Guest Blogged by Dori Smith of Talk Nation Radio
Failures in optical-scan voting systems made by Diebold Inc., as recently revealed in the state of Florida, might have occurred elsewhere across the country. In fact, any state where Diebold’s optical-scan voting machines are in use might be having the problem, and they may not have been made aware of it — either by the company, or the U.S. Elections Assistance Commission (EAC) which is mandated by law to be a “clearinghouse” for such information.
A faulty connector-pin on Florida’s Diebold optical-scan systems — affecting some 4.5% of the state’s machines, and as many as “one in 10 during the November 2006 election” in some Florida counties, according to a recent investigative report by the Daytona Beach News-Journal — has been identified by the company as a “J40 connector”.
It now appears that Connecticut’s voting machines, made by the same company, are being affected by the same problem.
“This is Connecticut, not Volusia County,” President of LHS Associates, Connecticut’s vendor for Diebold AccuVote OS machines, told me during a recent, bizarre late-night phone call. I felt pretty certain he was right, though his call came in to my home at 1:30 in the morning, so I was still a bit groggy.
LHS’s John Silvestro had picked a strange time to return my call from a week earlier. We’d been following LHS’s role in addressing memory card failures during Connecticut elections since 2006 on Talk Nation Radio, and I wanted to know how many memory cards LHS had to replace before, during, and after the 2007 election.
At that hour, I declined his offer to “do the interview anyway,” though we did touch on the basics of my inquiry. Did Registrars contact LHS to request replacements for “blown” memory cards as pre-election tests were being run? If so, a study currently underway at the University of Connecticut might not have been given the correct data to determine the true number of failing cards.
Silvestro asked me, “What would you say if I said Connecticut’s failure rate was less than one percent?”
Perplexed, I told him I would say “OK” and we agreed to discuss the matter at a later time when he might have access to his records, as he had said he didn’t have numbers in front of him at that late hour.
He wouldn’t be the first LHS official to exhibit bizarre behavior. After jarringly inappropriate comments left at The BRAD BLOG several months ago, one such official was told he was no longer welcome to work in the state of Connecticut.
Confusion in Connecticut
Diebold, and their New England distributor, LHS, failed to inform Connecticut officials about the J40 problem as found in the state of Florida where the hardware in the voting systems is the same, according to experts we spoke to. If the same problem is occurring in Connecticut, the company’s decision to keep quiet about it may have contributed to widespread confusion over just what may have caused recent voting machine failures.
Diebold’s Premier Election Systems has not yet responded to our calls seeking details.
Memory card failures can cause machines to fail outright, but they may also cause strange shifts in the numbers. Such a scenario may explain variations in results in recent elections in Connecticut. Audits of voting systems used during the 2006 election found significant variations in machine versus hand recounts and audits, but the problem was kept quiet, according to a report [pdf] by George Barnett of election watchdog, True Vote Connecticut. Similar fluctuating totals during the 2007 election have left town officials puzzled, as a number of memory card failures have been reported.
Diebold, whose elections division has now been renamed “Premier Election Solutions”, had been tracking similar data in Florida and knew about the memory card problems for some time prior to the 2007 election. Yet, neither they, nor the U.S. Elections Assistance Commission (EAC), established in part as a “clearinghouse” for such information, have warned states using the same equipment about the confirmed problems.
It is a story that smacks of cover-up, possible collusion between manufacturers and any number of vendors, and potentially criminal concealment of problems with voting machine technology in states which use Diebold’s optical-scan voting machines.
News-Journal reporter M.C. Moewe has been tracking this story in Florida, where she’s reported high failure rates for memory cards and where Diebold ultimately replaced cards just prior to the 2007 election in the wake of her report.
Following her initial story, Moewe published an interview on November 12th with Christopher D. Riggall, the communications manger for Diebold Inc.’s Premier Election Solutions. He said engineers had been able to identify faulty “J40 connectors” in their optical-scan systems, and he admitted Diebold might have to replace the parts in Florida’s machines.
More disturbingly, Riggall told Moewe the company would address problems in other jurisdictions only “on a case by case basis”. He said they would only respond after customer complaints had “identified…a concern.” In other words, if the states using the scanners didn’t ask, Diebold wouldn’t tell.
