Sticking with Punch Cards in Ohio
Most of Ohio's 88 counties will use "hanging chad" punch card machines in November, even thought that equipment has a demonstrated record of denying votes. According to Ohio Secretary of State Kenneth Blackwell: "Sixty-four counties will use [punch-card] equipment that is unfit, unreliable, and unfair to the disabled." (Disclosure: I'm co-counsel with the ACLU in a lawsuit against Blackwell and others challenging Ohio's continuing use of punch card voting machines.)
The Toledo Blade reports that Blackwell made this statement while addressing county election officials on Wednesday. He also criticized the Ohio legislature for passing a law to require electronic voting systems to produce a contemporaneous paper replica:
Addressing the officials for the last time before the November election, he also criticized the General Assembly for passing a law requiring new touch-screen voting machines to be fitted with printers by May, 2006, that can produce a paper receipt for voters to inspect before casting a ballot, saying the printers would be new, untested equipment for which there are no state or federal standards. The new requirement, he said, has "injected uncertainty" into buying machines.See here for my earlier discussion of the Ohio legislation.
That uncertainty, he said, caused many Ohio county elections boards to use old punch-card voting systems in the November presidential election instead using new touch-screen machines that might need to be retrofitted with printers. . . .
"This new standard established in the 11th hour that hasn't been proven and that has yet to be successfully designed and implemented has sort of gummed up the works," he said of the new voter receipt requirement. "There's this notion, which is cut out of fiction, that what we were considering in Ohio in terms of the new voting technology is to be breakthrough, untested, untried technology."
In related news the A.P. reports that most of the 31 counties authorized to switch to electronic voting machines in November have decided not to do so. Only five counties have decided to make the switch to electronic voting. The A.P. story incorrectly reports that: "All 88 counties will be required to have electronic voting systems by the May 2006 federal deadline." There is no requirement, under either federal or state law, that all counties switch to electronic voting machines -- although counties will be required to have at least one accessible voting machine for disabled voters in place by then.
The A.P. story also erroneously refers to the paper records required by the new Ohio law as "receipts." In fact, the Ohio legislation prohibits giving voters receipts that they could take with them, since that could lead to vote buying. See here. What is instead required by Ohio's legislation is a contemporaneous paper replica of the voted ballot that would be kept by election officials.
My take: As Blackwell notes, the move to require a CPR has led directly to counties choosing to stick with "hanging chad" punch card machines. They've done so despite the fact that these machines result in more lost votes than other systems, and have a particularly negative impact on voters of color and people with disabilities. The great irony here is that the move to require a CPR is largely driven by left-leaning folks, who learned the wrong lesson from 2000 -- namely that a recount is a good in itself. In fact, recounts are a means to an end: more accurate and reliable election results. That end is not served by requiring a CPR. To the contrary, the result of this requirement will be that counties stand pat with equipment that's been proven unreliable.
As I've said before, Ohio could well be the next Florida. If there's a close election here, and Democrats come out on the short end of the stick, they'll have only themselves to blame.
Nevada's Experiment with the CPR
The State of Nevada will be undertaking a high-stakes experiment during this November's presidential election. As reported in Wired.com, Nevada will be the first state to attempt implementation of a direct record electronic voting system with a contemporaneous paper replica (aka, "voter verifiable paper audit trail" or "VVPAT").
That's assuming that the DRE-with-CPR system that Nevada plans to use -- a Sequoia touchscreen with a printer add-on -- gets certified. The Wired.com story reports that the system malfunctioned during federal testing last week.
Wired.com's story opens by asserting that many county officials are hoping that Nevada's experiment with the CPR will fail. I don't know whether that's the case, but it's safe to say that many election officials are dubious about it's prospects for success. The story correctly notes that "county election officials oppose the idea, saying printers will create more problems for poll workers if they break down or run out of paper, and they will cause longer poll lines if voters take more time to check their ballots." It's for this reason, among others, that many people (including me) have opposed requiring a CPR.
The story quotes Stanford computer scientist David Dill, one of the most prominent and articulate critics of paperless DREs -- and advocates for the CPR/VVPAT. Dill acknowledges that Nevada's attempt to use the DRE-with-CPR "should be considered an experiment." He nevertheless asserts that: "If the printers fail, it will be because of problems with the implementation (of a paper trail), not problems with the concept. "
My take: Dill's right to characterize Nevada's decision to use a DRE with an attached printer as an "experiment." The problem is that he and others are seeking to require that all electronic voting systems to have this experimental device, which has yet to be proven workable in any real election. That's a big mistake.
Elections don't take place in a vacuum or in a laboratory. They take place in the real world. Just because an idea -- like the CPR -- looks great on paper doesn't mean that it will actually work in the real world. Unfortunately, CPR advocates have demonstrated a remarkable obliviousness to, if not outright disdain for, the realities of election administration.
