By Katie Wagner, Staff Writer
A Princeton University professor is proceeding with a study of inconsistencies in the printouts and back-up tapes of certain electronic voting machines used in New Jersey’s Feb. 5 presidential primary, including Mercer County, despite the refusal of the manufacturer to allow him to examine the machines.
The Constitutional Officers Association of New Jersey (COANJ) had asked Edward Felten, a computer science and public affairs professor at Princeton, to study the cause of discrepancies in the Sequoia Voting Systems machines that occurred in several parts of the state.
The COANJ includes all of the state’s county clerks, sheriffs, registrars and surrogates who must certify election results.
Professor Felten’s study was to have included experimenting with and analyzing the Sequoia voting machine models used in counties that found discrepancies in voter turnout until Sequoia’s compliance officer threatened to take legal action against Union County Clerk Joanne Rajoppi if she gave Professor Felten one of the county’s machines. In an e-mail message, the voting machine manufacturer also threatened to sue Professor Felten for conducting the study.
Ms. Rajoppi was the first to discover differences between the numbers of registered voters that appeared on the cartridge printouts and paper-tape backups from the machines used in the Democratic and Republican primaries.
Professor Felten said he thinks it’s unfortunate that Sequoia is trying to prevent independent studies of its machines and hopes that he and his team of researchers will ultimately be able to study them.
He said he has obtained all nine of the problematic paper-tapes from Union County and hopes to get paper tapes from all machines in the state that inconsistently recorded voter turnout.
Following notification of the discrepancies from Ms. Rajoppi, Sequoia examined some of the machines used in the New Jersey primary and claimed, in a statement, that the errors resulted from poll workers using the wrong buttons on the machines’ control panels.
”On my watch, for every primary that we’ve conducted over the past 10 years there’s never been a similar problem and I find it very odd that the discrepancies would occur as a poll worker error,” Ms. Rajoppi said. “If a poll worker initiated it, why did it happen, because a machine should not allow it to happen.”
Ms. Rajoppi, a member of COANJ’s executive board, and her fellow COANJ members think Sequoia’s self-evaluation is not adequate and have expressed this in a Wednesday letter to the state Attorney General Anne Milgram. The letter urges the attorney general to join COANJ in seeking an independent investigation of the machines and inform Sequoia that public confidence trumps a trade secret agreement that does not permit the machines to be released to outside parties for examination, which Sequoia claimed had been made between it and Union County in e-mails to Professor Felten and Ms. Rajoppi.
Of the 21 counties in New Jersey, 17 use the same Sequoia voting machine models, Ms. Rajoppi said. So far, officials in Union, Middlesex, Mercer, Gloucester, Cape May and Bergen counties have found differences between the machines’ two records of the numbers of Democrats and Republicans to have registered to vote on the machines in some of their precincts.
Mercer County had more machines producing errors than all the counties in the state, with 30 machines yielding inconsistencies in paper-tape backup and cartridge printout records of voter turnout, County Clerk Paula Sollami-Covello said.
She added that the Statewide Clerks Association had also sent a letter to the attorney general stating its recommendation that an independent study be conducted on the machines and has received no response.
Ms. Rajoppi said that since she only informed a “small sample” of the counties that use the machines of the discrepancies, that more machines probably produced the same errors in this year’s elections.
According to a statement released Thursday, Sequoia has commissioned Kwaidan Consulting of Houston to conduct a review of the software currently used on New Jersey’s voting equipment, with the results to be released simultaneously to Sequoia and Attorney General Milgram. Wyle Laboratories of Huntsville, Ala., or researchers from an unnamed academic institution will conduct similar reviews, according to the statement.
But Ms. Rajoppi said the Sequoia-commissioned studies were unacceptable because the consultants are being paid by Sequoia.
”Their unilateral attitude is unacceptable,” she said.
As of Thursday afternoon COANJ President Michael Dressler said he had not yet received a response from the state attorney general’s office to COANJ’s letter.
”It should be the attorney general that should be conducting an impartial investigation to find out what the problems were,” Mr. Dressler said. “The election process in this country is held in such high esteem. It’s one of our basic rights. We will not tolerate the election process in New Jersey being questioned. It is above reproach.
”We will do whatever we have to do to ensure that the integrity of the election process in the state of New Jersey is intact,” he added.
Michelle Shafer, a spokeswoman for Sequoia, and press officers from the state attorney general’s office did not return phone calls.
