More testing needed on Fla. voting gear: auditors
- By Kathleen Hickey
- Oct 05, 2007
The Government Accountability Office called for additional testing of voting machines used during the 2006 Florida general election to provide further assurance that the technology used did not contribute to the undervote in the state's 13th District.
The agency issued its statement to a special congressional task force now studying why that election produced an unusually large number of ballots that were cast but did not show a valid candidate choice.
GAO issued the information in an Oct. 2 statement
to the House Administration Committee's Task Force on Florida's District 13 Election. Nabajyoti Barkakati, GAO's senior-level technologist for applied research and methods, conducted the election system analysis. He found that prior tests and reviews of Sarasota County's voting systems didn't completely rule out the possibility that the voting systems contributed to the undervote.
Sarasota County, part of Florida's 13th District, used iVotronic direct-recording electronic voting systems (DREs) and the Unity election management system, manufactured by Election Systems & Software (ES&S).
The GAO statement added that additional tests might not completely eliminate the possibility that the machines helped cause the undervote.
"Although the proposed tests could help provide increased assurance, they would not provide absolute assurance that the iVontronic DREs did not cause the large undervote in Sarasota County," GAO said. Absolute assurance is impossible, Barkakati said, because tests cannot re-create the election conditions.
Barkakati proposed three additional tests:
- A firmware verification test to provide assurance that the programs embedded in the computer chips used in all the equipment match the certified version held by the Florida Elections Division.
- A ballot test to determine whether the ways voters were allowed to choose a candidate contributed to the undervote.
- Finally, a calibration test would analyze the effect of an improperly adjusted voting machine.
Once arrangements are made, all three tests would take two weeks to complete, Barkakati said.
The Florida Department of State provided technical comments on the undervote issue, which the congressional audit agency incorporated in the statement.
The equipment vendor, ES&S, also commented on GAO's findings. The company said that earlier test results show the systems worked properly during the election, and that the testing should be focused on the effect of the ballot display on the undervote.
Florida's 13th congressional district includes DeSoto, Hardee, Sarasota, and parts of Charlotte and Manatee counties, according to the statement.
The results from Sarasota County's November 2006 general election showed a significantly higher undervote rate than those of other counties in the 13th District.
GAO's analysis did not specify any particular voting machines or machine characteristics that may have caused the large undervote. GAO did find that the undervotes were distributed relatively equally across all machines and precincts.Kathleen Hickey is a freelance journalist based in the Washington area with extensive experience in covering technology issues.
Kathleen Hickey is a freelance writer for GCN.