It's illusory to draw a distinction between the "concept" of a printer attachment and its "implementation," as Dill attempts to do. A concept doesn't do us any good, unless it can actually be implemented . . . and it remains to be seen whether the CPR is either workable or effective.
More Problems in Florida?
This time, the reported problems are with the iVotronic DRE system used by Miami-Dade and Broward counties, according to this story appearing on law.com. (See here for Rick Hasen's post today linking to the story.) Law.com reports that an October 2003 internal memo from a Miami-Dade election official indicated that the system's "audit log" failed to account for 162 votes.
The audit log in the iVotronic machines is supposed to record all activity that occurs on the touch-screen voting machines; it can be used to verify total vote counts. Critics say that if the audit log does not work, the credibility of the election process can be thrown into doubt because there is no other way to verify that all votes cast were recorded.The machine's manufacturer, ES&S, asserts that all the votes were correctly tabulated.
"The audit log should verify that the machine is functioning properly and has been used properly to record the votes," said University of Miami law professor Martha Mahoney, a member of the election reform coalition. "If some machines are missing from the logs, that verification cannot be made for those machines."
DOJ Enters HAVA Consent Decree with California County
The United States Department of Justice announced yesterday that it has entered into a consent decree with San Benito County, California, in what appears to be the first action to enforce the Help America Vote Act of 2002. See here for the A.P. report. In a lawsuit filed in the U.S. District Court for the Northern District of California, DOJ alleged that the county failed to post notice of HAVA's requirements, to provide notice of voters' right to cast provisional ballots, and to translate election materials into Spanish. According to BNA's Money & Politics report, the consent decree requires San Benito to post notice of HAVA's requirements and to provide assistance to Spanish-speaking voters.
Civil Rights Groups Identify Top Five Election Risks
The League of Women Voters and Leadership Conference on Civil Rights today announced their list of the "Top Five Risks to Eligible Voters in 2004":
1. Voter registration problemsThe League and Leadership Conference note that most voters will continue to use the same voting systems they used in 2000 -- and that we can therefore expect to see many of the same problems that we saw in that election. It identifies poll worker training, voter education, and management safeguards as ways of minimizing the negative impact of such difficulties. The accompanying press release lists the NAACP, National Association of Latino Elected Officials, and American Association of People with Disabilities as among the groups signing on to the top five list.
2. Erroneous purging of voters
3. Problems with the new I.D. requirement imposed by HAVA
4. Difficulties with voting systems
5. Failure to count provisional ballots
My take: The League of Women Voters and Leadership Conference on Civil Rights have been consistent sources of wisdom in the post-2000 efforts to upgrade the infrastructure of our democracy. This statement is no exception. With all the controversy over electronic voting and paper trails, it's easy to lost sight of problems far more likely to result in lost votes -- particularly among traditionally disenfranchised populations such as people of color, voters with disabilities, and non-English proficient citizens.
Web Resources on HAVA
If you're looking for information on the Help America Vote Act, two good places to look are the Federal Election Commission and U.S. Department of Justice webpages on HAVA. The FEC page includes the text of HAVA, a guide to HAVA acronyms, deadlines for specific HAVA requirements, and a list of the Election Assistance Commission's responsibilities. The DOJ webpage includes letters to states providing interpretive guidance on HAVA's requirements in such areas as disability access.
The Election Assistance Commission also has its website up and running. Although parts of it are still under construction, several new items have been added in the past few weeks. Among this is this page featuring notices and agendas of the EAC's meetings.
Virginia Commission to Study Electronic Voting Problems
The Washington Post reports that the Virginia General Assembly has approved a resolution creating an 11-person commission to look into reported problems with electronic voting. They will join California, Maryland, and Ohio, which have commissioned similar reports. Fairfax County experienced problems with its electronic voting system in November, although the machines functioned better in February, as discussed here.
According to yesterday's Post, the commission could recommend statewide reforms to verify the accuracy of electronic voting. The story also quotes the co-author of Dirty Little Secrets, a 1996 book on corruption in American politics:
Larry Sabato, director of the University of Virginia's Center for Politics, said he believes no one will find a completely satisfactory solution to the problems confronting electronic voting. "Frankly, I don't know what they're going to find that everyone else hasn't already found," Sabato said.My take: I tend to agree with Sabato. Nevertheless, inquiry into the problems that occurred in Fairfax probably can't hurt, so long as the commission goes about its work in a fair and evenhanded manner.
"Even with paper ballots, as we learned in the 19th century, ballot boxes are often stuffed," Sabato said. "There is no foolproof system."
