Can we audit or recount by machine, rather than hand-counting? My conditional answer remains a strong NO. However, as we have discussed before it is quite possible in theory to develop voting machines or auxiliary scanners with capabilities that can greatly reduce the cost, while increasing the integrity of audits, and increasing the confidence in elections.
Now a team in Humboldt County, CA is providing a demonstration of technology and procedures that can provide all these benefits. The Humboldt County Election Transparency Project:
This type of system holds great promise for Connecticut and any other state with optical scan voting. We will address the possibilities for reducing the costs and increasing the integrity of our elections. Today we will address the Humboldt project as a demonstration of public auditing and recounting:
Our Project aims to provide images of each counted ballot, so that any person or organization wishing to do an independent count will have access to a complete set of ballot images.
They also note as we have pointed out many times:
Voting systems such as that used in Humboldt County, which use optically scanned paper ballots, do leave an audit trail of all cast ballots. This audit trail becomes far more valuable if it is actually used to verify the count.
Here is the basic plan:
- Scan all ballots into a special scanner shortly after the election.
- The Scanner uniquely stamps a number on each ballot.
- Images are scanned into .jpeg images.
- Ballots are posted to the Web
- A publicly available open-source program to read the images and independently count the ballots.
So far there are several articles posted on their blog:
- General overview
- A critical question addressed in a post of its own: Can’t Digital Images Be Faked?
- Ballot Browsing Program
Hats off to those involved in this project and for their generosity:
The project was initiated by Humboldt County County Clerk and Registrar of Voters Carolyn Crnich, former Green party Presidential candidate David Cobb, local citizens Kevin Collins and Parke Bostrom, and Tom Pinto of the Humboldt County District Attorney’s office. The custom scanning and counting software, along with the technical procedure used by the project, has been developed pro bono by Humboldt-based freelance programmer Mitch Trachtenberg, who may be contacted at mjtrac@gmail.com. The scanning software will use the SANE (“Scanner Access Now Easy”) protocol, and is built on top of the SANE command line program “scanimage”. These programs run on top of the Debian Linux Etch release, downloaded from a debian mirror in June 2008. All software used is available for download free of charge.
I understand from the project team that the scanner can scan about 5000 ballots per hour and cost $26,000 including maintenance.
Could such a system help Connecticut? We will leave that for another post on another day.
In all the discussions of increased audits in Connecticut, we need to take the time to assure that we don’t fall into the same pitfalls other states have fallen victim to.
In the selection of the technology we eventually adopted, Connecticut carefully avoided many of the existing vulnerabillities demonstrated nationwide. Many states are following Connecticuts lead with limited election audits.
We quickly determined that there be no communication between tabulators, or between tabulators and any central counting facility. While this greatly complicated the end-night tally process, it clearly avoided the many documented vulnerabilities of the GEM software. Hopefully, at some point we may see the necessary improvements in security, adequate to allow the state and cities to take advantage of this valuable tool, but not yet.
We have upgraded our security precautions after each election, and will continue this process. Exterior ports are covered with recorded seals. Hopefully, future models will not include these ports at all. Election officials in the company of voting technology experts have toured the LHS facilities, making and assuring procedural changes. The SOTS has instituted independant pre and post-election testing of memory cards. This testing is in tandem with the Audit program established in 2007. The Audit program has been, and will be subject to further changes including many of those reccommended in the last Legislative session; at a minimum, resulting in an independant board. which was the reccommendation of all parties.
Chain of custody of the programming cards has been substantially tightened. Further procedural changes will be made as we gain experience with the optical scan voting system. And,of course, the cards themselves once tested locally and set for the election are sealed in the tabulator. sealed in the transport bag with its own security procedures, and then eventually locked into the tabulator box, effectively blocking all access to any ports on the tabulator including the programming card. We may argue the extent of this testing of the program cards, but the groundwork has been set for independantly verifiable accuracy of the ‘equipment’.
The tabulators have a proven record of reliability. Many of the reported problems we have encountered here in Connecticut have been related to dampness of the ballots, unique problems with the printing of two sided ballots and the subsequent positioning of the ‘circles’, and bleed-through which relates to the pens used and ballot arrangement. Many of the reported ‘breakdowns’ were due to jammed ballots or ‘dirty’ or dusty scanningunits, which Moderators and Registrars wer reluctant to address during the election, instead switshing to a new unit. As the ‘comfort level’ increases, many of these perceived problems will be dealt with at the polls.
