Ohio Paper Trail Legislation
The Ohio Senate passed a bill that will allow electronic voting machines to be put in place by the November election, according to this story in the Cleveland Plain-Dealer. An earlier Plain-Dealer story, available here, reported that the bill would require a contemporaneous paper replica of the electronic ballot (erroneously referred to as "paper receipts") to be implemented by 2006. In fact,the legislation specifically states that machines "shall not provide to a voter any type of receipt or voter confirmation that the voter legally may retain after leaving the polling place." See here for the Legislative Services Committee's analysis of the legislation, and here for the text of the bill as passed by the Senate.
Of particular note is the requirement that: "On and after the first federal election that occurs after January 1, 2006, unless required sooner by the Help America Vote Act of 2002, if the voting machine is a direct recording electronic voting machine, it shall include a voter verified paper audit trail." Sec. 3506.10(P). That is in turn defined as "a physical paper printout on which the voter's ballot choices, as registered by a direct recording electronic voting machine," which "[t]he voter shall be permitted to visually or audibly inspect the contents of ...." In other words, a contemporaneous paper replica (CPR) of the electronic ballot is being required by 2006. The bill does require that "any system that produces a voter verified paper audit trail shall be accessible to disabled voters, including visually impaired voters, in the same manner as the direct recording electronic voting machine that produces it." It's unclear, however, exactly how this requirement can be met.
My take: This is very preliminary, since I've only had a brief opportunity to review this fairly long bill, but it appears to be one step forward and two steps back for the State of Ohio. The step forward is that the bill doesn't forbid conversion to electronic touchscreen machines or require the contemporaneous paper replica (CPR) by 2004. This is significant, because Ohio remains one of the last bastions of the punch cards, which were used by 70 of Ohio's 88 counties in 2000. (See here for my prior comments and disclosure that I'm cocounsel on the ACLU litigation challenging the use of punch cards in Ohio.) However, the bill doesn't appear to require replacement of this equipment -- so perhaps this is more accurately characterized as a non-step.
The two steps back are:
1. The bill appears to prevent the Secretary of State from compelling counties to get rid of their present machines. In the words of the Plain-Dealer, it "strips Blackwell's office of some of its authority over counties in the voting-machine selection process - including forbidding Blackwell from pressuring counties into selecting vendors in 2004 and wresting some of the office's future control over federal money earmarked for machines." This means that it's likely that the majority of counties which used punch card voting equipment in 2000 will likely end up sticking with it.
2. The bill requires the CPR, a.ka. "voter verified paper audit trail," which has yet to prove either workable or effective in any election. In the few jurisdictions that has tried it, serious problems have emerged, including printer jams and "appalling" user interface issues. It's also not likely to be an effective check on voting security, particularly given that there appears to be no requirement that the paper replicas are actually counted -- and absent such recounts, the CPR does you no good. The end result will thus likely be to dissuade counties from moving to electronic voting machines at all. This would leave voters stuck with less user-friendly voting technology, including paper-based systems that produce racial disparities in lost votes, and that don't accommodate non-English proficient voters or people with disabilities effectively.
Election Assistance Commission Hearing on May 5
The Election Assistance Commission will meet on May 5 to discuss electronic voting. Here are the details:
The Election Assistance Commission hearing will be:
When: May 5th, 9am-4pm with an hour lunch break
Where: EPA Building
1200 Pennsylvania Ave, Room 3000
Rachel L. Carson Great Hall
Ariel Rios North Building
What: Public Hearing on Use, Security and Reliability of Computerized Voting Systems
Panelists: Submit written testimony of any length by Wednesday, April 28, 2004. Send it to email@example.com.
Nonpanelists: Submit written testimony of any length by Wednesday, May 5, 2004. Send it to firstname.lastname@example.org. While email is preferred, it can be mailed to the following address: US Election Assistance Commission, 1225 New York Ave, NW, Suite 1100, Washington, DC 20005
Suggestions for testimonial format are as follows: double spaced, with an executive summary, highlight the important points, and details of the testimony, in that order as a recommended format. There will be a sign language interpreter and some set aside seating for people with disabilities. Bring ID and arrive early.
Recommendation from California Voting Systems & Procedures Panel
On April 21, April 22 and yesterday April 28,The California VSP Panel met to discuss whether electronic voting systems could be used in the November election. See here and here for prior discussion of the VSP proceedings. I'm informed that the VSP approved the following recommendations at yesterday's meeting:
The Voting Systems and Procedures Panel recommends to the Secretary of State that:
1. No new DREs can be used in the November 2004 election unless they include an AVVPAT and implement specified security measures (see No. 4).
2. Existing electronic voting machines in non-TSx counties cannot be used in California at the November 2004 Presidential Election unless the following steps are taken:
A. Implement a fully tested, federally qualified and state certified AVVPAT and implement specified security measures (see No. 4).
B. Or, utilize current electronic voting machines, in non-TSx counties, as long as specific conditions are met, including:
1) providing the option of voting on paper ballots at all polling places;
2) printing out as soon as possible after the closing of the polls all ballot images recorded on the machines;
3) parallel monitoring;
4) submit a technical security plan to the Secretary of State detailing compliance with the RABA report; and
5) implement further security provisions (see No. 4).
3. Funding for complying with the conditions should come from vendors and state and federal sources.
4. All electronic voting systems (DREs) must comply with the following security provisions:
Certification and Testing
1. Full federal testing and qualification and document review
2. Full state testing and certification
3. Documentation regarding processes and tools for development of system software and firmware
4. Working version of machines to Secretary of State
5. No last minute changes
6. No wireless connection
7. No Internet connection
8. Must submit a physical security plan to the Secretary of State
9. Must comply with all Secretary of State directives
10. Poll Worker Training Materials for each jurisdiction using the system must be submitted to the Secretary of State, and the training must provide adequate hands-on training for each poll worker for the DRE and any other device used
11. Must submit communications plan to the Secretary of State detailing how elections officials will communicate with workers at each polling place
13. Provisional voting must be on paper
14. Audio access connected prior to the polls opening
15. Results posted at polls
16. Tampering notice posted at polls
The bottom line appears to be that counties that are currently using DRE's can continue doing so, provided that they comply with the procedural requirements noted above; however, counties that have not yet converted will not be allowed to do so by November.
Here is the response from disability rights advocates:
VOTING PANEL DISENFRANCHISES VOTERS WITH DISABILITIES
Preventing counties to purchase new machines denies access
SACRAMENTO – Flying in the face of voting rights for people with disabilities, the Secretary of State Voting Systems and Procedures Panel today has recommended to the Secretary of State to halt all new purchases of electronic touchscreen voting systems by counties.
“The panel has issued an outrageous and discriminatory recommendation,” said Patricia Yeager, executive director of the California Foundation for Independent Living Centers. “This disenfranchises an entire population of people who just want to vote privately and independently like everyone else.”
For people with visual and manual impairments, touchscreen voting systems allow the user to access an audio component using headphones that allow them to vote on their own. Without touchscreen voting technology, these voters have had to ask complete strangers or their own young children to vote for them while poll workers and other voters looked on and listened.