Video of EAC Hearing Available on C-SPAN.org
Now available on C-SPAN.org is the May 5, 2004 hearing before the Election Assistance Commission on electronic voting. Click here to get to C-SPAN's homepage, and then type "electronic voting" in the video search box in the upper right-hand corner to get to the link. See here and here for discussion of that hearing.
San Bernardino County to Join Disabled Citizens in DRE Lawsuit
San Bernardino County has announced that it plans to join disabled citizens and Riverside County, in their lawsuit against California Secretary of State Kevin Shelley. See here for the county's press release. The lawsuit (Benavidez v. Shelley) challenges Shelley's decertification of the direct record electronic (DRE) voting systems used in San Bernardino, Riverside, and twelve other California counties.
The plaintiffs in Benavidez allege that Shelley's decertification order violates the rights of disabled citizens to vote independently. They also allege that Shelley exceeded his authority under state elections law. See here and here for earlier discussion of the Benavidez lawsuit. A preliminary injunction hearing is set for June 1, 2004.
San Bernardino's press release states that the county tried unsuccessfully to negotiate an agreement with the Secretary of State's office for two weeks, before deciding to join Riverside and disabled citizens in their lawsuit.
Update: See here for coverage of San Bernardino County's decision.
Breaking News: Florida Lawsuit Challenging DRE's Dismissed
Rep. Robert Wexler's lawsuit challenging the use of paperless electronic voting has been dismissed, according to this A.P. Story. U.S. District Judge James Cohn reportedly dismissed the lawsuit because it "would require the federal courts to become deeply involved with election procedures, which typically are left to the states." A prior state court action brought by Rep. Wexler was also dismissed. See here for my earlier discussion of Wexler's lawsuits.
EAC to Hold Second Hearing on Voting Systems
The Election Assistance Commission will hold another hearing on voting systems on June 3, 2004 in Chicago. Its first hearing in May focused on direct record electronic (DRE) voting, see here. This one will focus on other systems. BNA reports:
The Election Assistance Commission is set to hold its second hearing June 3 in Chicago to examine issues related to optical scan, punch card, and lever voting machines, as well as provisional ballots.
The planned session follows the EAC's initial public hearing in early May, which focused on the controversy over security and reliability of electronic voting systems....
The EAC guidelines are expected to be written upon completion of the hearing process. A statement from the commission said the Chicago hearing will feature two panels of witnesses--the first on the various types of voting machines and the second on provisional voting.
My take: It's encouraging to see that the EAC does not plan to limit its focus to the problems with electronic voting. Too much of the debate over voting technology has proceeded in a vacuum. More specifically, there has been scant analysis of how the risks and benefits of electronic voting compare to other systems. DRE critics have assumed that the devil we don't know (i.e., tampering with DREs) is worse than the devil we know (i.e., lost votes and other problems with existing paper-based voting systems). The trouble is, there's been insufficient attention to how bad the devil we know really is. This hearing should provide an opportunity for deeper examination of this issue.
It's also encouraging to see the EAC take a look at provisional voting. In the forthcoming election, this could be a very important way of preventing the loss of votes ... due to faulty voter registration records and other problems. Indeed, a well-functioning provisional voting system may in the end make a bigger difference, in terms of votes saved or lost, than the voting technology used. For a very helpful discussion of provisional voting, see this report from the League of Women Voters.
Shelley "Confident" That 10 Counties Will Be Allowed to Use DREs
That according to this A.P. story, which reports on California Secretary of State Kevin Shelley's statements before the Senate Elections & Reapportionment Committee. Shelley stated that he was in discussions with the 10 counties whose direct record electronic (DRE) voting machines he had provisionally decertified last month. See here for the L.A. Times coverage, and here for a prior post on the situation in California.
The A.P. reports Shelley as stating, "I'm very confident that all the counties eventually will come in." Less clear is the fate of the four California counties using the Diebold AccuVote TSx, which have been banned from using this system in November.
The Onion's Take on Electronic Voting
America's Finest News Source features this infograph on electronic voting. Only slightly less realistic than some of the conspiracy theories relied upon to attack electronic voting. I especially like this part: "Electronic voting machines could potentially be tampered with unlike paper ballots."
More from the NY Times
The Sunday New York Times features this story on the electronic voting controversy. It's better than most of their coverage, though not nearly as thorough as the piece that appeared on its Technololgy pages last week, discussed here. Although the article does mention the practical problems with the contemporaneous paper replica (or voter verified paper trail), it disproportionately quotes the most fervent paper trail advocates -- such as Rush Holt, Kevin Shelley, Kim Alexander, and David Dill -- while giving short shrift to those on the other side of this debate. And yet again, the NYT ignores the civil rights benefits of electronic voting technology, especially for people with disabilties and non-English proficient voters.
