Indications of Malfeasance and Election Fraud Surface in Anchorage
By Jeanne Devon and Linda Kellen Biegel
Friday’s Assembly work session at the Loussac Library was a chance for Anchorage Assembly members to ask questions of the key players involved in the botched Municipal election of April 3. They’d be speaking to Gwen Matthews, Chair of the Election Commission; Barbara Gruenstein, Municipal Clerk; Jacqueline Duke, Deputy Municipal Clerk; and Dennis Wheeler, Municipal Attorney.
Validity of Votes Cast
First up was Gwen Matthew who spoke of her role checking for the validity of the votes cast. She assured the Assembly that her team was examining everything with all the tools at their disposal. No illegitimate votes would be counted, she said. There were a lot more rejected ballots than last year. Over 500 ballots have been rejected so far, with many more votes to be counted in the coming days.
Ballots Will Be Replicated by Hand
Assemblyman Ernie Hall asked about the ballots that were photocopied for voters to use in precincts that had run out of ballots. “Only an official ballot will run through the AccuVote machine,” Matthew said. “So those votes need to be transferred to a ballot card that runs through the machine.” This means that each photocopied ballot will have its individual selections hand-copied by someone on the Election Commission from the photocopied ballot to an official ballot.
The replicated ballot will then be scanned through the machine. No one asked why the photocopied ballots cannot simply be hand-counted, which would no doubt be faster, and leave much less room for error. These votes do not “need” to be transferred at all.
“I cannot imagine having to transfer by hand, all of those marks, and get them right. I’m a little concerned about how you validate that all those marks are transferred correctly from each one,” Assembly member Harriet Drummond commented.
“I think Jacqueline, with the Municipal Clerk could answer that question better than I,” said Matthew. Jacqueline Duke had no answer.
The answer is that in an emergency such as the one we had last week, photocopied ballots should be allowed in all cases. Many polling places, including schools and churches, have copy machines right on the premises. Many who tried to vote were turned away, sometimes in the parking lot, and told they would be unable to vote at their regular polling place. They then went to another (sometimes two or three) other locations before being permitted to vote, or giving up along the way. Some ran out of time. Nobody should be turned away from the polls, ever. This is something the Assembly must address.
The original ballots filled out by voters are kept for only 30 days after the election is certified, and then they are destroyed.
Chris Birch Asks the Critical Question
There were other more specific and detailed questions from every Assembly member, and all the while the Chair of the Election Commission assured them, and us, that the vote was in good hands. Nobody would be missed. Every vote would be painstakingly scrutinized. She wanted the vote to be “100% accurate.” And she might have seemed credible if it hadn’t been for Chris Birch, who finally asked the question that has needed to be asked since 2004.
Matthew’s answer to Birch’s question flew under the radar of many there, but for anyone who has been following issues of Alaska election integrity, it was jaw-dropping, and seriously calls into question Matthew’s believability and integrity.
Birch - I just had a question kind of on the functionality of the hardware – the Diebold scanning units. As I recall, they were secured by the state in about 2002. It kind of went through a statewide process, through the Lt. Governor’s office. I know when I personally voted, it kicked out my ballot a couple times. And I know from having familiarity with laptops, and computers and stuff, you know… things from 2002, 10 or 11 years later sometimes the technology can get pretty outdated. So, I just wanted to know, as a Commission, do you have any engagement with the hardware side of this, the functionality of the scanning equipment?
Matthew - Well, I happen to have been Chairman of the AccuVote testing board for this, (laughs) so I do. Those are amazing machines – utterly amazing. You… they print out everything. It is impossible for them to go haywire. You can put it in any way – front, back – sometimes if it doesn’t read you just turn it over and put it in. We have tested each and every one of those machines, every year, and um, we run tapes, and double-check and just like at the commission, two people work at the same time. They are highly accurate. I think that I could almost say that they’re totally accurate. I’ve never found a discrepency.
Birch - Well, that’s good to understand. What is that bounce-back – you know, when you feed it in and it kind of kicks back, and that seemed to be a recurring issue.
Matthew - Occasionally there are marks on the side, in the margin… those little dots that are on the side of the margin is what they read, and um, sometimes it’s bent or torn, or there’s a stray mark that’ll kick it back. But it also records… it also tells you on the front of the machine what is the problem. Like perhaps it’s an over-voted race, perhaps they voted twice in the same candidacy, or under-voted and it will tell you – um – it’s a very smart machine.