In March, CFILC, along with the American Association of People with Disabilities, the California Council for the Blind and 11 individual plaintiffs filed suit against the Secretary of State and Los Angeles, Sacramento, San Francisco and Santa Barbara. According to the complaint, the Secretary of State has failed to require accessible voting machines in four of the state’s most populous counties. This lack of accessible equipment violates the equal protection provisions of the US Constitution, the Americans with Disabilities Act and the Rehabilitation Act of 1973.
“We’ll see what the Secretary of State truly stands for,” said Yeager. “Is he for access for people with disabilities as he claims or has he succumbed to the hysteria and will deny voting rights.”
We've not yet heard the last of this. Secretary of Shelley must still decide whether to approve these recommendations. And even if he does, the disability rights community seems likely to litigate the decision to disallow conversion to electronic voting by November. Stay tuned ...
Dee's Paper on Punch Card Errors
Thomas Dee of Swarthmore's Economics Department has posted this paper on punch card errors in the California recall election, noted in Rick Hasen's Election Law Blog. What's new about this paper is that he doesn't just look at residual votes -- i.e., overvotes and undervotes. Instead, he examines whether voters mistakenly cast votes for the wrong candidate, by looking into whether candidates positioned next to major candidates on the ballot garnered more votes than would be expected.
The paper presents evidence that a "relatively small but non-trivial number of voters in the recent California recall election mistakenly voted for the four candidates positioned next to the two major candidates on the ballot." Dee proceeds to inquire into whether the voting machine used affected the incidence of error. Consistent with other studies of punch cards, Dee finds that the use of punch cards significantly increased the likelihood of errors. In particular, he finds at least a third more voter errors with Votomatic-style punch card machines.
From India: A Real-Life Example of Voting Fraud
Concerns about voting fraud with electronic voting machines in the United States are thus far speculative. While mistakes can happen with electronic machines, there have been no documented instances of electronic voting machines in the U.S. being rigged. By contrast, there are numerous documented instances of ballot-stuffing and other forms of fraud with paper ballots.
That's not to say that fraud with electronic voting machines is beyond the realm of possibility. During Monday's voting in Bihar, India, "political hooligans" are reported to have seriously disrupted the election. Their methodology, however, was not the sort of high-tech fraud that some academics have focused upon. As described in this story from yesterday's New York Times:
In what appeared to be a carefully planned series of events, two small bombs exploded near the polling place and party workers threatened the five policemen guarding the booth and then brazenly took control of it. As poll workers and policemen averted their eyes, young party workers pushed the button for their party on the electronic voting machine over and over again, casting vote after fraudulent vote.
The men had carried out a new version of a storied Indian electoral trick: "booth capturing," in which armed thugs hired by political parties seize control of a polling place and stuff ballot boxes.
Most of the recent attention in the United States has focused on the possibility of sophisticated, high-tech attacks on electronic voting systems. This incident should remind us that low-tech attacks -- whether reliant on paper-based or electronic equipment -- are at least as likely.
We can hope that the U.S. will be spared the sort of violence that occurred in India's election. At the same time, it's important to keep in mind that no amount of hardware can guarantee election integrity. That includes the contemporaneous paper replica (a.k.a., "voter verified paper trail") advocated by some electronic voting critics. Unfortunately, the obsession with the CPR has blinded many election watchers to more significant problems in our election system, for which better procedures -- not a redundant paper replica -- is the immediate remedy.
Thanks to Michael Waterstone for the pointer to the NYT story.
Time Magazine Story on Electronic Voting
This week's Time magazine features this story on the electronic voting controversy. There's not much new here. To its credit, Time does mention the upsides of electronic voting -- including better access for disabled and non-English proficient voters -- though it gives more credence than is warranted to some of the more hyperbolic (and thus far unsubstantiated) assertions of DRE critics. Still, it's a pretty good introduction for those who've not been following this debate closely.
Orange County Approves Plan to Improve Electronic Voting Process
Acting in response to problems that occurred in the March primary, the Orange County Board of Supervisors has approved a "sweeping" plan to improve the process in time for November according to this story in today's L.A. Times. During the March primary, some voters were given the wrong electronic ballot by poll workers. The improvements include better poll worker training and "placing a paid and highly trained county employee at every polling place."
My take: From the sound of this story, Orange County is doing precisely the right thing. With all the attention that's been focused on electronic voting machines, it's easy to forget that machines are only one part of the voting system. Poll workers and election administrators are also crucial parts of that system. As noted here, the problems that occurred in OC wouldn't have been ameliorated through the contemporaneous paper replica (CPR), or "voter verified paper audit trail" advocated by some DRE critics. If we really want to avoid losing votes in November, counties should follow Orange County's lead and focus on procedures. Decertifying electronic voting systems, or requiring the CPR, would only make things worse.
Decertifying Touchscreens Won't Address the Underlying Problem
That's the headline of this op-ed in the Oakland Tribune, by Mary Lou Breslin and Diana Honig. The authors focus on the advantages of electronic voting machines for people with disabilities. They also note that touchscreens allow voters with limited English proficiency to vote independently. And they have error notification that allows all voters to "check their work," correcting any mistakes they might have made before casting their votes. While noting that problems did occur in a three jurisdictions that used electronic voting machines in California's March primary, the authors point out that problems also occured with paper-based voting ... but have largely been ignored. The authors argue that "decertification only sends the message that continuing to disenfranchise constituents with disabilities and members of language minorities is acceptable to create the political perception that the integrity of the voting process has been preserved, whether or not that is actually achieved."
Radio Debate on the Connection
This morning, I debated Stanford Computer Scientist David Dill on "The Connection," an NPR show in Boston. See here to listen to the show. Prof. Dill argued for decertification of existing electronic voting machines, which is presently being discussed in California as discussed here. I focused on the advantages of electronic voting from a civil rights perspective, and on the importance of focusing on procedures -- with both paper-based and electronic voting systems -- to make sure that votes are counted accurately in November.
The View from the Left Coast
I'm still in California, where I yesterday appeared on a panel regarding the electronic voting controversy at the Computers, Freedom and Privacy Conference. As expected, it was a lively debate. I focused on the need to understand not only technology, but also election administration and voting rights. I also discussed the comparative advantages of touchscreen voting, including accuracy, user-friendliness, and accessibility. Also present was Michael Shamos of Carnegie-Mellon, whose recent paper on electronic voting I discussed here. Among those representing the other side were Kim Alexander of the California Voter Foundation, who supports a ban on paperless electronic voting systems.
What's largely been lost amid the intense discussion of electronic voting systems over the past few days is the fact that, on the whole, electronic voting systems performed very well in the March 2004 California election. This is borne out by the parallel monitoring that was conducted by order of the Secretary of State's office. Parallel monitoring is a form of testing, in which randomly selected machines are pulled out of service to make sure they perform as they're supposed to. As set forth in a report from the Secretary of State's office: "Results of the reconciliation analysis indicate that the DRE equipment tested on March 2, 2004 recorded the votes as cast with 100% accuracy."