More on California's Electronic Voting Controversy
CNN News offers this story on the latest developments in California. Last month, Secretary of State Kevin Shelley decertified the Diebold AccuVote TSx system used by four California counties. He also decertified the electronic voting systems used by 10 other counties, unless they meet 23 specified conditions. According to CNN, Shelley stated during a Senate committee hearing: "I'm very confident that all the counties eventually will come in." It goes on to report, however, that:
[F]our counties -- San Diego, San Joaquin, Kern and Solano -- remain banned from using their machines this fall as punishment for manufacturer Diebold Election Systems, which the state says lacked federal approval for machines used March 2. (Voting machine business stormy for Diebold)
Shelley said the early "disappointing" reaction among several counties to his 23 conditions is "changing now" -- including in Riverside County, which filed a federal lawsuit last month over Shelley's order.
Shelley declined comment about the lawsuit, but said: "I've had good positive conversations with members of the board of supervisors of Riverside County, as well as the registrar. I'm fully confident we'll all be in this together."
Reading between the lines, it sounds as though Shelley is making some constructive efforts to soften the hostility that appears to have developed between his office and the counties that are subject to his directives. Let's hope they can work things out. If they don't, the risk of a disorderly and dysfunctional election is very high -- and of course, the big losers will be voters.
Civil Disobedience by Disability Rights Activists
In New York City, disability rights activists blocked the door to a hearing room, to protect a state panel's failure to address the need for accessible voting equipment. According to the N.Y. Times,:
[P]rotesters said they had been disappointed that the legislators, on a bipartisan joint committee studying ways to fulfill the requirements of the Help America Vote Act, had yet to publicly discuss access to voting machines for the disabled. They want at least one voting machine in each polling place to have an audio prompt for the blind and an attachment allowing quadriplegics to cast votes with their breath, using a sip-and-puff system.The NYT further reports that protesters chanted "Voting is a civil right! Give us access now!" as they blocked the doors. Two were arrested for disorderly conduct.
Mike Godino of the Queens Independent Living Center Inc., who helped organize the protest, said the Assembly had ideas on the matter that might work.
"They're not providing us the ability to access the machines,'' Mr. Godino said. "We wanted to get them to come to the table and start to discuss it. They've been ignoring it and they said that today was supposed to be the last meeting."
N.Y. Times Story on Electronic Voting
Today's New York Times offers this story on the controversy surrounding Direct Record Electronic (DRE) voting equipment. It's the most thorough and balanced coverage we've seen yet on this issue from the NYT. The story is co-written by Robert Lemos and Paul Festa of CNETNews.com.
Here's a sampling:
Riverside County, Calif., has bet millions of dollars on high technology to secure the future of its elections, but officials are now discovering that there won't necessarily be a payoff anytime soon.
The county, in Southern California northeast of San Diego, signed on early to adopt touch screen voting machines, known officially as direct recording electronic, or DRE, systems. But despite using them in 29 elections over four years, Riverside County may not be able to use them the next time around, because of objections from state election officials.
Bottom line: The battle threatens confusion at the polls, leaving voters as the ultimate victims.
My take: I've been very critical of the NYT previous coverage of this issue (see here), but this story is definitely worth reading. Rather than simply parroting the claims of the anti-DRE technolgists, this article quotes people who actually know something about election administration. While addressing the legitimate security concerns that have been raised, it also addresses (at least briefly) the concerns of disability rights activists who stand to benefit from DRE technology.
Let's hope the N.Y. Times editorial page finally catches up, and recognizes that improving our voting systems is a lot more complicated than just strapping printers onto voting machines.
NIST Report on Accessibility and Usability
The National Institute of Science and Technology has issued a lengthy report to Congress, summarizing its plan for developing voluntary standards to encourage more user-friendly systems. The report can be found on the Election Assistance Commission website, along with the EAC's first annual report.
Roll Call describes the NIST report as follows:
The 97-page report, written by the Commerce Department's National Institute of Standards and Technology, includes 10 specific recommendations for developing guidelines on everything from better ballot design and voting-facility layout to uniform testing procedures that measure the "accessibility and usability" of voting equipment.
But the document also warned that there is a "lack of specific research on usability and accessibility" - particularly with regard to voters with disabilities - and added that until more applied research is done, there is "little basis upon which to include many detailed specifications."
"Accessibility has been addressed by generic design standards that intended to remove barriers to access, but usability by persons with disabilities has not been addressed by research," the report stated. "In fact, we know very little about users' experiences with voting systems including those people with disabilities."
The technical report further suggests that to address unaided use of voting machines by persons with disabilities, federal standards "must address the removal of physical and cognitive barriers to accessibility" and should outline unambiguous design requirements for vendors, "particularly in the area of external physical requirements."
Roll Call also quotes EAC Chair DeForest Soaries as saying that his commission has adopted all ten recommendations, but needs funding to move forward. See here for my thoughts on the EAC's request for additional funding.