State Officials Left in the Dark
As the EAC, Diebold and LHS had all failed to notify jurisdictions outside of Florida of the possible problems, the exchange between Moewe and Riggall was a revelation for Connecticut’s Deputy Secretary of State Lesley Mara. When I asked Mara if Diebold had ever notified her office about potential problems with memory cards on the machines her response was a forceful, “no!”
She and Democratic Secretary of State Susan Bysiewicz had recently commissioned a top down review of memory cards from the University of Connecticut’s Voting Research Team, as led by computer scientist Alexander Shvartsman. The team is under contract to provide voting machine security assessments for Connecticut.
The Secretary of the State’s request for research into memory cards was partly due to Talk Nation Radio’s 2006 reports about card failures that came up during last year’s Congressional race. We had conducted interviews with three LHS staff members who had worked the polls last year in Connecticut with back up memory cards in their pockets and back up voting machines in the trunks of their cars. That, despite the fact that the state’s voting machine security protocols, under Connecticut Statute 9, prevents them from even touching the machines.
One by one, the employees filled my radio listeners in with details on how they’d address voting systems failures by opening machines to make memory card replacements on the spot. An illegal memory card switch — swapping out the sensitive cards which store voting data and which can also be used to insert malicious code into the systems — was made by an LHS staff member in Montville, CT during the recount of the 2nd Congressional District.
The election moderator in Montville, Lorraine Elliot, told us an LHS staff member used a card that had been hand carried to the polls. From her description, it seems Elliot may have come across an incident which sounds as if it involved the same kind of problems with pin configurations that has been identified in Florida’s machines.
She told us that she “had set up a new machine and we put the memory card in and it didn’t work. It didn’t print anything out. So we took that one out and put another one in and the same thing with the second one. It didn’t print. I don’t know how the memory card actually works but the little prongs, I guess, sometimes they don’t actually touch the inside? So the third card that we got from LHS worked fine.”
As they prepared security protocols for 2007, officials from the Secretary of the State’s office and the UConn Voting Research Team team visited LHS Associates in Methuen, MA. There, according to Deputy Secretary of State Lesley Mara, LHS staff members showed them the company’s rapid response center where LHS staff members rush to replace failing memory cards as elections continue. They rush the newly programmed cards out to the polls.
Some states allow the practice. Yet, the same practice may have resulted in an artificially low rate of reported memory card failures. Hard data from the company is difficult to come by. Moewe’s initial story for the News-Journal detailed Diebold’s refusal to release data they had collected on memory card failures in Florida, citing it as “proprietary business information.”
My findings during the 2006 election were also revealing. LHS was in position to keep memory card failures quiet, and there has been conflicting information about whether or not they had informed Secretary of the State Bysiewicz when cards had failed.
Our preliminary information on the 2007 election is that the UConn team has identified memory card failures during their initial testing. The failures occurred in consistent ways that the team is said to be looking into. They have yet to produce a report on the memory card research, as cards were still arriving at the lab to be tested as of last Friday afternoon, according to Deputy Secretary Mara.
After Moewe’s revelations in the Daytona Beach paper, we phoned Deputy Secretary Mara to let her know that Florida’s problem could be affecting Connecticut machines as well.
It’s a good thing we did, since neither the vendors, nor the EAC had done so.
The EAC, as reported multiple times by The BRAD BLOG (eg. here and here) has refused to send out such alerts themselves. Two reports by the Government Accountability Office, in 2005 and again in 2006, have been highly critical of the federal agency for their “continued absence of a national clearinghouse for voting system problems” as mandated by the 2002 Help America Vote Act (HAVA).
Per Mara’s request, I remained on the line while sending her the transcript of Moewe’s interview with Diebold’s spokesperson admitting that the company would only respond after customer complaints of problems. She immediately forwarded it to UConn’s Professor Shvartsman, who quick replied back to say he was aware of “noise about memory connectors on the Internet and non trivial faulty ones,” she said.
Mara explained that “noise” was Shvartsman’s way of describing the type of issues we were talking about. She added that, at some point, LHS mentioned a problem to the team having to do with “bent pins.”
