In most of Florida's larger counties, repairing the damage from last year's chaotic presidential contest will mean spending millions of dollars on the very latest election technology _ computerized "touch screen" voting machines.
Now comes a study from some of the nation's most respected scientists and engineers, who reported Monday that electronic voting machines _ including touch screens _ have a "comparatively poor" track record in general elections from 1988 to 2000.
In presidential races, voters who used electronic machines erred almost as often as those who used the much-maligned punch cards, according to the six-month joint study by the California Institute of Technology and the Massachusetts Institute of Technology.
In Senate and gubernatorial races, electronic machines performed worse than punch cards, the report said.
The group urged counties to buy optical scan voting systems instead, which require voters to mark a paper ballot by filling in an oval or an arrow, much like a standardized test.
In contrast, touch screens operate much like ATMs.
More than half of the state's 9-million voters, including those in Pinellas, Hillsborough and Pasco counties, could be using touch screens next year as elections supervisors scramble to meet the legal deadline for changing their outdated systems.
Among the group's other ideas: a federally funded research lab that would work with companies to develop new election equipment and test it in small-scale elections.
The report also estimated that 4-million to 6-million people did not vote or have their votes counted in the 2000 general election because of registration mix-ups, poor conditions at polling places, problems with absentee ballots, faulty equipment and confusing ballots.
It offered several ways to recover most of those "lost votes" in the future. Among them: Cut down on mix-ups by using computer laptops at each polling place with access to a central voter registration file, and institute "early voting" at a polling site as a way to cut down on problematic absentee voting at home.
Touch screens are "promising" in many respects, the report said. They make it impossible for people to vote twice in one race. They count votes more accurately than other systems. They are the only systems with listening and puffing devices that allow blind people and those with other disabilities to vote without help.
But the Caltech-MIT team nevertheless branded the sleek, new machines as "untested" and "confusing" and called for the federal government to develop "a coherent national approach" to new voting technology.
In the meantime, the scientists said, counties should choose optical scan systems, which had the fewest voter errors nationally. Indeed, optical scan systems have problems with paper jams, and the ballots are costly. "But it is the best of what is," the report said.
Touch screens have gotten too complicated for voters, said MIT professor Stephen Ansolabehere at a news conference Monday. "One of the important things we see here is the need for simplicity," he said.
"I think the data they used is not accurate," countered Kathryn Ferguson, a former elections registrar who is a spokeswoman for Sequoia Pacific Voting Equipment, one of a handful of firms vying for Florida's touch screen business.
She said the findings included data from older electronic voting systems from the 1980s and did not account for later improvements. Most notably, she contended, the study did not accurately capture the results from the 2000 general election in Riverside County, Calif., widely recognized as the largest, most successful test of an electronic system to date.
The Caltech-MIT report praised that election but said touch screen elections last year in Pennsylvania and New Mexico had voter error rates as high as 5 percent. A rate of 1 to 2 percent is generally considered acceptable.
Caught in this debate are the 41 Florida election supervisors who see problems with both optical scan and touch screen systems, but who, under the state's new election law, must choose one of them in time for the 2002 September primary. It's a tight schedule that forces the supervisors to get the systems delivered, train poll workers and re-educate voters in a few months' time.
Optical scan technology is fine for smaller and mid-size counties, but it "doesn't become very viable for a big, urban county like ours," said Hillsborough County elections supervisor Pam Iorio, who also is immediate past president of the Florida State Association of Supervisors of Elections.
Once a fan of optical scan, Iorio recently crossed into the touch screen column. In a big county, one problem is the inflexibility of the optical scan ballot, she said.
"We have to make the best of what we have available," Iorio said. "Nothing is ever a perfect situation."
"This has not been a whole lot of fun," said Pinellas elections supervisor Deborah Clark, who also is considering a touch screen system. "The timetable is so short, the options are not great, and yet we're mandated to have them in place by next fall. We wish we had better options, but we just don't."
The other counties considering buying touch screens are Pasco, Miami-Dade, Broward, Palm Beach, Sarasota, Sumter, Lee, Collier and Indian River. All have punch card systems.
Recommendations from the report
Major findings and recommendations from the Caltech-MIT report:
4-million to 6-million voters did not vote in the 2000 election or did not have their ballots cast because of election system problems nationwide.
A "National Elections Research Lab" is needed to develop better voting systems.
Optical scan systems are recommended over electronic voting systems such as touch screens.