Legislators Ask GAO to Review Electronic Voting
That according to this AP story, which reports that the "lawmakers said that while existing data indicates that the machines can be more accurate than outdated punch card voting machines, experts are becoming increasingly concerned that many of these electronic voting machines have other flaws."
California Secretary of State Clarifies Decertification Order
California Secretary of State Kevin Shelley has issued a "clarification" of his order decertifying electronic voting systems, unless 23 conditions are met. Among the points of clarification are:
- In the event of an unexpected failure on election day, "paper ballots must be available at all polling places as a back-up." Provisional ballots may be used for this purpose, and the Secretary states that he's "committed" to ensuring that counties don't bear the costs of this.
- Permanent records of each ballot cast must be kept, two copies of which (on CD-R or comparable formats) must be sent to the Secretary of State's office.
- Parallel monitoring will be conducted, with costs to be borne by the Secretary of State's office.
- A technical security plan must be submitted to the Secretary of State's office.
- All components of the system must pass federal and state testing.
- Official results should not be received "through a public telephone."
- Wireless connections and internet connections may not be part of any electronic voting system.
- A poll worker training plan must be in place and submitted to the Secretary of State's office.
This clarification doesn't appear to be posted on the web, but other documents -- including the original decertification order -- may be found on the Secretary of State's website here.
For discussion of the reverberations of Shelley's decertification decision, see this story from the Christian Science Monitor (reprinted in the Anchorage Press).
California Counties vs. Shelley
The fight between California Secretary of State Kevin Shelley and the counties subject to his electronic voting ban continues to become more acrimonious. According to this story in the San Francisco Chronicle: "Officials in most of the 14 counties doubt they can comply with Shelley's order before they must make final decisions on how to conduct the election by early June." Three counties (Riverside, Kern and Plumas) have now sued Shelley over his order decertifying electronic voting systems -- despite the fact that they were shown to be 100% accurate in parallel monitoring conducted during the March election. See here for my prior commentary on this battle.
Recent Editorials on Electronic Voting
In the past few weeks, we've started to see some more balanced editorials on the electronic voting issue, apparently prompted by the EAC's deliberations. Some newspapers are starting to recognize the civil rights advantages of touchscreen technology -- especially for people with disabilities -- and the practical problems with the contemporaneous paper replica (CPR) or "voter verified paper audit trail," that some seek to require. The St. Petersburg Times offers this editorial, noting the irony that the problems with paper-based voting in Florida sparked the call for election reform in the first place, but now have led to some demanding a paper replica of electronic ballots. The SPT editorial is critical of Diebold, yet calls for the more breathless critics of electronic voting to "calm down" in light of the facts that concerns of tampering are thus far hypothetical:
Electronic voting is evolving, and new technology could in time address the concerns of critics who are demanding a paper trail. Meanwhile, it makes no sense to attempt to avert a perceived crisis by acting hastily and creating a real one.Other newspapers have also begun to recognize the problems with requiring a CPR. The Washington Post, while expressing nervousness about possible breakdowns or tampering, notes:
About 30 percent of registered voters are expected to use the machines in 2004, and the systems offer many advantages. When functioning properly, they're more accurate than such alternatives as optical-scan or punch-card ballots; voters can be alerted to make certain that they intended to leave a space blank, or prevented from marking two choices in a single race. The machines offer improved access for voters who don't speak English or who are blind (audio hookups give blind voters a truly secret ballot).It also expresses skepticism about whether it's possible to implement a CPR by November, given the practical problems with paper.
The Denver Post offers a similar perspective. It notes the growing alarm in some quarters over electronic voting, while acknowledging that paper presents its own difficulties:
Paper, though, might be a step backward. Improving machines to allow voter verification of selections before recording the vote, secure ways to store a specific machine's results in case of dispute, and not connecting machines to central databases until after polls are closed might make e-voting more secure.
The federal standards should create a system as foolproof and tamperproof as possible to restore public faith that every vote does, indeed, count. Anything less undermines the American political system.
EAC Seeks Additional Funding
The newly constituted Election Assistance Commission is asking Congress to double its funding, according to this A.P. story and this story from electionline.org. President Bush's budget would give the Commission only $10 million and the EAC seeks $10 million more, so that it can provide states with guidance on voting systems by 2006. In addition, the President's proposed budget would provide only $40 million to the states for election improvements -- a small fraction of the $600 million authorized by HAVA, according to electionline.org.
Amazingly, at least one congressman appears to be balking at making additional funds available to the EAC. According to electionline.org, Appropriations Subcommittee Chair Ernest J. Istook Jr. (R-OK) "asked the panel why, if the November elections are only months away and if the states are already spending federal money on voting systems, the research would have any meaning this year." EAC Chair informed him that the process of conducting research and promulgating standards is ongoing, and wouldn't end with this year's election.