Unfortunately, this crucial finding is buried within a lengthy report released by the California Secretary of State's office at last week's hearing, described here. It was also lost on the New York Times, which today published this editorial calling California's electronic voting system compromised. The editorial repeats its mantra that the contemporaneous paper replica, or "voter verified paper audit trail" is the panacea for the supposed flaws in electronic voting -- completely overlooking the fact that it wouldn't have done anything to avoid or resolve the real-life problems that did occur in the most recent California election. See here for my prior thoughts on the NYT's position.
Most galling of all is the NYT's assertion that: "Bad decisions by voting machine manufacturers and local election officials have left California with a seriously compromised election system." While Diebold may have engaged in inappropriate conduct, if reports are correct (see here), I wonder what election officials the NYT is referring to? Riverside County Registrar Mischelle Townsend, who switched from an optical scan system to Sequoia's touchscreen in 2000, and has seen this system perform well through 27 elections? Or perhaps Alameda County registrar Brad Clark, who saw his county's residual vote rate decrease dramatically between the 2000 elections to among the lowest in the 2003 California recall, after switching to a touchscreen system? Again, I don't mean to dismiss or minimize the fact that problems have occurred with electronic voting machines in some counties, including Alameda. But the solution to these problems is NOT the CPR. Moreover, these problems shouldn't distract us from the fact that electronic voting machines result in fewer lost votes, and have accessibility advantages that other systems lack.
Michael Shamos said it best at the Computers, Freedom & Privacy conference yesterday. The call for a "voter verified paper trail" isn't based on science or evidence, but has instead taken on the character of religious zealotry. We've apparently reached the point where facts and evidence no longer matter. I had the opportunity to meet some of the CPR's proponents yesterday, including David Dill and Avi Rubin. They seem like genuinely nice people, and I have no doubt that they're very well meaning. But their advocacy, I fear, has done far more harm than good. It has already resulted in jurisdictions opting to stick with inferior punch card and central-count optical scan systems, but may now result in decertification of electronic voting systems used by 14 California counties. That would leave those counties scrambling to implement new voting systems between now and November -- a prescription not only for lots of lost votes, but for a potential election-day disaster.
Democracy in Action: The California Voting Systems Panel
Yesterday, I attended the California Voting Systems & Procedures Panels hearings regarding electronic voting. The first item on the agenda was Diebold's alleged misfeasance, including the use of uncertified software, discussed here. The eight-person panel voted unanimously to recommend decertification of Diebold's Accu-vote TSx, which was used by four counties (Kern, San Diego, San Joaquin and Solano) in the March 2004 elections. More details can be found in stories that appear in today's L.A. Times and San Diego Union Tribune. As set forth in the SDUT story, panel members asserted that Diebold:
Misled state officials about the prospects of the voting machines' receiving federal approval.
Installed uncertified software in machines in 17 counties.
Downplayed security concerns that had been raised in four separate studies.
(Thanks to Trent Douthett for the pointer to the SDUT story.)
It seems likely that Secretary of State Kevin Shelley will follow the panel's recommendation and decertify this model. Note that this decision does NOT affect all counties using electronic voting, or even all those using a Diebold DRE, but only those four counties that use the TSx.
The discussion then turned to whether California should vote to decertify ALL electronic voting machines. I testified against this proposal, noting that it would be a huge mistake to throw the baby out with the bathwater, particularly given the considerable advantages of electronic voting from a civil rights perspective. Numerous county registrars also testified against a blanket prohibition on touchscreen voting. In a particularly interesting moment, Jim Dickson of the American Association of People with Disabilities testified that a blanket ban on electronic touchscreens would violate the legal rights of people with disabilities, given that electronic machines are the only ones to allow accessible voting. Mr. Dickson, who is Vice President of the AAPD, held up legal papers that would be filed against the California Secretary of State in the event that the state banned electronic voting.
The panel will further discuss a blanket prohibition on electronic voting next Wednesday, according to the San Francisco Chronicle.
My take: Overall, I was impressed with the quality of the deliberations on the panel. The panel members generally seemed knowledgeable about the problems relating to electronic voting. The same, unfortunately, cannot be said about those in the legislature who are insisting upon a complete ban on electronic voting -- notably Senators Perata and Johnson, whose proposed bill I discussed here and here.
As is to be expected, there were a fair number of conspiracy theorists in attendance at the panel hearing, who seemed to lack either knowledge or a willingness to understand the complexity of the election process. But there were also some more intelligent and articulate critics of electronic voting there -- particularly people who were understandably upset by Diebold's alleged wrongdoing. If the allegations against Diebold are true, it's understandable why the panel would want to take strong action against the company. The problem, however, is that the decertification of the TSx leaves four counties holding the bag, and left to procure and implement a whole new voting system in less than seven months. DRE opponents underestimate what a tall order this is.
If the Secretary of State were to decertify ALL electronic voting machines, as some DRE critics urge, then the state would be asking for trouble on a much greater scale. Not only would this likely violate federal and state disability rights law, but it would also create a chaotic situation -- with some 14 counties scrambling to get a new voting system in place in a very short period of time. This would be a genuine recipe for disaster.
Yesterday's hearing reminded me of Winston Churchill's famous quip: "Democracy is the worst form of government except for all those others that have been tried." Though democracy can sometimes be messy, it's a good thing for there to be a lively public dialogue on issues of great public importance, such as our voting machines. It's too bad we didn't have a similar debate on the integrity of punch card voting BEFORE the 2000 election fiasco. And it's a shame that there has been so little attention paid to the comparative advantages of electronic voting, relative to other systems. Fortunately, yesterday's meeting did feature some discussion of this issue, and the panel seems to recognize the considerable advantages that electronic voting carries -- especially for traditionally disenfranchised populations.
Churchill's admonition about democracy is equally applicable, in my view, to electronic voting: It's the worst form of voting out there ... except for all the others that have been tried.
Chronicle of Education Story on Electronic Voting
The April 23 Chronicle of Higher Education features this article on the electronic voting debate. The article features quotations from Avi Rubin, discussed here, as well as Michael Shamos whose work I discussed here. Key excerpts are below.
Wary of E-Voting, Some Professors Sound the Alarm
Others say we'll byte the ballot eventually, so let's make electronic elections tamperproof
By PETER SCHMIDT
It is hard to dismiss Aviel D. Rubin as a conspiracy nut. He speaks in calm, measured tones, is an associate professor of computer science at the Johns Hopkins University, and is regarded as one of the nation's leading experts on computer security.
But the warnings that Mr. Rubin issues to the news media from his office here at the university's Information Security Institute sound as if they have been lifted from a Hollywood thriller with a trust-no-one plot.
The November 2004 presidential election could be rigged, he says. Worse yet, the public might never even know it. The culprit could be a company that manufactures electronic voting machines, or a rogue computer programmer on that company's payroll, or an election official, or a poll worker who has been bribed by foreign agents. Or someone else entirely. There's no telling just who might tamper with electronic voting machines to manipulate the democratic process ....
Even proponents of electronic voting, who feel that Mr. Rubin and Mr. Dill are crying wolf about the potential for fraud, acknowledge the critics' influence.