Brad Friedman, award-winning investigative journalist and noted election integrity expert, responded to Ms. Matthew’s description of the Diebold AccuVote scanners. “Anybody who says it’s “impossible” for these machines to fail either knows absolutely nothing about how these machines work or is simply lying. I don’t know Ms. Matthew, so I won’t guess which one of those two things is true, but I promise you one of them is.”
Let’s have a little recap of what Matthew described as a “very smart machine” to fully understand why her answer is literally unbelievable. The state began using Diebold AccuVote scanners in 2002.
“A Very Smart Machine”
After the 2004 election, the Division of Elections posted the vote totals out on their website. They posted both a “statewide summary,” and the breakdown of the House district votes. But when vote totals of individual districts were added up, the total didn’t match the statewide summary total. Here are a couple examples:
- George Bush received 190,889 votes according to the statewide summary. But, if you added up all the House district votes, he received 292,268 – a difference of 101,379 votes.
- Lisa Murkowski received 226,992 votes according to the statewide summary, but only 149,446 if you added up all the House district votes – a difference of 77,546 votes.
- Precincts across Alaska were reporting voter turnout exceeding 200%.
The Democratic Party (ADP) asked the Division of Elections (DoE) to explain what had happened and what the correct totals were. They couldn’t. Then, they asked if the Division of Elections could simply turn over the data. They refused.
By the spring of 2006, David Shoup, an attorney specializing in civil litigations representing the ADP, asked for the DoE to turn over the central tabulator file which is the device that actually tabulates all the vote totals from the GEMS (Global Election Management System) database. This is public information, and a matter of public record. We the people already own this information, we just wanted a copy of it.
The State refused stating the information was proprietary to Diebold. When it was pointed out that all the information they said was proprietary to Diebold was readily available on the internet, they changed their reason. Citing a post-9/11 code, they said they were withholding the information because it posed a “security risk.” A judge wasn’t buying this new reason, and the State couldn’t adequately explain how turning over vote data was a security risk. Finally, the ADP sued the state in the spring of 2006, and the state was forced to turn over the information.
They discovered that when individuals entered the totals from the voting machines into the GEMS database, there was no code or key indicating who had entered the votes, or where they were coming from. The username in each case was “admin” and the password was “password.” The votes could never be reconciled, and we would never know why the numbers didn’t add up. And it got worse.
“One of the things we discovered ,” Shoup said to a shocked studio audience of the TV show Moore Up North, “was that you could hack into the GEMS database from a phone in Cleveland, and insert votes into it for an Alaska election.”
These are the very same ballot scanners that we use today, and that were used in the 2012 Municipal election.
Machines “Routinely Miscount Votes”
Friedman had more to say about Matthew’s assertions that the AccuVote was virtually infallible.
“To say they are “totally accurate” is simply astounding to hear from an election official who claims to understand these machines. Diebold themselves have admitted, time and again, that the machines routinely miscount votes, drop entire sets of ballots from the count without notifying the operator, can be rigged in such a way that pre-election testing won’t reveal the tampering, and can have their entire “audit logs” deleted (to hide any such gaming) without notice to the administrator. Worse, in addition to all of the above, they can also be gamed by insiders in such a way that no testing in the world would reveal.”
There are many reports of problems with Diebold machines, and in particular the AccuVote scanner and software. An optical scan system using the same technology as ours recently declared the wrong winners in Palm Beach county Florida in several races. Those races were ultimately settled by hand count.
During the Question-And-Answer period with the Municipal Clerk’s Office, it was clear that the member of the office in charge of the election, Deputy Clerk Jacqueline Duke was quite defensive. Each Assembly Member who attempted to establish a clear timeline of events for April 3rd seemed to end up frustrated. The session generated more questions than it answered. It was especially confusing when they attempted to determine who was giving the marching orders that “troubleshooters” were relaying to the voters. Ms. Duke claimed that she did not tell them to send voters away…or to absentee ballot polls UAA or the Airport…when the ballots ran out at 53 precincts. However, Lindsey Spinelli, assistant to Mayor Sullivan, must have felt comfortable enough with that information to post it on the Mayor’s Facebook Page at 6:44 pm:
And then to post an hour later that they should wait for ballots:
When asked by Assemblywoman Harriet Drummond why the Mayor’s Facebook page posted those instructions, Ms. Duke had no explanation.
In Jacqueline Duke’s sometimes feisty exchange with members of the Assembly – in particular Bill Starr and Debbie Ossiander - she made one thing quite clear…she personally conducted the training sessions for election night poll workers this year.
