Skip the navigation

Paper Trail Flawed in Ohio Election, Study Finds

E-voting has benefits, but still a calculated risk, report says

By Marc L. Songini
August 21, 2006 12:00 PM ET

Computerworld - A report questioning the accuracy of Diebold Election Systems' e-voting equipment in a recent Ohio election gives more ammunition to critics who doubt the viability of electronic voting technology.

The report, issued publicly last week, was based on a study funded by the Board of Commissioners of Cuyahoga County, Ohio. It claimed that even backup paper records meant to assure voters that their votes were tabulated correctly can prove inaccurate.

Nearly 10% of the paper copies of votes cast in the election were "either destroyed, blank, illegible, missing, taped together or otherwise compromised," the report said. Ohio law requires that each machine include a so-called voter-verified paper audit trail listing each vote -- and considers that the official ballot.

"There are some serious process issues that need to be addressed," said Steven Hertzberg, project director at the Election Science Institute, which conducted the study of a May 2 primary election in Cuyahoga County.

ESI is a San Francisco-based nonprofit organization founded in 2002 to promote the development of auditable election systems and help ensure that all votes cast are counted accurately.

"What we found is that when you take this [technology] out of the lab and put it in a real work environment with real voters, you're going to have some issues you need to resolve," Hertzberg said.

The May 2 election marked the first use of Diebold's Accu┬┐Vote TSx touch-screen systems in the county, which includes Cleveland and surrounding communities.

In a letter to the county commissioners, Hertzberg said the study found that voters did benefit from the e-voting systems, noting that the Diebold machines are easier to use than the punch-ballot systems they replaced. However, use of the TSx equipment should currently be viewed as a calculated risk for the county, he warned.

For example, the report said that 72% of the polling places demonstrated a discrepancy between the electronic record on memory cards and the paper ballots; 42% of the discrepancies involved problems with 25 votes or more.

ESI also told county officials that some of the voting equipment, including 87 paper rolls and 28 voting machines, was found to be missing prior to the start of the study. Therefore, the institute's report concluded that it is "unable to give a definitive opinion of the accuracy of the Diebold TSx system."

The report also suggested that printer malfunctions could cause "profound" election problems. Such problems could be caused by paper jams or rolls improperly loaded on to the machines. The report urged extensive training of personnel, printer testing and the creation of contingency plans in case of printer malfunction.

E-Voting Troubles
A study of 467 of 5,407 e-voting machines used in the May 2 primary election in Cuyahoga County, Ohio, found that:

One-third of the booth workers had problems setting up the machines.
45% had problems closing out machines.
38% had problems with printers or spools.
90% of the voters liked the new systems.
10% of the voters reported problems with the machines.
24 of the storage devices held no data from the election.

Source: Election Science Institute, San Francisco
A Diebold spokesman last week questioned the methodology used in the study. For instance, he contended that the discrepancies between the paper ballots and electronic records were caused by matching paper votes with the wrong memory cards.

The spokesman also contended that ESI failed to take into account special procedures used for some voters, such as 17-year-olds allowed to vote on selected ballot questions who used separate memory cards.

In addition, the Diebold spokesman charged that ESI and the county commissioners released the report publicly despite hearing of the possible flawed methods earlier from Diebold officials.

In an August 16 letter sent to the county commissioners, Michael Lindroos, Diebold's vice president and counsel, said that the Allen, Texas-based company was "surprised and dismayed" by the publication of the report and noted that ESI did not let the vendor participate in the analysis of the election. "Diebold Election Systems equipment is reliable and accurate," Lindroos said.

The county expected some problems to be found in the study, considering that the election marked the first use of the TSx machines there, said Hugh Shannon, government service coordination manager for the county.

Shannon said the commission, ESI, Diebold and the election board "have agreed to meet and work through the issues of the report and have some more definitive answers by the end of the month to plan for [the] November [election]."

A spokesman for Ohio Secretary of State Kenneth Blackwell said the Diebold machines have been tested and certified by both the state and federal governments.

"The problems in the primary in Cuyahoga County were problems with the procedures and poll worker training," Blackwell said.

This report underscores that voting machines aren't used in a vacuum, noted Michael Shamos, a professor who specializes in e-voting and security issues at Carnegie Mellon University in Pittsburgh. The devices are used as part of a huge system of people, laws and procedures, he said.

Shamos noted that the paper trails didn't guarantee a safe, reliable election. "When machines fail," he said, "the paper trail doesn't work, either."

Read more about Government IT in Computerworld's Government IT Topic Center.



Our Commenting Policies
Blog Spotlight
Sharky

This pilot fish is a contractor at a military base, working on some very cool fire-control systems for tanks. But when he spots something obviously wrong during a live-fire test, he can't get the firing-range commander's attention.

Sharky
Sharky