Guest Blogged by John Gideon of VotersUnite.org
First I think congratulations are in order for Humboldt County California Registrar of Voters Carolyn Crnich. Crnich worked with local voters and began the “Humboldt Election Transparency Project”. The Eureka Times-Standard describes the program like this:
Software, created by volunteer Mitch Trachtenberg, then allows viewers to sort the ballots by precinct or race to conduct recounts at their pleasure.
Crnich also deserves a big thank you, along with Trachtenberg, from the voters of Eureka Precinct 1E-45 because if not for the project and Trachtenberg’s software their votes in the recent general election would not have been counted. The Times-Standard explains:
And this “glitch” in the Diebold/Premier GEMS central tabulator software has been known to the company since 2004, or so Crnich learned after exchanging several calls with the vendor…
The paper goes on to report:
“(Eureka’s) Precinct 1E-45 seemed out of kilter,” she said. “The count just wasn’t adding up.”
After double checking all of the precinct’s logs and ballots, Crnich said she discovered a deck of 197 vote-by-mail ballots for the precinct that had been run through the ballot counting optical scanner, but did not seem to appear in the final vote tallies.
After exchanging several calls with Premier Elections Solutions, Crnich said she was told that the software begins counting decks of ballots at zero, and that sometimes when a deck is deleted from the machine due to normal complications, the software also deletes the Deck Zero, which in this case was the vote-by-mail ballots from Precinct 1E-45.
Crnich said she then called the Secretary of State’s Office.
“They were very interested and actually offered great congratulations on this project,” Crnich said.
Crnich said she later learned from the Secretary of State’s Office that two other California counties, Santa Barbara and San Luis Obispo, use the same version of GEMS elections software (version 1.18.19), as well as several entire states, including Maryland.
Apparently both Santa Barbara and San Luis Obispo Counties had been notified of this problem by Diebold/Premier but they neglected to notify both Humboldt County and the California Secretary of State even though the system used in California was a part of the state’s “Top-To-Bottom Review” of voting systems.
“Has this happened in other counties or other states?” he asked. “How can we know?”
That is a very good question. How many states/counties know about this problem? In fact, did the U.S. Election Assistance Commission (EAC) know about this problem? If not; why not? This voting system miscounts votes due to problems that are not generated by the voter. This clearly makes the use of this voting system a violation of the Help America Vote Act of 2002 Section 301(a)(5). If the EAC knew about this problem why was it not referred to the Department of Justice for investigation? If they did not know about the problem will they take action now? I know I will be asking those questions.
The news article concludes:
Kevin Collins, who volunteers with the transparency project and is one of its charter members, said this never would have been uncovered without Crnich’s dedication to transparent elections.
“She deserves a huge amount of credit for devising a system for doing something in Humboldt County that isn’t being done anywhere else, and that’s auditing 100 percent of the ballots,” Collins said.
The uncovered glitch means little for Humboldt County’s election, as it won’t change the outcome of any races and, consequently won’t even require a re-certification of the election’s results, but it has implications that could reverberate throughout the world of elections.
“You just can’t trust a secret program to count this stuff because programers [sic] make mistakes,” Trachtenberg said. “People have been complaining about secret machine counts and the companies have said these folks are nuts. But, the first time (the transparency project) is done in a general election, it comes up with a problem — a problem (Premier Elections Solutions) has known about for four years.”
UPDATE 12/9/08: Much more now here…
OT John, but you might find this interesting.
It appears China is going to Smoot-Hawley us…not good
Publishing actual scans of ballots leaves a system wide open to improper influence of voters, such as by coercion or vote buying. (Marks are easily made that appear legitimate but that are readily recognized as bearing a unique “signature” of a voter, such as in patterns of the ways ovals are filled.) Publishing ballots does, however, make available to all the inter-contest correlations that are generally not made public and that can give advantage to a party that has exclusive access to them. Thus, the extreme transparency of this scheme goes too far and erodes ballot secrecy but does solve an issue with opaque elections that is often ignored.
“Anonymous” –
California already casts nearly 40% of it’s ballots via absentee. While your point about offering the possibility of vote buying/selling via this scheme may be true, it’s certainly dwarfed by the existing vote-by-mail possibilities already in place in CA.
Thus, I’d suggest the transparency offered by the Humboldt project far out-weighs the concerns you note (which is likely one of the reasons that SoS Bowen, who had initially opposed the idea, for the reasons you mention, changed her mind eventually.)
First, get rid of all the touch screens that do no produce a paper ballot, and do the procedure that Crnich developed.
This procedure would certainly work well for mail-in ballots.