Peculiar to Connecticut is our insistance on local management of elections. No other state has this local control of voter registry lists, election management, and ballot preservation (Storage). In tandem with this is the need to secure all election materials, especially the ballots in the event of a close vote recount of a court challenge to the results of any election. This is why central counting or electronic transmittal of results have never been a factor in Connecticut.
Deadline and timetables for the Audits have been subject to the required sequestering of all election material including ballots for the two weeks after the declaration of winners or of the recounts. We can squeeze these dates a little but the need to secure the ballots until after the legal requirements remains. Mentioned on this site is perhaps a more interesting question of post Audit challenges in the event of a tabulator found to have malfunctioned, affecting the race in question and possibly other races on the ballot. Under current regulations, the SOTS can order further counting if they fell it is warranted, but how does this interface with challenge deadlines?
This leads directly to the conduct of Audits. We agree with the need for and value of a strong and independant Audit program. One of the proposals being presented would require either the transport of ballots to some central counting location, or the use of some type of local independant auditors.
Either of these options creates serious problems. We have repeatedly seen (Most recently in the August Florida temporary loss of 18,000 ballots) that the most serious problem with recounts have been lost or misplaced ballots. As many precautions and procedures that they have taken, the transport problem has remained a constant point of failure. Whatever eventual system we use, the sanctity and securitry of the ballots must remain paramount. From what I have read here and elsewhere, the transport of ballots is precarious at best.
Efforts of post election study of the ballots, with current Records Retention requirements for 13 or 22 month retention of the ballots and related election materials, must take into consideration the varying storage methods used across the state. Increased space requirements for storage of several elections worth of ballots will force many towns and cities to use off site storage. Whether the State can afford any relief with central storage (and perhaps some form of access) is open to question. Any post election ballot review plan must be specifically limited to technical study, rather than never ending attempts to challenge the results of an election. While the benefits of such a program are real, the threats of misuse are also everpresent.
There appears to be a sea change occuring in election management. Prior to the optical scan system, the vast majority of effort went into pre election verification, setup, and election day security. The existence of the individual votes in the form of the paper ballots now raises the opportunity and requirement for post election review of elections, and the potential of never ending hand recounts and ballot review.
Where all this will lead is unknown, but we need to remember that every election must have an end, especially when they occur on the heels of the last (Over 15 times a year in referrendum towns).
I agree with some of your points however, let me point out:
Most New England states have local storage of ballots and local control of elections very similar to Connecticut.
Yes there are dangers in storage and transport of ballots to central locations. Just as there are dangers in local storage as well. I would point out that in many towns in Connecticut, the ballots are stored in vaults or cabinets to which each registrar has a key. In some towns each staff member of the registrars’ office has access to that key. I doubt we would accept evidence in a criminal trial for which even one person had single key access for days. Other states guard their ballots in vaults with human guards and video surveillance.
I have been advised not to ever be alone with ballots as it would open me to charges of tampering if there were ever a subsequent problem with the ballots (some of our observers have been left alone with the ballots by election officials, I will be advising them not to be left in that position) — our registrars’ and their staffs should also not take that risk.
Having extra seals on the machines was a very good idea. Unfortunately, the Moderators’ report forms do not require they be tracked. What good are tamper evident seals if we don’t track and record them? As to sealing ballots, they need to be sealed until all possibility of the need to investigate election integrity as passed. Many people disagree how long they need to be sealed but many claim state law limits the requirement of a seal 14 days after the election (before the audits start).
Since UConn has not given the report of the February Audit to the Secretary of the State as of this date, the ballots, memory cards, and tabulators should still be sealed in case UConn reports an issue to be investigated.
Speaking of the Pre and Post Election Memory Card tests, while it is a good idea, the tests so far have shown that procedures are not being followed and there continues to be a problem with card reliability, and the tests are not really a random sample.
Many many other states do their audits much sooner. They work around all the other requirements to do recounts, audits, and contested elections so all are done soon after the election. Just because it has never been done in Connecticut does not mean it is not possible.
I would love to see a copy of those improved procedures for handling memory cards. When were they implemented? What were the changes in the procedures?
Yes the lever machines were more convenient to election officials since they had not paper record, there was no opportunity to verify the results. Every election should have an end – audits should happen quickly after the election – if our tabulators work, they are programmed correctly, the ballots are appropriately preserved, and the election officials use proven techniques for counting ballots…all elections will end quickly with the result reflecting the intention of the voters.
I share your support of the need a lot of changes along the line of those proposed last year, including an Independent Audit Board of some type. Lets hope that it indeed does become the law in the next session.