It was the Deputy Secretary’s impression that Shvartsman is now aware that there could be a problem with Connecticut’s machines. She said she’d consider writing to Diebold immediately in order to ask company executives why they had failed to inform Connecticut officials about the problems they had found in Florida.
The State Election Enforcement Commission is already investigating a complaint we filed on October 30, 2007. It has been assigned as docket number 2007-385. The complaint asks Connecticut officials to investigate LHS Associates and the handling of memory cards, as well as to provide research into Connecticut’s voting machine security protocols. The complaint points out that the vendor and/or manufacturer have a strong motive to conceal such problems.
Meanwhile we are continuing our own investigation into several reports of memory card failures. We have done just a handful of interviews so far. Deputy Secretary Mara said she knew of two card failures during the election.
LHS Official Barred from Work in CT
LHS has proven to be less than agreeable to concerns about the voting systems they distribute across New England. Recently, their Director of Sales and Marketing, Ken Hajjar left a surprising comment in reply to a BRAD BLOG story about Diebold. Hajjar’s impolitic comment began “Dear Brad, you are totally full of shit.” He went on to charge, “You have no idea how elections are conducted and how many safeguards are in place, including human oversight,” before referring to Election Integrity advocates as “paranoid” and “deluded”.
Unfortunately, the comments were not out of character from previous interviews I’d done with him.
Hajjar’s underlying resentment cropped up during my 2006 interview with him, when he charged that my questions were, “an indication of just how dangerous the Internet has become”. And when we touched on the UConn team’s reports he blurted out, “I’ve read every report in America, some of which are frankly superfluous”.
It was just plain impossible to follow all of Hajjar’s various contradictory, and often confusing statements. For example, Hajjar admitted he had seen Connecticut’s voting machine security protocols, which had been faxed to LHS in MA. He said, “I was given one sheet of paper which was the Secretary of State’s rules and I was just told ‘don’t touch anything just answer questions.’ So I don’t have that with me, and I’m not even sure what I did with it. I might have just thrown it away once I got through”.
A few minutes later, when asked who would be on site, to either repair or replace the machines if there were problems during elections, Hajjar said, “Either a representative of the town or a representative of the vendor.” He then explained, in some detail, about how they would open machines, in apparent violation of state law, to change memory cards during the elections.
Other LHS staff members we spoke with, including Mike Carlson and Tom Burge, provided similar comments. They said they would open machines up during an election and swap memory cards as needed. This is illegal under Connecticut law and Deputy Secretary Mara told us she has since informed LHS that such actions were in violation of Connecticut election laws.
In 2006, as Hajjar argued in favor of their policy to change cards during elections, I asked him about about the laws which govern chain of custody issues. His response: “I mean, I don’t pay attention to every little law. It’s just, it’s up to the Registrars. All we are is a support organization on Election Day”.
He said he had three memory cards in the trunk of his car and, in the event they had to be used, the chain of custody issues wouldn’t matter since, “once you run the [pre-election] test deck through, you’re golden”.
“We would have a whole bunch of machines in the trunk in the car and we hope the phone doesn’t ring, but if it does somebody tells us where to go, we replace the machine and then we go on our merry way,” he declared.
Hajjar also argued that the first use of the new machines in 2006 “was perfect”. But LHS President John Silvestro has since admitted that there were failures of memory cards in the state that year and he has offered his interpretation that his staff was confused between laws in Connecticut versus other states where the company provides services.
As I struggled to pin Hajjar down on these matters, he cried, “Madam you’re the reporter I’m the machine guy! I don’t tell you how your presses run, don’t try to get me on semantics. The machine either sees or does not see. If it sees it counts. If it doesn’t see it doesn’t count.”
Ken Hajjar’s bluntness finally caught up with him this year. Following his less-than-appropriate comments at The BRAD BLOG, he has informed by Connecticut officials that he is no longer welcome to work in the state.
Just as well. His patience for dealing with these issues was clearly running thin long ago. “You cannot…you wanna go into the booth and tell the voter how to fill in the oval?,” he asked me in 2006. “You wanna slap ’em across the face if they don’t fill it in right?”
It was his boss, Silvestro, who had called me at 1:30am recently. The one who had informed me that “This is Connecticut, not Volusia County.”
He was mostly right.