RE: Comment #2 by Anonymous:
At one time, a stamp was used to mark ballots. Under or by each election item was a line that you sat the stamp on, then depressed the stamp. It made a fairly uniform mark. I would not mind touch-screens if a ballot (preferably on special paper) were spit out of each touch screen unit, to be verified and counted. The automatic counting of ballots without a paper trail is fraught with problems. Over the past several years, problems with the machines themselves have been numerous. So far there has been no proof of which I am aware that show election fraud with these paperless machines, but I would not be surprised to learn of lots of fraud. However, matching up voter counts with votes counted should always be done. These disappearing and reappearing envelopes, boxes, and the like demonstrate a lack of respect for the voters.
72dawg
Brad has typed his fingers blue against such a thing. People don’t verify the correctness of the printout, whether it’s just a damn receipt or a ballot. Plus, it’s just the machines doing the reporting of the vote and ballots don’t come into play unless and until a recount is ordered. And, as we have seen, those precious ballots have a way of turning up missing when it’s recount time. The chain of custody issues from place to place are a damn nightmare.
Everyone votes on a ballot with a pen, puts it in a clear plastic box in front of everyone who wants to participate in the by hand counting or oversee the by hand counting, and the by hand counting is done right there by these people, recorded by hand, then ballots sealed, put somewhere safe, with strict chain of custody, and then we can have faith in the results.
Period. That’s the only way.
I worked as a pollworker in the Humboldt County Elections. In our precinct, we worked for an hour or two after we closed to reconcile our paper ballot count with the machine’s count. When we found the machine counted 2 ballots we couldn’t find, we counted again and again without finding 2 paper ballots that reflected the machine’s count. I asked and other workers told me they don’t always reconcile. We berated ourselves and left it to the central counters to rectify…
Now I wonder about those two votes…maybe it wasn’t us. Maybe it was the Diebold/Premier machine. It’s odd that workers have found being a couple of ballots off a ho hum abnormality. A couple of votes… how many precincts in the nation? Hmmm. Maybe more lurks beyond the absentee glitch.
This is great news –congrats to all involved –I too vote in California and I kinda thought California got rid of diebold…am I wrong Brad?
While the benefits outweigh the disadvantages at this time, I think Anonymous’ point is well taken. We should fight against the erosion of the secret ballot. There is potential for harm here.
While the Transparency Project is a wonderful thing in terms of keeping track of Diebold, its dangers should be addressed. I would prefer to view this project as a temporary fix, not a permanent solution, due to the possibility of being able to identify a particular voter’s ballot.
Back to paper ballots NOW! DIEBOLD cannot be trusted, as they have proven time and time again, we cannot have rigged elections in this country!
hey joe i think it proves we do have rigged elections
did ya ever notice that after these hackable machines were used to count the vote that exit polling was completely discredited as a proven method of validating an election because the exit poll didnt match what the hackable computer said was the official tabulation?
Brad:
You wrote in COMMENT 3 above: “California already casts nearly 40% of it’s ballots via absentee. While your point about offering the possibility of vote buying/selling via this scheme may be true, it’s certainly dwarfed by the existing vote-by-mail possibilities already in place in CA.”
This type of reasoning is big trouble. For instance, it could be maintained by a similar argument that, because of the lack of transparency of the registration and poll-book processes, we should not be concerning ourselves with the integrity of the counting process.
It is also worth noting that the level of integrity achieved by posting photographs of paper ballots is still based on a kind of “trust us” model and without any effective dispute resolution procedure. Whatever enhancement of integrity or auditability it may provide hinges on the full “chain of custody” through to the website.
Perhaps it’s worth considering if there might be a variant that can reduce reliance on chain of custody while increasing ballot secrecy.
So a person can see their own vote. Doesn’t mean votes are tallied correctly. Lets stop looking to compromise. Lets go back to hand counted paper ballots. Get rid of mail in, scanned and every other possible trick or we will be here at Bradblog in 20 years when we are all grey or dead and democracy will be a nebulous mysterious memory. The election officials here look like they were doing their job in which case this system would work. Otherwise as is the case with computers, all sorts of stuff can happen that we cant see. Yeah we can see the ballots and look at what the computer tells us just like the touch screens. Corrupt officials and companies could still subvert this system big time I think. While we have democrats in office lets go for the gold standard. Hand counted paper ballots.
Naomi Klein is saying the Canadian election system is broken. How could that be? I thought they had the election system that was about perfect. Think I read her post on huffpo or democratic underground. I live near the canadian border so have been interested in their country as I can get their tv. It got boring overnight after harper was elected as ours did. ..esp. the news.Some have said stolen elections kinda’ went world wide.
Kim asked @ 8:
Yes, you’re wrong, unfortunately. In ’04, CA SoS Shelley decertified their touch-screens in the state. In ’06, CA SoS McPherson re-certified them. In ’07 CA SoS Bowen decertified both their touch-screens and their op-scans, but “conditionally” recertified both for use.
In the case of the touch-screens, they could only be used one per precinct to marginally meet HAVA’s disabled voter accessibility standards, and would have to have all of their “paper trails” counted 100% by hand.
The op-scans were allowed for continued use, under slightly beefed security and post-election audit requirements. Humboldt County uses Diebold’s op-scan systems under those standards.