The story also reports on the slow start that the EAC got off to. Although created by the Help America Vote Act of 2002, the four commissioners weren't even appointed until December 2003 -- and only $1.2 million was appropriated for its operations in FY 2004. The A.P. story also reports that $2.3 billion for voting improvements is, at long last, set to go out to states this week.
My take: Whatever one feels about the risks and benefits of electronic voting, I hope we can agree that responsible election reform can't be done on the cheap. Congress and the President must provide the Election Assistance Commission with the money it needs to do its job properly. And they must make available to the states the money that HAVA promised for improvements to our decrepit election systems.
Electronic Voting in India
While the controversy over electronic voting in the U.S. rages, 380 million people in India have voted electronically over the last three weeks, according to this A.P. story. The story reports that:
To avoid software glitches, viruses or hacking, the Indian machines use basic software and are not networked. They are different from voting machines manufactured by private companies in the United States, for example, which resemble ATMs and have touch-screen monitors, smart-card readers and more sophisticated software.It also reports on some security concerns similar to those that have been raised in the U.S. . . . and others of a much different variety:
In a country where politicians routinely hire goons to bolster their count, newspapers have been full of reports of thugs taking away voting machines and tampering with booths. In one case, henchmen were able to enter 50 votes for their candidate because one of them knew how to run the control machine. Some of these incidents led to repolling in 1,879 voting booths throughout India on Wednesday.
The Electronic Corp. of India, one of the two government-owned companies that manufactured the voting machines, said failures and vandalism were "negligible," given the scale of the elections.
More on Diebold
CNN offers this story on Diebold's troubles.
California County to Defy Electronic Voting Ban
San Bernardino County plans to continue using electronic voting, despite Secretary of State Kevin Shelley's order decertifying these machines, according to this story in today's N.Y. Times. Another California county, Riverside, has joined with disabled voters in a lawsuit against Secretary of State Shelley, challenging his decertification decision. (See here.) A preliminary injunction hearing is set for June 1 in that case.
My take: It's become clear that Shelley's decertification decree, far from restoring order to California's plans for November, has thrown things into disarray. It's one thing to condemn bad behavior on the part of vendors who've acted badly, as Diebold allegedly has. It's quite another to punish voters in 14 counties, by decertifying the equipment to be used in November -- including those systems which have worked well.
Diebold Head Admits Mistake
Wally O'Dell, the CEO of Diebold, admitted to making a "huge mistake" by expressing support for President Bush's reelection campaign, the N.Y. Times reports. O'Dell says he's now dropped out of any partisan political activities.
My take: I'm glad that O'Dell has admitted his mistake. Unfortunately, his statement did a great deal of harm -- not just to his company, but to our voting system. It's fueled conspiracy theorists' claims that the use of electronic voting is a ruse by which Republicans plan to steal votes in November. The reality, however, is that converting to electronic machines would prevent the loss of votes, especially in minority communities that are especially hard-hit by the "hanging chad" punch card still used in many places. It's surprising to me that so many Democrats have bought into the anti-electronic voting hype, which threatens to do the greatest harm to that party's core constituents, such as people of color, immigrants, and people with disabilities.
Florida Judge Hears Argument on Electronic Voting Lawsuit
On Monday, Federal District Judge James Cohn heard argument on Rep. Robert Wexler's lawsuit to prohibit the use of paperless electronic voting in Florida's election in November, according to stories in the Sun-Sentinel and Palm Beach Post. Wexler claims that electronic voting violates the Equal Protection Clause of the U.S. Constitution, by denying voters the right to have their votes manually recounted. The state's attorney labelled Wexler's lawsuit "a near midnight-hour rush to fix something that isn't broken."
The Sun-Sentinel story notes that Wexler filed a lawsuit in state court, making similar claims under Florida law, which was dismissed in February. For my previous commentary on the state lawsuit, see here.
According to the Palm Beach Post story, Judge Cohn stated that federal courts are supposed to intervene in state election disputes only as a "last resort," and asked "How can this action filed in federal court be a 'last resort' if an almost identical case is pending in state court?" He's expected to issue a ruling within two weeks.
My take: Rep. Wexler's lawsuit should be dismissed ... not because of the parallel state lawsuit, but on the merits. If there were a violation of equal protection, or some other federal law, I would agree that dismissal of Wexler's action would be improper. Federal courts have a responsibility to adjudicate cases raising violations of federal rights, subject only to narrow exceptions.
