least two states where poll workers were refusing to hand out emergency ballots despite significant machine failures," he said referring to New York and Pennsylvania. "In one case a caller claimed that voters in line were "fighting with poll workers" over emergency ballots," he said. In addition, some voters were upset their vote was not secret because election officials could see their backup ballots.

It is impossible to know how many votes were affected by the issues cited in these incident reports. However, depending on the state and location, individual paperless voting machines could be used by 200 to 600 voters, and paper-ballot scanners could be used to count even larger numbers of ballots. While the presidential results would not appear to be undermined by any of these problems, they do reveal that machinery-related problems are more extensive than many people assume.

Machinery Still Unfamiliar

Hall's report noted other categories of voting machine issues. Since 2002, federal law has encouraged the use of paperless voting systems, especially for people with disabilities. However, Hall said "disability access equipment simply didn't work or was not set up properly" in Arizona, California, New York, Missouri, Minnesota and Maryland. He also noted a report of a poll worker not helping a blind voter because too many other voters were in line.

Machinery malfunctions were a common complaint, Hall said, saying New Jersey and Pennsylvania experienced "numerous reports of lights and buttons not working on machines." In addition, he said there were reports of machines that kept rebooting, "would work only after periodic shaking," or did not work with other computers in the network.

In 14 states, voters reported "vote flipping," where the machines selected another candidate other than their pick (FL, OH, PA, VA, GA, MD, MS, TX, NV, MO, NC, SC, IN, WV).

Voters raised the issue of who was authorized to fix broken machines. In South Carolina, "individuals removed a voting machine from the polling place and took it out to a car to tinker with it," Hall said. In New York, others, including a policeman, apparently "fixed" voting machines. Voters in Ohio and Pennsylvania also noted clocks on some machines were still set to Daylight Savings time, which prompted them to question whether their votes would count.

The process of paperless voting also confused some voters. "A 'fleeing voter' is a voter who leaves a voting machine without having cast their voted ballot," he said, citing incident reports from New Jersey, New York and Pennsylvania where that was an issue. "A 'premature voter' is one who accidentally casts their ballot (or has it cast for them) before they are finished voting their ballot," Hall said, citing reports from Ohio, Pennsylvania, Virginia, Florida and Chicago.

Straight-party voting or selecting candidates from one political party also had glitches, he said. In some instances, not all the candidates were selected, as in Washington, D.C. In Virginia and Pennsylvania, there were reports of ballots where the presidential race "was the only contest available," or the opposite, where the presidential race "was the only race missing."

Computer scanners that read paper ballots had other problems. Reports from Ohio, Virginia, Minnesota, Texas and North Carolina found counters on scanners did not record when new ballots were inserted, an issue that raised concerns about vote count accuracy. In Ohio, Missouri, Illinois and California, printers attached to paperless voting machines to create a paper record of electronic votes failed. In Virginia, scanners could not read wet or humid paper ballots. In Florida and California, using the wrong kind of pen caused votes to be misread, the incident reports said.

Poll workers also were confused with how to handle back-up paper ballots, Hall said. In California, New York and Pennsylvania, it was not clear where to put these ballots after people voted, he said. Some security seals on boxes were broken, he said, and in some cases ballots were "just laying around" or "stacked on top of machines," as was the case in Minnesota.

"In a few cases, poll workers intentionally or mistakenly cast a voter's ballot before they are finished voting or before they've had a chance to revise their ballot," Hall said, citing examples from New York, Virginia, Illinois and Arizona.