Study shows higher voting accuracy

Electronic voting machines reduce undervoting in Georgia

From staff and wire reports   Macon Telegraph   03 December 2004

Undervoting, a key measurement of voting accuracy, ped precipitously in Georgia during the 2004 elections, according to figures released by the Secretary of State's office.

That means Georgia's electronic voting systems are the most accurate yet, elections officials said.

Georgia went from having the second-highest number of undervotes in the country to the lowest, according to the Georgia Secretary of State's Office. In addition, individual precincts with a high percentage of black voters saw the greatest improvement, according to a Secretary of State's Office study that focused on majority-black precincts in several counties, including Bibb.

An undervote refers to a ballot that, for whatever reason, doesn't include a valid vote for one or more races. In this case, the study looked at voters who failed to vote in the presidential race. There's no guarantee that ballot represents a misread or an error by the voter, but common sense suggests that if a voter is going to the trouble of voting in a presidential election year, they're going to vote in the presidential race, said Chris Riggall, a spokesman for Secretary of State Cathy Cox.

To that end, the number of Georgia residents who cast ballots, but failed to vote in the presidential race, ped from 3.5 percent in 2000 to .39 percent this year. Likewise, in precincts where more than 80 percent of registered voters are black, the undervote rate ped from 6.7 percent to .69 percent.

Precincts in poor and minority neighborhoods have traditionally had higher undervote rates, largely due to old or faulty voting equipment, experts say. A similar look at precincts with more than 80 percent white voters showed a smaller, but still significant, decrease - from 3.15 percent in 2000 to 0.21 percent this year.

Bibb's overall rate ped from 4.4 percent to .47 percent and Houston County's went from 3 percent to .31 percent.

Undervotes occurred more often in 2000 "because people were screwing up the optical scan or they were misreading the punch card," said Elaine Carr, Bibb's elections supervisor.

"The electronic ballot has made it a more accurate voting system," Carr said.

The Nov. 2 election marked the fifth time that the state's $54 million electronic voting system was used for a statewide ballot, and the first time in a presidential election. Adopted in 2002, the touch-screen machines replaced a hodgepodge of punch cards, optical scan machines and paper ballots.

"This was the first opportunity to have an apples-to-apples comparison," Riggall said. "To be able to see dramatic improvement in the accuracy of the vote count is not only a positive outcome, but just confirms what we set out to do four years ago."

The greatest reduction in undervotes was in a Muscogee County precinct, where presidential no-votes ped from 21.8 percent in 2000 to 1.3 percent this year. The North Lumpkin precinct in Columbus used an optical scan voting system four years ago.

Critics complain that the new machines do not create a paper record of each vote, saying such records are necessary for recounts and in case any results are questioned. Some fear computer hackers could break into the voting system and tamper with results, despite Cox's assurances that the system cannot be accessed by the Internet.

"We're still left with the question of how do we know they counted up the actual vote," said Andrew Appel, a computer science professor at Princeton University who teaches a course on electronic voting. "An absence of evidence of things going wrong is not the same as knowing they didn't."

Appel said Georgia's increased accuracy rates sound consistent with voting on computerized systems in other states. He said he expects electronic voting companies - including Diebold Election Systems, which makes the Georgia machines - to eventually offer touch-screen equipment that also offer paper ballots.

Cox has said she is not closed to the idea of a paper ballot, but that the technology currently available is expensive and unwieldy.

Regardless, a majority of Georgia voters have given high marks to the new voting system.

In a University of Georgia poll conducted a year ago, 70 percent of voting-age respondents said they preferred the new computerized system to any other voting method. Twelve percent chose paper ballots and eight percent opted for punch cards.

Support was even higher among the 1,618 Georgia voters in the Nov. 2 elections who participated in an exit poll by the Associated Press. About nine out of 10 voters polled said they were confident that their votes were accurately counted using the touch-screen machines.