"They have slowed progress, for sure," says Michael I. Shamos, a professor of computer science at Carnegie Mellon University who has spent two decades evaluating electronic voting systems for state governments.
However, Mr. Shamos dismisses many of the arguments used by critics of electronic and online voting as "purely emotional" and devoid of science. "If I give them a scientific explanation of why they are wrong," he says, "they simply ignore it." He characterizes the debate as "a shrieking match," in which detractors have gained the upper hand through a slick, well-organized public-relations campaign that falsely brands demurring voices as ignorant.
"You only hear the people who are yelling," he says. "You don't hear the silent majority."
Diebold's Use of Uncertified Software
Tuesday's Tri-Valley Herald features this story on Diebold's use of uncertified vote-counting software in Alameda County. The story reports that faulty software was used in the March 2004 election, despite the fact that Diebold's attorneys "warned in late November that its use of uncertified vote-counting software in Alameda County violated California election law and broke its $12.7 million contract with Alameda County." The story also reports on the Secretary of State's deliberations over whether to allow the continued use of electronic voting, noting that a decision to prohibit the use of Diebold's system "could send two of California's largest counties -- Alameda and San Diego -- scrambling for other ways to count votes six months from now."
My take: This story describes behavior on Diebold's part that, if true, is totally unacceptable. But it does not remotely justify the baby-with-the-bathwater approach of banning electronic voting entirely, which some have advocated. See here. It's important to emphasize that the problem isn't limited to electronic voting, since vote-counting software is used to count paper ballots as well. Although the story is not entirely clear on this point, it sounds as though the software in question is used for counting optical scan as well as electronic votes. Thus, the solution of banning electronic voting (or alternatively requiring a contemporaneous paper replica of electronic ballots) is both too broad and too narrow. Too broad, because it would ban electronic voting equipment that has worked well. Too narrow, because it wouldn't deal with the problems that can occur with paper-based systems, and may not be detected through the 1% manual recount required under California law.
The bottom line is that California voters shouldn't be punished for Diebold's bad behavior. Yet that's exactly what the "solutions" now being proposed would do.
Computers, Freedom & Privacy
Blogging will be sporadic for the rest of the week, as I'll be in Northern California for the Computers, Freedom & Privacy Conference. On Thursday, I'll be in Sacramento attending the Voting Systems & Procedures Panel hearing on electronic voting. See here for my comments to the Secretary of State's office. On Friday, I'll be appearing on a panel at CFP on the Great Paper Trail Debate, along with Michael Shamos (whose paper is discussed here) and others. It should be a lively exchange. My focus will be on the civil rights advantages of electronic voting -- especially for minority, non-English proficient, and disabled voters -- and on the practical difficulties of implementing the contemporaneous paper replica, which have largely been overlooked by DRE critics. I hope that we'll also have a chance to explore alternative means of enhancing the integrity of elections more effective than the CPR.
Lynn Landes' Piece on Voting Machines
Lynn Landes offers this opinion on voting machines in the Baltimore Chronicle & Sentinel. Landes notes that 99.4% of voters use some type of voting machine, either paper-based (optical scans or punch cards), or paperless (electronic or lever machines). She proceeds to make the unremarkable -- though often overlooked -- point that no voting machine is perfect. From this fact, however, she draws the remarkable conclusion that: "The only fix that will give Americans back their constitutional right to vote is to ditch the machines." And she doesn't mean only electronic voting machines, but ALL voting machines.
My take: For anyone who knows anything about contemporary voting, Landes' opinions are bound to seem outlandish if not insane. To suggest that we all go back to paper ballots -- now used only in the smallest jurisdictions -- is not practicable. Nor would it solve the problem, since even paper ballots have to be read and interpreted, and we all know that human beings, like machines, sometimes make mistakes.
At the same time, Landes' argument does take to their logical extreme the arguments made by some electronic voting critics. Overlooking the civil rights advantages of electronic voting, these critics focus on the hypothetical threat posed by the possibility of someone tampering with software. I don't mean to deny that there's a possibility of fraud and error with electronic voting, just as there is with paper-based voting. But in their thirst for a perfect solution, Landes -- like DRE voting's critics -- have overlooked the serious problems that exist with paper-based voting. They've also given short shrift to the considerable advantages of electronic voting for people of color, non-English proficient voters, and people with disabilities.
The perfect is the enemy of the good.
DRE Critics to Sue Registrars Across the Country
I've been receiving reports that the electronic voting critics will be filing lawsuits across the country against registrars who use or have approved the use of electronic voting systems, including those in Maryland, North Carolina, and New Mexico. While it's not yet clear what the basis for these lawsuits are, previous challenges to paperless electronic voting in California and Florida have been rejected. See here. In the only published decision of which I'm aware to consider such a claim, Weber v. Shelley, the Ninth Circuit found that there was no evidence that paperless DRE's were "inherently less accurate or ... less verifiable, than other systems." The court proceeded to reject the argument that electronic voting violated the equal protection rights of voters. In reaching this conclusion, it observed that touchscreen voting results in "numerous positive changes" including "increasing voter turnout, having greater accuracy than traditional systems, being user-friendly, decreasing the number of mismarked ballots, saving money, etc." By contrast, the concerns raised about electronic voting were entirely "hypothetical."
More on this as details emerge ...
Michael Shamos' Paper on Paper v. Electronic Voting Record
Michael Shamos, Distinguished Career Professor of Computer Science at Carnegie Mellon, has posted this paper addressing the controvesy over electronic voting. While acknowledging that there are security concerns with electronic voting, Shamos concludes that much of the criticism of electronic voting -- particularly the call the contemporaneous paper replica (or "voter verified paper audit trail") -- is misguided. I've pasted the abstract, the opening paragraph, and concluding paragraph of his article below:
There has been much discussion in the popular press concerning the use of contemporaneous paper trails to plug various perceived security risks in electronic voting. This paper examines whether the proposed paper solutions in fact provide any greater security than properly maintained electronic records. We conclude that DRE machines pose a number of security risks but that paper records do not address them. A number of alternatives to paper trails are suggested to respond to DRE security concerns.
Among the arguments that have been advanced against the use of direct-recording electronic (DRE) voting systems are the following:
1. Voting machines are “black boxes” whose workings are opaque to the public and whose feedback to the voter is generated by the black boxes themselves. Therefore, whether or not they are operating properly cannot be independently verified and the machines should not be used.
2. No amount of code auditing can ever detect malicious or even innocently erroneous software. Therefore the machines should not be used.
3. No feasible test plan can ever exercise every possible combination of inputs to the machine or exercise every one of its logic paths. Therefore the machines should not be used.
4. Hackers can break into the FBI’s servers and deface its website. It ought to be child’s play for them to throw an election. Therefore the machines should not be used.
5. DRE machines have been plagued by a host of failures all around the country. Therefore the machines should not be used.
6. The DRE industry is dominated by a small number of companies, some of whose executives are announced supporters of the Republican party. An executive could command his programmers to add code to each machine manufactured by that company to move votes to a favored candidate, thus determining the outcome of the election. Therefore the machines should not be used.