Duke’s training of the election night poll workers made election worker Wendy Isbell’s revelation regarding her poll worker training even more shocking.
In an interview with The Mudflats, Isbell said that she had attended one of the training sessions for the 2012 Municipal election workers, and she has worked as one for four years. She was well aware of the instructions in the election book, given to every poll worker, especially the one regarding the seal in front of the memory card of each Diebold voting machine (Accu-Vote Opening Instructions on page 7, Step 5 #2):
Isbell claims that the verbal instructions coming from the trainer, now determined to be Jacqueline Duke (by Duke herself), were quite different. According to Isbell, Duke told the trainees at the session she attended that (paraphrasing) ‘if the seal over the memory card was broken when they picked up the Accu-Vote machine they should not be concerned. The seals break all the time.’ Further, Isbell states that the seal on the Diebold machine for her precinct was, in fact, broken.
Other than the fact that the verbal instructions Wendy Isbell alleges directly contradict the 2012 Elections Manual, such lax security has been known in other states to leave the machines open to all kinds of mischief.
It is also interesting to note that in the diagram of the Accu-Vote machine in the MOA Elections Manual, the location of the important memory card seal is nowhere to be found:
When asked to comment on the broken seals, Brad Friedman stated emphatically
“If and when any seal on these machines are broken they are to be immediately taken out of service and quarantined for forensic investigation. If that is not already the law in AK or Anchorage, then it is a grave security hole in the law.
Anyone who instructs someone to not report a broken seal and use such a machine anyway should be investigated for malfeasance, misfeasance and/or criminal election fraud.”
None of the above will be visible to anybody unless a full hand count of the paper ballots is carried out. And such a public hand count this late after an election is only as reliable as the chain of custody of the ballots since election night.
It’s not only actual election failures and Diebold admissions that instruct us of the many ways these machines fail, but also independent study after independent study by world-class computer science and security experts in states like CA, OH and CO who have all found the very same thing over and over in each and every test.
To get a full understanding of why security seals must be intact, take a look at the final scene of HBO’s Emmy-nominated 2006 documentary Hacking Democracy. Alaskans may recognize the voting machine being tested. It’s our very own Diebold AccuVote scanner – the one that Election Commissioner Gwen Matthew called “utterly amazing” and “highly accurate.” The security seal is the piece of metal that holds the memory card in the vote scanner. A broken seal means there is no guarantee the card stayed in the machine untouched, before arriving at the polls. And this is what can happen.
The issue of how many ballots were printed, and location of ballots is a critical piece to this puzzle, but it is not the only thing that needs to be investigated. Nothing else really matters if we are faced with elections where there is no way of proving that the results are accurate. Anchorage should demand:
1) An independent investigation of the election to determine why 53 of 121 polling places ran out of ballots, where the ballots were stored, why they were not at polling locations, why the troubleshooters were late in responding, and many other questions concerning the logistics of the vote.
2) A full audit and recount by hand of paper ballots, and the chain of custody of those ballots.
3) Clarification that photocopied ballots are acceptable, and should be the first step taken in such an emergency. It is never acceptable to turn voters away from their polling place.
4) Based on facts known, the Assembly needs to demand that we no longer use the faulty, and easily compromised equipment we are currently using. The most accurate and honest way to count a vote is the old-fashioned way – by hand. Anchorage (and Alaska as a state) is perfectly capable of performing an open and transparent hand count of ballots. It is cheaper, more accurate, and allows the public to have confidence in the most critical process of our democracy.
5) An investigation of the Clerk’s Office and the allegation that poll workers were instructed to ignore machines with broken seals.
5) A new election.
There is no way of knowing how many voters were disenfranchised at the polls, how many gave up and went home, how many were told by friends not to bother, how many ran out of time racing from polling place to polling place, etc. Even a hand recount is only as good as the chain of custody of the ballots this long after the election. Easily tampered-with machines that arrived at polling places with broken seals are enough to render any results meaningless.
It is critical for conscientious citizens to demand an accurate and thorough vote count, and to question any result that relies on this equipment. The Democratic Party was right to question it in 2004, and 2006. Joe Miller was right to question it in 2010. The Anchorage Assembly is right to question it in 2012.
The Municipality is in the position to be able to demand answers, and to change the way we do business on election day. This change is far easier to make on the Municipal level than on the state level. It is absolutely critical to contact the Assembly (firstname.lastname@example.org), share your concerns with them and tell them what you want them to do. This isn’t about who won or lost. Democracy only works when it works for everyone.