Optical-scan beats touch-screen in '02 voting study

S.V.Date in the Palm Beach Post
Saturday, August 7, 2004

TALLAHASSEE Even as state and local elections officials defend "touch-screen" voting machines against charges of inaccuracy, an 18-month-old state report detailed the machines' higher "undervote" rate in the 2002 election.

A Division of Elections analysis of that November's governor's race found that the two touch-screen systems used in 15 counties produced undervotes in which no vote is recorded at all at a rate between two and three times as high as optical-scan systems used in the other 42 counties.

Undervote rates for optical-scan systems:

• Diebold Accuvote, used in 30 counties including St. Lucie, 0.31 percent;

• ES&S Optech, used in seven counties, 0.30 percent;

• ES&S M100, used in 14 counties, 0.53 percent;

• Sequoia Optech, used in one county, 0.48 percent.

Undervote rates for touch-screen machines:

• ES&S iVotronic, used in 11 counties including Martin, 0.92 percent;

• Sequoia EDGE, used in four counties including Palm Beach, 0.93 percent.

The optical scan systems produced "overvote" rates ranging from 0.01 percent to 0.32 percent.

The touch-screen systems had no overvotes they are impossible to cast on touch-screen machines.

The report is required of the division following every general election as a result of the 2001 elections reform law passed after the contentious 2000 presidential election.

It concluded that the use of both systems helped reduce the error rate from the 2000 election.

"Overall, the percentage of uncounted ballots decreased from 2.93 percent in the 2000 election presidential election to 0.86 in the 2002 gubernatorial election," the report stated, adding that the error for the now-infamous punch-card machines in 2000 was 3.93 percent.

The report did not attempt to explain the higher undervote rate for the touch-screen systems.

Since then, state elections officials have said that undervotes on a touch-screen system are intentional, since voters are prompted by the machine when they have failed to make a choice in a particular race.

That higher undervote rate was actually foreshadowed by Gov. Jeb Bush's election reform task force, which in 2001 recommended optical scan machines for all Florida counties for the 2002 elections, leaving open the option of touch-screen machines for future years, as the technology improved.

In its report, the group cited a CalTech/MIT study that found that optical scan machines had lower error rates than touch-screen machines, which its authors attributed to the difficulties some voters had dealing with high-tech equipment.

"The differences in error rates among various kinds of voting systems are much too high to be accounted for solely by uneducated, uninformed or disinterested voters," the report said.

Notwithstanding these recommendations, election machine manufacturers, hiring top lobbyists, successfully sold the state's largest counties on the touch-screen machines, which cost several times as much as optical-scan units.