7. Many prominent computer scientists have said that DRE machines cannot be trusted. Therefore they should not be used.
8. If added to a DRE machine, a voter-verified paper trail allows the voter to satisfy herself that her voting preferences have been recognized correctly by the machine. Therefore, the voter-verified paper trail solves every one of the aforementioned problems and every DRE machine should be required to have one.
Each of these arguments will be examined in this paper and found fatally flawed, at least to the extent that it implies that machines cannot be relied upon to count votes in real elections. The numbered statements above all share the property that their first sentence of their premise is true, yet their consequent, that DRE machines should not be used, does not follow from the premise.
DRE machines have been described, somewhat dramatically, as a threat to democracy. A far greater threat to democracy is a return to any form of paper ballot, but both of these pale in comparison to the fact, not widely known, that in each presidential election more than 5 million Americans who are eligible to vote and want to vote are unable to cast a ballot because they happen to be outside their home districts on election day and cannot comply with their state’s absentee procedures. Many of these people are overseas. The claim that tens of thousands of Floridians were disenfranchised in the 2000 election because of butterfly ballots, though probably true, is insignificant when measured against the millions who were unable to obtain any ballot at all. If computer scientists are truly concerned about threats to democracy, that’s one they should work on.
My take: This is an excellent paper that is a must read for anyone with even a passing interest in the controversy over electronic voting. He injects a needed dose of reality into a debate in which a real-world understanding of elections has been astonishingly rare. He also suggests genuine means of improving security -- such as better procedures, open source, and parallel testing -- that are much more likely to enhance election integrity than the CPR.
Election Integrity and Voting Machines
With all the attention that's been focused on electronic voting security in recent months, it's easy to forget that serious problems can also result from the use of paper-based voting. Just as there can be problems with electronic voting if proper procedures are not followed, paper ballots can be lost, misplaced, or miscounted -- to say nothing about the possibility of deliberate fraud, through stuffing or "losing" ballots.
Problems with paper-based ballots are not just hypothetical, but actually happen in real-life elections. In Lucas County, Ohio, for example, some 300 ballots were "forgotten" and found in a storage closet, according to this story in the Toledo Blade. This threw into question the result of a school funding measure, decided by 33 votes. And in Orange County, California, some 2,800 paper ballots used by absentee voters were miscounted, according to this story from the L.A. Times. Most attention in Orange County has focused on the electronic voting system used for in-precinct voting, see here, but the miscounted paper ballots remind us of what we should have learned in 2000: that paper-based systems are also subject to large-scale error.
The response that comes from electonic voting critics is that, with paper ballots, we can be sure that our votes were counted accurately. If there's a paper ballot, the argument goes, there will be a record that we can rely to be sure that the election was conducted on the square. But is this really true? Paper ballots and even ballot boxes can be lost or misplaced, intentionally or otherwise. Can we really be sure that no other ballots were misplaced in Lucas County? And, as we all know, stray marks and hanging chads can lead to uncertainty in what voters intended.
That's not to say that all paper-based systems are inherently unreliable, any more than problems with electronic systems prove that those systems are inherently unreliable. The point is that any voting system is subject to fraud and error, if proper procedures aren't followed. And a voting system consists not only of machines, but also of the people who use them. With the November election fast approaching, it's vital to ensure that proper election procedures are in place and are followed. If this doesn't happen, then no machine -- whether paper-based, electronic, or an attached printer -- will save us.
California Hearings on Electronic Voting
Today's L.A. Times features this story on the California Secretary of State's office hearing set for next Wednesday and Thursday, on voting machines to be used in November's election. After three days of hearings, Orange County officials decided that they want to keep their electronic system, despite problems experienced in the March election. See here for my previous discussion of O.C.'s problems. The contemporaneous paper replica, or "voter verified paper trail," urged by some wouldn't have done anything to resolve these problems. My comments to the California voting panel can be found here.
Vote Counting Widespread, Not Confined to Florida
That's the headline from this story from Scripps Howard News Service. The story reports that over 1.6 million ballots didn't record a vote for President in 2000. While some of these were intentional, others were the result of "antiquated voting equipment, mechanical failures, improperly programmed tabulation devices and faulty accounting methods." In fact, seven states had higher undervote rates than Florida. Some election officials argue that the number of ballots that don't record a vote -- also known as "residual votes" -- are meaningless. Others recognize that they can be a sign of problems with voting machines.
My take: I don't agree with those who contend that the residual vote rate tells us nothing about the reliability of voting equipment. To be sure, isolated counties -- especially small ones -- may have high or low residual vote rates that can be explained by something other than the equipment used. But a consistent pattern of high residual votes with a particular type of equipment, such as the "hanging chad" punch card, is a sign of trouble. For example, Henry Brady of U.C. Berkeley has this analysis of California counties' residual vote rates in 1996, 2000, and 2003. He finds that punch card jurisdictions consistently have higher residual vote rates than those using other voting equipment, and that those which switched from punch cards to other systems lowered their residual vote rate. While there's no question that some residual votes are intentional, there's also no real question that some are unintentional. In light of this evidence, it can't plausibly be disputed that punch cards result in more unintentional non-votes than other equipment.
Catching Up: News from Ohio
I've spent the past few days working with the ACLU of Ohio, in its voting rights lawsuit challenging the use of punch card voting machines. Ohio is one of the last bastions of the "hanging chad" punch card, used by 70/88 counties in 2000, and has been proceeding at a glacial pace in replacing that system. The ACLU voting lawsuit alleges that the continuing use of punch cards violates equal protection and denies the voting rights of African Americans.
Last week, a state legislative panel recommended that electronic voting machines be required to generate a contemporaneous paper replica ("CPR") by 2006. See here for the Toledo Blade article on the state legislature's decision. Because no CPR system has successfully been implemented anywhere as yet, this is likely to delay -- if not derail completely -- the transition to electronic voting. The Mansfield News Journal and Morning Journal have both editorialized against the legislative panel's decision, noting that it will slow down the transition to more modern voting equipment and complicate the voting process, while providing dubious security advantages. As the Morning Journal puts it: "There's no reason to keep wasting time."
Leadership Conference Testimony to U.S. Commission on Civil Rights
The Leadership Conference on Civil Rights has now posted Executive Director Wade Henderson's testimony at Friday's USCCR hearing here. As usual, Mr. Henderson's thoughts on civil rights are worth paying attention to. He emphasizes the advantages of electronic voting machines from a civil rights perspective -- including lowering error rates -- while at the same time recognizing the need to deal with the security issues that have received such prominent attention. He emphasizes three principles that should guide the debate: 1) equal access for all voters, including people with disabilities and non-English speakers, 2) second-chance voting, so that voters may "check their work" before votes are cast, and 3) compliance with national certification standards, to make sure that all votes are counted accurately throughout the country. He notes that it would be impractical to require a contemporaneous paper replica (the "voter-verified paper trail") in time for the 2004 election, and suggests a number of other means by which to promote secure electronic voting. These include thorough testing, poll worker training, parallel monitoring, and preventing tallies from being transmitted via the internet. Finally, he emphasizes that technology is only one aspect of the problem, and that we need to focus on polling place operations and poll worker training.