Hope that clarifies the god awful mess.
Still looking forward to Brad’s response to COMMENT 12.
In Humboldt, the ballots were scanned at the Elections office by non-elections staff citizens on an office-style off-the-shelf scanner, by a Linux system, using open source, inspectable code.
The images were then grouped into batches of ten thousand which were zipped into Windows’ style compressed folders. These compressed folders were digitally signed using Gnu Privacy Guard.
Yes, it would be better if each precinct’s independent scan could be done right at the voting precinct, before the ballots are even returned to a central count location.
Thanks for doing so much to keep these stories in the news. Looks like al franken will lose in MN – ironic because he and “air america” have pooh-pooh’d the idea of any voting problems, refusing to allow it to be discussed as a legitimate issue.
Now they will have an election stolen from them in MN – wonder how they feel about these issues now.
Anonymous @ 16 asked for a response to his/her question @ 12, where he/she made two different points.
I didn’t respond originally, because I saw no need to. Your points speak for themselves, and though I don’t necessarily agree with them, I didn’t think it necessary to respond, particularly as any such response gets into the weeds and might take a while.
But since you’ve requested, specifically, that I respond, I’ll try to do so as quickly as I can.
The first question you asked, in my opinion, sets up a false dichotomy by arguing that my lauding of additional transparency and the ability for more citizen oversight (which led to the discovery of 200 uncounted ballots, as reported in the original story) is akin to saying that because we’ve lost transparency in the registration process, it’s unnecessary to count ballots accurately. That, was meant in reponse to my suggestion that any potential for vote buying/selling via secret marks on the ballot, then shown later to someone else online (after sifting through thousands of ballots to find it) is an issue dwarfed by vote buying/selling concerns that already exist via the 40% absentee ballot rate in CA and the additional transparency gained by posting the ballots for citizen review on the web.
I disagree that your suggested parallel is anything close to “a similar argument” as you argued. It’s simply a false dichotomy. And I maintain that the additional transparency offered far out-weighs the concerns you mention. That’s my opinion, of course, and you’re welcome to yours (which I had originally hoped to simply leave stand, so that it could speak for itself if others wished to concur with you).
On your second point, you point out problems with the scanning and posting of ballots on the web, correctly pointing out that it still requires a measure of “trust us” by officials.
Mitch Trachtenberg, one of the developers of the software used in the Humboldt Citizen’s Transparency Project (and the one who noticed the missing ballots) responded w/ additional detail above that may speak to some of your concerns.
Still, as I’m sure he’d admit, it’s decidedly imperfect in its current iteration, will likely see improvement as the program moves forward, and even when “perfected” will still be decidedly imperfect, and far less desirable that fully transparent hand-counting at the precinct before ballots move anywhere, and where secure Chain of Custody is much easier to assure.
But my personal hope here is to improve the possibilities for citizen oversight and transparency, while not causing more harm than good, which I believe this program does at this time.
If you wish to make the argument that such programs may prevent or delay more transparent and secure counting processes (such as hand-counting), you are welcome to make such an argument. It’s not an unreasonable one.
I’ve not gone on the record to say that the Humboldt Project represents what should be done everywhere. But I am on the record lauding the project for the additional transparency it brings, as well as the cooperative efforts for pilot projects between citizens and election officials.
I’d like to see more such projects, including pilot projects for hand-counting as I’ve pushed folks such as LA County Registrar Dean Logan (here I am asking him about it on video) and OH SoS Jennifer Brunner (here’s my interview with her about it) to carry out. Both expressed open-mindedness about the idea.
I’d be delighted to see more HCPB folks push those officials and others to carry out such pilot projects, since I can’t do it all myself. And, the fact is, for hand-counting to become a widely-accepted alternative to our current electronic means of tallying votes, we need to establish a larger body of data to demonstrate that it can work, work well, work accurately and securely on varying ballot styles in varying locations around the country.
One criticism I have of the tactics used by many HCPB advocates out there, is that many are making the same mistake that voting machine companies made by saying “trust us, this method of counting will work and work accurately” without first providing ample data to prove that case.
Those who support hand-counting need to make their case with hard data to back it up, rather than simply making the case by saying (accurately) that “secret vote counting sucks” and (speculatively) that “hand counting is more accurate”.
While I happen to believe the latter, it’s not good enough to say NH does it in some precincts, as do some other places, and oh, yeah, we did it about a hundred years ago everywhere (before long ballots, with initiatives etc. and multi-language ballots, etc.) existed.
We need a lot of data, from a lot of different ballot styles and locations, from a lot of different pilot programs to help make the case to officials this is a viable and necessary alternative.
To that end, I wish those supported hand-counting would get to work on that level and start working with officials to that end!
Again, I can’t do it all! So I hope hundreds and thousands of you out there will take that ball and run with it! I’ll happily have your back in the process!