In reading this story I was struck by the numerous references to swapping out memory cards on election day. It appears to be standard operating procedure. This raises two serious problems:
1. The memory cards contain both the ballot definitions which instruct the optical scan machine which candidate to credit the mark it reads when scanning the ballot, and it then stores those read marks (votes) in a database for later upload into GEMS, the central election management system.
If the memory card is removed during the election after ballots have been scanned and votes credited to candidates, those votes are now lost. The only way to recover those votes would be to re-scan the paper ballots. I did not read that was the practice when memory cards failed and were removed. If there was no reporting to the Registrar exactly which batch of ballots data was lost there would be no necessity for re-scanning the ballots. That is one of the problems when vendor personnel replace the card from stock they have on hand. It is in their interest to not report failures, which it appears was the case.
2. The other problem is the introduction of memory cards into an optical scan machine during use. As we know from the Hursti I attack (and subsequently confirmed by the State of California and others), that memory card could have been preloaded with votes that could effect the outcome, it also could have an altered ballot definition that would credit scanned votes to the wrong candidate. If there is no random audit of sufficient size to catch such manipulations, or no reporting of such swap-outs to the Registrar, it would be relatively easy for elections to be compromised.
In the real world of elections both the vendors and the election officials have a vested interest in claiming there were no problems, or the problems were minor but did not effect the outcome of any race. We have seen that spin repeated hundreds of times throughout the country over the last decade.
It has only been because of citizen activists who have volunteered their time and money that the lid has been removed from the septic tank that is our current electronic voting processes. If not for the activists, the only time election officials and vendors admit there is a stench is when the toilets back up. It is time to bring in the pumps and remove these waste products from our elections. Spraying air freshener will no longer work in reassuring the voters of this country.
Jody, good points. If you read through my interview with Deputy Secretary of State Mara you find that Connecticut rejected a service LHS offers to provide memory cards and back up machines during elections. Connecticut rejected that service, and when I first began looking into this back in 2006 it was because of a report from the UConn team on vulnerability to the machines whenever chain of custody were broken over the cards.
It is my understanding that the UConn Voting Research Team is looking into the programming of the cards for any errors, checking the GEMS. It seems the voting machine security protocols are key and that is indeed something activists have been trying to get across for years. Thanks to the work of M.C. Moewe in Florida the researchers can add a hardware component to their study and can see if testing memory cards on a variety of machines may be in order.
This page contains an interview with Michael J. Fischer, a Yale computer scientist and President of True Vote Connecticut have been monitoring this topic and pressing for acute control over the machines and tight security protocols. You can find my interviews with him easily.
It’s an ongoing process.
Dear Brad,
Thanks for this superb work and the calm, cool and collected approach.
My idea may be redunant, but besides chaotic malfunctions, faulty election machines habituate local election officials to the “normality” of highly irregular procedures, such as switching out memory cards and manual manipulation.
Just a thought.
Diebold’s New England vendor, LHS Associates, has a ‘rapid response’ service. During elections poll workers and officials would let them know of a failed memory card. They would program a ‘new’ one and provide it immediately. That is as I understand it, the standard operating procedure in some states.
To clarify, according to Deputy Secretary of the State Lesley Mara Connecticut refused that service.
During 2006 when I contacted them to discuss the memory card swaps made during the recount of the 2nd district race, the Secretary of State and Deputy Mara said they were not aware that this was happening. They said they were surprised by what the LHS staff members told me and it became a matter of inquiry. A meeting was held with LHS President John Silvestro.
The state received my formal complaint on this matter. I continue to be interested and actively researching.
Dori Smith Jan 10, 2008
I have just found this blog and have favorited it. I would ask anyone interested in supplimental information to please visit
BlackBoxVoting.org
This includes details on Mr. Hajjar’s past narcotics convictions.
Best Regards!
I am watching with great interest and concern from Canada. We all know this is a global issue, which will in time affect other citizens.
Your country( and the world) cannot afford one more election to be lost due to corruption and treason. Are there no avenues for a class action suit( from the citizenry) against these blatant criminals? Your court system is your last defense against these corporate criminals.
The secrecy involved in electronic voting systems can be tolerated no longer.
K Mitchell B.C, Canada