My take: This is a terrific statement. The USCCR would do well to pay careful attention to the suggestions that Mr. Henderson and the Leadership Conference have made. We don't have to choose between having secure elections and protecting the civil rights of traditionally disenfranchised voters.
Prior Sixth Circuit Opinion on Accessible Voting Systems
Law week, I blogged on the recent decision in AAPD v. Hood which appears to be the first case to hold that the Americans with Disabilities Act requires voting equipement that allows people with manual and visual impairments to vote independently. See here. My friend Kevin Russell alerts me to a prior opinion, Nelson v. Miller, 170 F.3d 641 (6th Cir. 1999), which considered a claim under the ADA and the Rehabilitation Act. This is an odd opinion. As the Sixth Circuit characterizes plaintiffs' claim in Nelson, its "lynchpin" was a provision of the Michigan Constitution guaranteeing a secret ballot. Thus, plaintiffs did not claim that the ADA or Rehabilitation Act independently confer the right to vote independently. The Sixth Circuit rejected the claim in Nelson, based upon a Michigan statute that provided blind voters with third-party assistance rather than guaranteeing their right to vote independently. The court found the Michigan statute to be consistent with the "secret ballot" provision of the Michigan Constitution -- thereby avoiding the issue whether federal law protects the right of people with disabilities to vote independently.
As I said, this is an odd opinion. Unless I'm missing something, it doesn't squarely address the question whether the ADA or Rehabilitiation Act require the provision of equipment to allow people with disabilities to vote independently. Nor does it consider the technology that's currently available -- such as electronic touchscreen machines -- to allow independent and secret voting for people with disabilities. Thus, even in the Sixth Circuit, it doesn't seem to me that Nelson forecloses the sort of claim that we've recently seen succeed in AAPD v. Hood, against jurisdictions that fail to provide touchscreen voting systems to people with manual and visual impairments.
L.A. Times Report on U.S. Commission on Civil Rights Hearing
Saturday's L.A. Times featured this story on Friday's electronic voting hearing before the U.S. Commission on Civil Rights, discussed here. The LAT story discusses testimony from Meg Smothers of the League of Women Voters, that electronic voting reduced the rate of unvoted ballots significantly, when implemented in Georgia. It also reports on testimony regarding the advantages of electronic voting for disabled and non-English speaking voters. Finally, the LAT discusses conflicting testimony on the security issue, with Rebecca Mercuri taking the position that electronic voting is inherently insecure and Michael Shamos of Carnegie-Mellon taking the contrary position position -- and urging thorough testing and poll worker training as ways of promoting safe implementation. Additional solutions are to be explored at the USCCR's September meeting.
Seattle Times Editorial
I'm presently in an airport on the way to Seattle, so it seemed appropriate to note that Lance Dickie has this column in today's Seattle Times regarding the paranoia surrounding electronic voting, which he compares to the Y2K craze.
Dickie notes that many jurisdictions are now replacing their paper-based systems with more user-friendly electronic systems, which do away with the amibiguous markings that caused so many problems in Florida. He also discusses the experience of Snohomish County, which has successfully used such a system since 2002. These machines, Dickie emphasizes, are stand-alone units that are not hooked up to the internet or to a network, and run on software that has been independently tested and certified. The county also conducts rigorous logic and accuracy testing before elections. That's not to say the system is foolproof -- no system is -- but if Dickie's account is accurate, then Snohomish County appears to be doing the sort of things that should be done, to promote safe electronic voting.
U.S. Commission on Civil Rights Hearing on Election Reform
Tomorrow, the U.S. Commission of Civil Rights will hold a hearing on "Integrity, Security and Accessibility in the Nation's Readiness to Vote." See here. Scheduled to speak before the commission are electronic voting critic Rebecca Mercuri, Carnegie-Mellon computer scientist Michael Shamos, Meg Smothers of the League of Women Voters, Maria Rodriguez Salazar of LULAC, Wade Henderson of the Leadership Conference on Civil Rights, Jim Dickson of the American Association of People with Disabilities, and Larry Gonzalez of NALEO. The USCCR has recently released this briefing paper on the state of election reform, entitled "Is America Ready to Vote?" Mary Francis Berry, the USCCR's chair, states that the hearing is to address questions that have been raised regarding the security of voting systems. It should be an interesting hearing.
The Right to Accessible Polling Places
In yesterday's blog, I discussed the recent federal district court decision in AAPD v. Hood, concluding that voters with visual and manual impairments have a right under the Americans with Disabilities Act to equipment allowing them to vote independently. This appears to be the first decision ever to apply the ADA to voting equipment. Michael Waterstone reminds me that there's a prior district court decision in People of New York v. County of Schoharie, 82 F. Supp. 2d 19 (N.D.N.Y. 2000), holding that the ADA protects the right to accessible polling places. Thanks for the tip Michael.
Federal Court Orders Accessible Voting
A federal court in Jacksonville, Florida has ordered Duval County to provide accessible voting machines in time for the 2004 elections. The decision in American Association of People with Disabilities v. Hood is, to my knowledge, the first ever to hold that the failure to provide accessible voting equipment violates the Americans with Disabilities Act of 1990. The decision doesn't appear to be published on a publicly available website but, for those with Westlaw access, can be found at 2004 WL 626687. (Thanks to Michael Waterstone for the pointer.)
The case arose from Duval County's decision to purchase a paper-based optical scan system in 2002. That system did not allow people with visual and manual limitations to vote independently. The AAPD v. Hood court found that, at the time this paper-based system was purchased, "it was technologically and financially feasible to employ a technology readily accessible to visually impaired voters" -- namely, electronic touchscreen machines. The court also found that touchscreen systems would have permitted at least some voters with manual dexterity impairments to vote independently. It therefore concluded that the county's use of an inaccessible optical scan system violated ADA regulation 28 C.F.R. 35.151(b). The court's order requires that accessible voting machines be available at 20% of the county's polling places in this year's elections.
My take: This is a landmark decision for voters with disabilities. For the first time I'm aware of, a court has held that federal law protects their ability to cast a secret ballot just like any other voter. The decision should also serve as a wake-up call to those jurisdictions which have not yet made accessible voting technology available to citizens with disabilities. If election officials don't provide accessible voting machines -- and the only ones presently available are electronic -- they're likely to be found in violation of the ADA. The decision also means that proposed legislation seeking to prohibit electronic voting, such as California SB 1723 which I discussed here, would almost certainly violate the ADA. In light of the decision in AAPD v. Hood, we should expect see more lawsuits like the one recently filed by disability rights advocates in California, described here, in coming months.
Comments to California Voting Systems Panel
As noted yesterday, the California Voting Systems & Procedures Panel will be conducting a hearing on April 21-22, on the voting systems to be used in November's election. This comes amid calls by some legislators to ban the use of electronic voting entirely in that election. Below are my written comments to Secretary of State Shelley and the panel. For those uninterested in reading the whole letter, my position is that foreclosing the use of electronic voting machines in the November 2004 election is a bad idea, one likely to do more harm than good. A better approach is to make sure that adequate procedures are in place -- for both electronic and paper-based systems -- so that everyone's vote is accurately counted.
