Leo / Friday, September 8, 2006 / Categories: 2006 Initial Results Show Pilot Parallel Testing Program A Success West Palm Beach, FL - September 7, 2006 - A simulated polling place was created at the Supervisor of Elections Office (SOE) on primary day for the county’s first ever pilot parallel testing program. This experiment was done without bias to show the integrity and reliability of the Sequoia voting machines. “It is important the constituents of Palm Beach County (PBC) are confident and comfortable with the equipment and know it actually records the voter’s intent, because every vote does count!” said Supervisor of Elections Dr. Arthur Anderson. “This is unprecedented…no other county in the country has ever taken the initiative, or made the effort individually from the states request, to satisfy the concern of critics regarding the existing logic and accuracy tests of the voting units,” said Jocelyn Whitney, from the independent consulting firm JBS Associates who oversaw the project. “The purpose of this test is to show the results inputted into the voting units will be the actual results recorded and reported when they are tabulated.” To obtain accurate results, a Pilot Parallel Testing Committee (PPTC) made up of community representatives from different political parties as well as SOE staff members was formed. They hoped to gain experience, develop and perform a methodology with the opportunity to refine it’s understanding and execution of the voting process in preparation for November’s General Election. A simulated polling place was set up from 7am to 7pm on Tuesday November 5th using three machines where 73 pre-scripted votes were to be cast on each machine by the parallel election committee members throughout the day for a total of 219 votes. They worked in teams: one member voting, one observing and one signing off that they voted according to each script after completing their assigned ballot. Every minute was recorded on film by a professional independent videographer so the group could go back and look for errors if there were discrepancies during the review process. The Brennan Center for Justice in Washington D.C. recommends using two machines when testing. During an election, the average machine receives 100 - 150 votes per precinct. Early morning during this mock election, there was a random power failure in the elections office. While the lights went out briefly, the machines continued to function without error. In a separate occurrence, one of the machines did break-down due to unknown causes. However, the 24 votes that had been cast prior to the machine’s decline, were retrieved, reported and recorded without incident. The PPTC utilized one ballot style of the 25 possible in PBC with a representative from the Democratic, Republican, and Green parties. The ballots and candidates were selected at random by JBS Associates. A variety of scripts for the committee to follow were also written and provided by Whitney, known to those in the industry as the mother of parallel elections. Human error was intentionally written into the scripts. When the poll closed, all cartridges from the Sequoia Edge AVC™ voting unit and video tapes were put under lock and key in a secured vault until the results would be tabulated the following day after the consultants on the west coast had a chance to review and analyze them. Wednesday afternoon, the Committee regrouped for a debriefing to report the test findings where the ‘tally tape totals’ would be compared with the expected totals sent by the consultants. This phase consists of documenting the findings of the parallel testing; describing all test results, tester errors, usability of the equipment by the tester, any equipment failures, power outages, or other anomalies or events that occurred during the testing. The final report will be published once all reconciliation activities have been accomplished. “There shouldn’t be specific expectations for the results today,” said Whitney. “This is only part one of the process. Initially, the results do not tell the whole picture and additional analysis may be needed. What is important is that we rule out any tester error, and 9.9 times out of 10, that’s where the mistakes come from,” she added. There are only a few possibilities as far as potential errors that could occur: · Script Error - made by the person who wrote the script · Tester Error - made by the person who executed the script · Machine - contains the reports and commands After a meticulous review of each script from each machine, it was concluded that no discrepancies in the tabulation function voting units tested were evident. Although initially there was an under-vote discrepancy with one of the scripts, upon further review of the video tape, it was found to be human error. Therefore, the PPTC can say with certainty both units accurately recorded and reported the predetermined tally results. “Being able to present objective, audited results from a controlled, secured testing environment using randomly selected voting units concurrently with a live election, should offer the best evidence yet that the voting system being used in Palm Beach County is reliable and accurately record votes as the voter intended,” said Dr. Anderson. “We are proud to be the first county to assume this challenge.” There will be a discussion with the vendor (Sequoia) and the Pilot Parallel Elections Committee to discuss the outcome, but overall the testing was a complete success. An additional parallel election is planned for the General Election on November 7th. Previous Article Notice Next Article Glades Central High Participates In iVote Project Print 404 Rate this article: No rating Please login or register to post comments.