Wexler's problem is that he can't show that the use of punch cards violates equal protection. Although his complaint is not terribly clear, his claim appears to be that the use of paperless electronic voting machines prevents touchscreen voters from having their votes recounted, in the event that such a recount is necessary. But the reasons that often necessitate a recount with paper-based systems -- such as hanging chads and stray marks -- don't exist with electronic voting. What's significant, for equal protection purposes, is whether votes are counted accuracy. Yet Wexler's complaint is devoid of any factual allegations that would show electronic voting to be less accurate than other systems.
Moreover, the relief that Wexler is seeking would do far more harm than good. His goal is to require counties using electronic voting systems to implement a contemporaneous paper replica. But given that there's no system capable of doing this that's certified and ready for use, he seeks to force these counties to use paper-based voting systems. This would require them to implement optical scan voting between now and November. The consequence would not only be to squander about half of the $133 million in HAVA money allocated to the State of Florida, but also to put counties in an impossible time crunch. Procuring and getting in place a new voting system is no simple matter. It takes months of preparation, including procurement, poll worker training, testing of equipment, and voter education. Though Wexler may not realize it, attempting to put in place a new voting system in this compressed time frame is a recipe for disaster.
Visually Impaired Voters Sue State of Ohio
The National Federation for the Blind has brought suit against the Controlling Board and Secretary of State of Ohio, challenging the failure to provide accessible voting technology. A press release issued by the NFB indicates that the complaint states claims under the Americans with Disabilities Act and the Rehabilitation Act. According to this release:
The complaint argues that, although the barriers imposed by the current voting systems and ballots can be readily eliminated through the use of existing technology or through modifications to existing technology, the Controlling Board has refused to approve the request of the Secretary of State and the Defendant counties to use federal funds available to Ohio to purchase accessible voting systems that use this technology because of a dispute over methods to achieve ballot security.
“Obviously the National Federation of the Blind wants a secure voting system,” said NFB President Marc Maurer. “But the Federation believes that electronic voting systems have been proven to be more secure than either punch-card or lever voting machines. The prolonged security debate only continues to disenfranchise blind voters,” Maurer said.
As mentioned yesterday, voters with disabilities have also brought suit in California to demand accessible electronic voting technology. The Ohio NFB lawsuit is discussed in this editorial from the Cincinnati Enquirer, which urges Ohio to stop dragging its feet on conversion to accessible electronic voting equipment.
Disabled Voters Sue California Secretary of State
A group of disabled voters, along with the County of Riverside, have brought suit against California Secretary of State Kevin Shelley for his decision to decertify electronic voting systems used in that state, according to this A.P. story. The L.A. Daily Journal (not available on the web) reports that a hearing is set for May 24, 2004, at which plaintiffs will seek a T.R.O. Electronic voting machines are the only certified voting technology that allows people with disabilities to vote independently. Many of the plaintiffs in this new lawsuit previously sued Shelley and several registrars, for failing to provide enough accessible electronic voting to them, as discussed here.
Update: A preliminary injunction hearing is set for June 1, 2004, before Judge Cooper in the Central District of California.
Ohio Governor Signs Electronic Voting Bill
Governor Bob Taft of Ohio has signed a bill to require a contemporaneous paper replica of electronically cast ballots by 2006, and to allow 31 counties to move to paperless electornic systems by 2004, according to this A.P. story. It's not clear how many, if any, of those 31 counties will actually make the switch to electronic voting. Ohio is among the states in which the "hanging chad" punch card remains most heavily used. In 2000, 69 of 88 counties used this system. (Disclosure: I'm co-counsel with the ACLU in a lawsuit challenging Ohio's continued use of punch card voting.)
California Senate Committee Votes to Ban Electronic Voing
A California Senate Committee has voted 3-1 in support of a bill to ban electronic voting , according to this A.P. story. Secretary of State Kevin Shelley has already decertified one touchscreen voting system, and has provisionally decertified all the others used in the state -- although, as discussed here that decision will be challenged in court. The bill will next move to the Senate floor.
More News on Yesterday's EAC Hearing
Yeterday, the Election Assistance Commission heard testimony on the election voting controversy, as discussed here. Fore more coverage yesterday's hearing before the Election Assistance Commission on the topic of electronic voting, see stories in the Cleveland Plain-Dealer, Newark Star-Ledger, and K.C. Star.
Daily Journal Op-Ed
Below are excerpts from an op-ed that I wrote, which was published in today's Daily Journal, a California legal newspaper. It responds to California Secretary of State Kevin Shelley's decision on Friday to decertify electronic voting systems, discussed here.
California and the rest of the country are in the middle of a once-in-a-generation transformation of our voting systems. A critical component of that transformation is the shift to electronic touch-screen voting.
Thanks to a federal court order, California has abolished the punch-card voting equipment that caused so many problems in 2000 - and disproportionately harmed African-American, Asian-American and Latino voters. Punch cards performed especially poorly in the California recall election, resulting in 170,000 lost votes statewide. It was only the large margin by which the recall succeeded that saved the state from a Floridalike catastrophe in October.