April 6, 2004
VIA E-MAIL, FACSIMILE, AND U.S. MAIL
Secretary of State Kevin Shelley
Voting Systems and Procedures Panel
1400 11th Street, 5th Floor
Sacramento, CA 95814
Attn: Michael Wagaman
Re: Electronic Voting
Dear Secretary Shelley and Members of the Voting Systems and Procedures Panel:
I request the opportunity to present an oral statement at the April 21-22 Voting Systems and Procedures Panel (“VSPP”) meeting, regarding the voting systems to be used in the November 2004 elections (Agenda Item 3). I also ask that you accept and consider the written comments set forth below.
As these comments explain, electronic voting technology provides significant advantages over paper-based systems from a civil rights perspective. The experience of California counties has shown that it is possible to implement electronic voting safely and effectively, if proper procedures are followed. I therefore urge that you resist the calls that some have made to forbid electronic voting during the coming election, a drastic and unnecessary step that would have profoundly negative consequences for countless voters throughout the state and thrust the state into conflict with voting rights laws. Instead, counties should adopt and follow procedures – including ones recommended by the Secretary of State’s electronic voting task force – to ensure secure and accessible voting by all citizens.
As both an advocate and a scholar, I have a longstanding interest in the voting equipment used in California and other states. From 1995 through 2003, I was a staff attorney at the ACLU of Southern California, in which capacity I served as co-counsel for plaintiffs in Common Cause v. Jones. That case, as you know, resulted in decertification of pre-scored punch card machines effective March 2004. I am also the immediate past Chair of California Common Cause. Presently, I am an Assistant Professor of Law at the Ohio State University, where my research interests include voting rights, racial equality, and disability rights. My recent writings in the area of political equality include “First Amendment Equal Protection: On Discretion, Inequality and Participation,” 101 Michigan Law Review 2409 (2003) and “Political Equality After Bush v. Gore: A First Amendment Approach to Voting Rights,” in Final Arbiter: The Consequences of Bush v. Gore for Law and Politics (SUNY Press, forthcoming 2004). I am also the author of a weblog (www.equalvote.blogspot.com) which addresses civil rights and voting technology issues. The comments that follow are, of course, made only on my own behalf and not on behalf of the ACLU, Common Cause, or the Ohio State University.
There are three dimensions to the current debate over electronic voting: 1) technology, 2) election administration, and 3) voting rights. Unfortunately, the media coverage of this debate has focused almost exclusively on technology, with relatively little attention to the election environment within which this technology is implemented. It can scarcely be denied, however, that the security of any voting equipment – whether paper-based or electronic – cannot intelligently be assessed in isolation from the procedures followed by election officials.
So too, the impact of any particular technology cannot be assessed without a firm understanding of the voting rights protected by federal and state law. Foremost among these are laws that prohibit racial discrimination, that require materials for non-English proficient voters, and that provide for accessible voting for citizens with disabilities. Unfortunately, these voting rights have largely been disregarded amid the intense controversy that has surrounded the introduction of electronic voting.
From a voting rights standpoint, electronic voting machines offer significant advantages over their paper-based counterparts. Electronic voting reduces the racial disparity that tends to exist with punch card and optical scan voting systems. Michael Tomz & Robert P. Van Houweling, “How Does Voting Equipment Affect the Racial Gap in Voided Ballots?” 47 Am. J. Pol. Sci. 46 (2003). It also provides for easier multi-lingual access, allowing non-English proficient citizens to cast votes without assistance.
In addition, electronic voting machines are the only type of certified equipment that allows people with disabilities to vote independently. They can be outfitted with devices that allow people with manual dexterity impairments to vote independently. In addition, they have an audio capacity that not allows visually impaired citizens and those who are unable to read to cast secret ballots.
Electronic voting machines can also help in reducing the rate of non-voted ballots. They do not allow overvotes, and have verification screens that allow voters to check their ballots, to make sure that they have cast their votes as intended. As a result of these features, electronic voting machines fared much better than paper-based systems in the October 2003 recall election. According to one analysis, the rate of non-voted ballots with electronic machines was 1.5%, compared to 2.7% with optical scan equipment and 6.3% with punch cards. At least 160,000 to 170,000 votes were lost through the use of punch card systems in the recall election.
Despite the considerable advantages of electronic voting over paper-based systems, some have called for a prohibition on the use of these systems in the November election. They base their opposition to the use of electronic voting systems on problems that occurred in three counties (Alameda, Orange and San Diego) during the March 2004 election.
Although the calls to stop the use of electronic voting may be well-intentioned, they ignore the fact that electronic voting systems appear to have worked well in 11 of 14 counties. In fact, in one county (Napa) the electronic system seems to have functioned flawlessly, while serious problems occurred with the optical scan system used for absentee ballots. While most of the attention has been focused on alleged problems in electronic voting, this incident demonstrates that paper-based systems are not immune from error. To the contrary, like any other system, optical scan voting is vulnerable to problems if adequate procedures are not followed.
None of this should be understood as minimizing the importance of preventing a recurrence of the problems that occurred in three counties that used electronic voting in the March 2004 election. At the same time, experience teaches us that some problems are inevitable any time jurisdictions move to a new voting system. There is no such thing as a perfect election system, and it is to be expected that there will be kinks to be worked out when new voting equipment is put in place.
Such problems cannot remotely justify the drastic step of foreclosing counties from using electronic voting systems in the forthcoming election. To do so would throw out the baby with the bathwater, depriving California voters of the considerable advantages that electronic voting systems carry. It is also completely disproportionate to the magnitude of the asserted problem. For example, in Orange County’s election, a recent Los Angeles Times story reported that 33 out of 16,655 voters in the 69th Assembly District received the wrong ballot. While the loss of any vote is undoubtedly serious, this small number of lost votes (approximately 0.2% of all votes cast) pales in comparison to the number of votes lost due to the continuing use of paper-based systems. As noted above, the lost votes in the California recall are estimated to have been in the hundreds of thousands statewide.
A ban on electronic voting would also violate the voting rights of citizens with disabilities throughout California. At present, electronic voting machines are the only type of system that allows people with disabilities to vote independently. There is already a lawsuit pending against the Secretary of State regarding the lack of accessible voting. American Association of People with Disabilities v. Shelley. In the event that electronic voting is prohibited in November, these citizens’ voting rights are certain to be violated and the citizen will be exposed to substantial liability.
The baby-with-the-bathwater approach also disregards the fact that the problems that have been reported could have been avoided altogether through improved election procedures. For example, the problem that occurred in Orange County – a small number of voters receiving the wrong ballot – is one that could occur with any type of voting equipment. So too, the difficulties encountered in Alameda and San Diego counties with Diebold’s system could have been avoided through better procedures. In both these counties, the precinct control modules (“PCM’s”) used to encode smartcards reportedly failed to boot properly. In particular, poll workers should be given training in advance on the relatively simple “four-click” process for resolving the problem in a matter of minutes.
