Voting machine faults ignored

Elections officials were aware that touch screen systems recorded higher undervotes than other machines, but purchased them anyway.

Published July 21, 2004

TAMPA - Florida elections officials knew before they bought the first touch screen voting machine that the devices had a history of problems.

The machines recorded cases in which no vote was cast, known as undervotes, at a higher rate than some other machines.

But election officials bought them anyway, partly because they didn't think undervotes would become a major problem.

Now, undervotes are at the center of the latest controversy surrounding Florida's troubled elections process.

Democratic U.S. Sen. Bill Nelson on Monday called for an independent audit of touch screen machines because of the high rate of undervotes in the March presidential primary.

A task force appointed by Gov. Jeb Bush to recommend changes in Florida elections reported in March 2001 that touch screen machines had a higher rate of undervotes than optical scan machines.

Touch screen machines are similar to ATMs, except they don't produce a receipt. Optical scan systems are similar to standardized tests, in which voters use a pencil to fill in ovals on a ballot that is fed into a scanner that records the votes.

A Caltech/MIT study cited by the task force concluded that the "results probably indicate how "people relate to technologies more than actual machine failures."'

Mark Pritchett, the executive vice president of the Collins Center for Public Policy and the executive director of the Governor's Select Task force on Election Procedures, Standards and Technology, said the undervote issue prompted him to recommend that counties wait for touch screen technology to improve before buying the machines.

Pritchett cited the Caltech/MIT study which said that in presidential elections from 1988-2000, touch screens were "unreliable" an average of 3 percent of the time - the same rate as the much-maligned punch card systems. Those statistics included undervotes and overvotes in the same category of unreliability.

The higher rate of undervotes on touch screens "was a well-known fact in 2001," said Pritchett.

But 15 large Florida counties, including Hillsborough, Pasco and Pinellas, bought touch screens because they are more accessible for disabled voters, eliminate paper ballots and were ready to go.

Pritchett stops short of saying that buying the machines was a mistake. "A more deliberate approach would have been better," he said.

The machines, however, have stirred controversy in recent months among Democrats who complain that they do not leave a paper trail that can be used in manual recounts. The state has told county elections supervisors not to conduct manual recounts with touch screen machines.

Howard Simon, executive director of the American Civil Liberties Union of Florida, says Florida moved too quickly to embrace touch screen technology "before they were adequately tested."

But a leading expert on voting machines says undervotes are nothing to worry about.

"A small but significant number of voters in every election intentionally undervote," said Dan Tokaji, an Ohio State College law professor and the author of a recent report to the Election Assistance Commission on touch screen machines. "Why would somebody do that? I'm not sure if anyone has a clear answer to that question."

Still, undervoting seems peculiar in elections with only one issue on a ballot. Why would anyone take the time to go to a polling place, get a ballot and not vote? Yet that's exactly what happened in March. In Pinellas, for example, 211 voters cast blank ballots in the March primary in precincts that had just one race.

With optical scan voting, Tokaji said, some voters may have deliberately chosen to leave their ballots blank, but felt pressured to select a candidate when their ballots were spit back by mechanical card readers or when a poll worker instructed them that they had made a mistake.

Touch screen machines, on the other hand, include a review screen. If the voter does not cast a ballot in a race, that race and the words "No selection made" will appear in red letters on the screen. If a person actually chooses not to vote at all, the review screen asks "Are you sure you want to cast a blank ballot?"

Secretary of State Glenda Hood said undervotes are not a problem. "That is clearly the intent of the voter not to make a choice," she said. "It is not a lost vote ... I've done it myself."

* * *

-Staff writer Alisa Ulferts contributed to this report. Tamara Lush can be reached at 727 893-8612 or at lush@sptimes.com