When more than 10,000 votes are lost in three counties in a presidential election, and when a review of the balloting shows a large number of undervotes, it is up to elections officials to reassure the public that the voting system can be trusted. But when that reassurance can't be backed up with positive test results, a decision must be made about the system's viability.
That's what officials of the Pennsylvania Department of State did when they conducted a re-examination of the Unilect touch-screen voting machines that were used by Mercer, Beaver and Greene counties in last year's election that featured the race for president.
The state concluded that the system was too inaccurate and inefficient and instructed the boards of elections in the three counties to select other voting methods for this year's primary and general election.
Irregularities in Mercer County in last November's general election turned the spotlight on the touch-screen voting system. Among other things, a computer coding error resulted in 13 precincts in the 4th Congressional District experiencing malfunctions. The foul-up resulted in the resignation of the director of the board of elections, James Bennington, and hearings by the Independent Elections Commission, which was appointed by the three county commissioners. There was testimony from voters regarding the problems they encountered with the machines.
While last week's decision by the Department of State to revoke permission for the continued use of the machines has put Mercer, Beaver and Greene counties behind the eight ball -- they will have to scramble to get optical scanner systems ready for the May 17 primary -- there was no alternative.
A study conducted Dr. Michael Coulter and one of his students at Grove City College found a large number of undervotes in the three counties in the 2004 general election. In other words, many ballots did not contain a choice for president, even though selections were made in other races and issues. When more than a handful of ballots fail to show a vote for president, it's an indication of something wrong.
In light of Dr. Coulter's study and the outcome of the actual testing of the machines, continued use of the Unilect system would certainly have given rise to public skepticism and complaints on primary election day. In addition, any close race would have triggered allegations from the losing candidates that the touch screens had failed. After all, that's what the test of the machines found.
Residents of the three counties should remember that the decertification of the touch-screen system came on the recommendation of Michael Shamos, an expert consultant retained by the department of state. The displeasure voiced by Mercer County Commissioner Michele Brooks over the state's delay in announcing the decision -- the re-examination of the machines occurred Feb. 15 -- is justified. After all, elections boards need lead time in preparing for an election.
But it is also clear that once the state found the Unilect system to be flawed, it could not permit its continued use.