It bears emphasis that neither of the problems reported with electronic voting would have been resolved by the CPR, so often touted – wrongly in my opinion -- as the solution to the security issues surrounding electronic voting. There is as yet no CPR system that has been proven in a real-life election environment, and considerable problems emerged when such a system was tested on a limited basis in Sacramento’s 2002 election. Moreover, no CPR system is certified in California to my knowledge. And the CPR would have done nothing to remedy the problems that have been reported in San Diego, Alameda and Orange counties. Rather than insisting on hardware that will do little if anything to enhance security, we should focus on implementing procedures to make sure that votes are counted accurately.
It is simply wrong to assert that electronic voting is more susceptible to error than paper-based systems. The Ninth Circuit has firmly rejected the argument that paperless electronic voting compromises citizens’ voting rights. Weber v. Shelley, 347 F.3d 1101 (9th Cir. 2003). The unanimous court in Weber concluded that plaintiff had raised “at most a hypothetical concern about the ability to audit and verify election results.” Id. at 1103. There was no evidence to show that electronic systems were “inherently less accurate” or “less verifiable” than other systems. Id. at 1105. To the contrary, the evidence noted above shows that they are more likely to be accurate than other systems.
It would therefore be a grave error to require all of the election jurisdictions using electronic voting systems to procure and implement paper-based systems in time for the November election. If there is anything that we should have learned over the past few years, it is that it takes time and money to implement a voting system effectively. To force counties to change their plans at this late date, and put in place a different voting system in less than seven months, is completely unrealistic. It is the classic case of a cure far worse than the disease.
This is not to deny that there are things that can be done to improve voting security between now and November. In fact, the Secretary of State’s office already has available to it a document that contains excellent suggestions on how the security of electronic voting can be protected without sacrificing the voting rights of minority, disabled, and non-English proficient voters. In particular, the July 2003 report of the Secretary of State’s Ad Hoc Touch Screen Task Force contains several recommendations for enhancing security that are worthy of consideration. These include parallel testing of machines, and a “chain of custody” to ensure the integrity of software.
Rather than adopting a hasty and drastic “fix” – a ban on electronic voting -- that is certain to cause more problems than it solves, the Secretary of State’s office should focus on improving the procedures used in counties throughout the State of California. This includes not only those used to implement electronic voting, but also those used to implement paper-based systems. By so doing, the Secretary of State’s office can enhance the integrity of the state’s voting systems without sacrificing the voting rights of the hundreds of thousands of Californians who stand to benefit from electronic voting.
There can be no question that safeguarding the integrity of our voting system should be among the State of California’s highest priorities. While we should obviously take seriously the security concerns that have been raised in recent months, the VSPP should reject calls to decertify electronic voting systems – and especially to avoid doing so in advance of the November 2004 election. Such precipitous action would not only place the state into clear conflict with voting rights laws, but would also throw the forthcoming election into a state of disarray. It is far too late in the day to force the counties using electronic voting machines to procure and implement paper-based voting systems, without risking an election-day disaster.
Nor should the Secretary of State’s office accede to calls to force implementation of the unproven contemporaneous paper replica (the CPR, or “voter verified paper audit trail”) by November 2004. Instead, the Secretary of State’s office should work to ensure that counties have in place adequate procedures to ensure that all voting equipment – whether paper-based or electronic – operates as intended, so that all Californians can enter the polling place in November with confidence that their votes will be counted.
I look forward to speaking before the VSPP and addressing any questions you may have at your meeting later this month.
Daniel P. Tokaji
Deadline Tomorrow for Comments on California Voting Systems
The California Voting Systems and Procedures Panel ("VSPP") will be meeting on April 21 and 22 to discuss voting systems to be used in the November 2004 election. The agenda for this meeting can be found here. This meeting comes amid calls by some California legislators to prohibit electronic voting altogether in this election, a move likely to have disastrous consequences as described here and here. The meeting will probably be heavily attended by the most vocal critics of electronic voting, including those who think that a contemporaneous paper replica should be required. It's important that this not be the only perspective represented before the VSPP. In particular, the VSPP should be reminded of the considerable advantages of electronic voting, from a civil rights perspective, especially for voters of color, non-English speakers, and people with disabilities.
Written comments must be submitted by 5 pm tomorrow, April 6. I plan to submit written comments to this panel and hope that others will consider doing so as well. Comments may be submitted by email to Michael Wagaman in the Secretary of State's office, at email@example.com.
ES&S Press Release on Accessible Optical Scan Voting System
Election System & Software (ES&S) and Automark issued a media release yesterday regarding a new ballot-marking system that they claim will allow independent voting for people with visual impairments. Until now, electronic voting machines have been the only type of voting equipment to allow independent voting for people with visual and manual dexterity impairments.
The "ES&S Automark" is not yet federally certified, and it does not appear that it has been used in any actual elections yet. Nevertheless, the press release claims that "The ES&S AutoMARK will be the only product in the marketplace that will enable a greater number of voters to mark a vote privately and independently using an optical scan voting system. Once certified, ES&S plans to market the unique product to all jurisdictions in the United States that currently use or plan to upgrade to optical scan equipment." The press release further claims that this system can be used to comply with HAVA's requirement that at least one accessible device be available in each polling place by 2006.
My take: This is a promising development but, as with any other system, the vendor's claims should be taken with a grain of salt. In particular, we should avoid rushing to any conclusions about this device until it has been examined, certified, and implemented in real-world election environments. Like any other voting system, what sounds great on paper may or may not turn out to be so great in practice. It would be a big mistake to lock into place any particular fix to the accessibility/security conundrum, at least until it has been proven that the system works as advertised. Nevertheless, I hope that some jurisdictions will experiment with the ES&S Automark system in coming elections.
Three Dimensions to the Electronic Voting Debate
There are three dimensions to the current controversy over electronic voting 1) technology, 2) election administration, and 3) voting rights. To understand this debate, one must pay attention to all three of these things.
Unfortunately much of the coverage of this controversy has focused almost entirely on the technology itself. It has paid little attention to the realities of election administration, including the environment in which this technology is implemented. That includes the calls for a contemporaneous paper replica (the so-called voter verified paper trail), which have largely disregarded the real-world administrative problems with such a fix, let alone whether it will actually prove effective. And almost no attention has been paid to the voting rights of people who are hurt by our current paper-based systems -- most notably, people of color, linguistic minorities, and voters with disabilities.
A notable exception is this article earlier this week by Johnathan Briggs of the Baltimore Sun. (Registration required.) It's one of the most balanced stories I've read on the subject. Mr. Briggs addresses the concerns that some have raised about electronic voting. But he also notes the election administration concerns raised by the proposed paper replica. Most importantly, he discusses the civil rights dimension of the controversy, including the advantages of electronic voting for people with disabilities.
We need more coverage of this nature, rather than the partial perspective that has been offered by such papers as the New York Times (see here) and, even more egregiously, by Wired.com (see here).