From a voting rights perspective, electronic touch-screen voting is vastly superior to the paper-based equipment that it is replacing. Touch-screen voting results in far fewer lost votes than the infamous "hanging chad" punch-card machine that, until this year, was used by half of California voters. Touch screens provide particularly great advantages for traditionally disenfranchised voters, such as people with disabilities and non-English speakers.
Unfortunately, many Californians will be denied the benefits of touch-screen voting in the coming election. Thanks to a recent decision by Secretary of State Kevin Shelley, citizens in at least four counties will be forced to use paper-based voting systems that may result in many more lost votes in November . . . .
Much of the fear over touch-screen voting rests on a lack of information. Contrary to popular belief, touch-screen machines are not hooked up to the Internet. They are instead stand-alone units, with redundant internal backups. Speculation that a clever hacker somehow could infiltrate the system on election day is the stuff of conspiracy theories. Even in the event of a power loss, votes cast on touch-screen machines can be preserved - and counted - through the redundant internal backups . . . .
Voting equipment is only one part of our election system. A widely respected CalTech/MIT report found that, while technology was responsible for many lost votes, an even greater number of votes was lost because of registration and polling-place problems.
Nor is paper any guarantee of election integrity. In Napa County, for example, the electronic system functioned very well; on the other hand, serious problems occurred with the optical-scan system used for absentee ballots. This incident demonstrates that paper-based systems are not immune from error. Paper is equally vulnerable to fraud and error if adequate procedures are not followed.
The myopic focus on software and hardware has, unfortunately, sidetracked us from far more likely sources of lost votes. If we really want to avoid lost votes in November, we should focus on implementing better procedures, training poll workers, and educating voters, rather than implementing a false fix like the CPR [or "voter-verified paper trail"] that is likely to cause far more problems than it solves.
Election Assistance Commission Meeting on Electronic Voting
The Election Assistance Commission held a public meeting in Washington, D.C. today regarding electronic voting. See here for the announcement of the meeting. Among those scheduled to testify in person were California Secretary of State Kevin Shelley, Jim Dickson of the American Association of People with Disabilities, L.A. County Registrar Conny McCormack, Prof. Avi Rubin, and Kay Maxwell of the League of Women Voters. For early coverage of the hearing, see this Reuters story and this story from IDG News. The League of Women Voters' testimony is available here. My written testimony to the EAC appears below.
My Written Testimony to the Election Assistance Commission
Below is the written testimony that I submitted to the Election Assistance Commission yesterday.
May 4, 2004
Election Assistance Commission
Attn: Bryan Whitener
Re: Electronic Voting
Dear Members of the Commission:
This letter is submitted for the hearing on electronic voting, to take place on May 5, 2004. I respectfully request that you accept the written comments set forth below and that these comments be made part of the record.
As these comments explain, electronic voting technology provides significant advantages over paper-based systems from a civil rights perspective. Experience and studies that have been conducted over the past several months have shown that it is possible to implement electronic voting safely, if proper procedures are followed. I therefore urge that you resist the calls that some have made to forbid paperless electronic voting during the coming election, a drastic and unnecessary step that would have profoundly negative consequences for countless voters throughout the country and thrust the state into conflict with voting rights laws.
I would also urge against adoption of a requirement that all voting machines generate a contemporaneous paper replica (the CPR or “voter verified paper audit trail”). Given that this device is unproven, and has thus far proven problematic in the few places that have attempted to implement it, the Commission should avoid locking into place any particular hardware “fix” at this time. It should certainly avoid doing so between now and the November 2004 election, given that there is insufficient time for jurisdictions to obtain and implement CPR technology. Instead, the focus should be on ensuring that appropriate procedures are in place for all types of voting technology, including but not limited to electronic voting, to make sure that all votes are accurately counted in November.
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 resulted in decertification of pre-scored punch card machines in California 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’s Moritz College of Law, 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.
The United States is in the midst of major changes in the infrastructure of our democracy, spurred by both legislation and litigation. Although progress since 2000 has proceeded more slowly than some hoped, many jurisdictions have already switched (or are in the process of converting to) electronic voting machines, also known as “Direct Record Electronic” or “DRE” equipment. Touchscreen voting machines are one type of DRE. Approximately 10% of voters nationwide used DRE voting machines in the 2000 elections.
There are three dimensions to the current debate over electronic voting: 1) technology, 2) election administration, and 3) voting rights. Unfortunately, the current 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 central-count 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 cast secret ballots. 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 vote independently. The Help America Vote Act, of course, requires the implementation of accessible technology by 2006. Moreover, a recent federal court decision, American Association of People with Disabilities v. Hood , 2004 WL 626687 (M.D. Fla. 2004), holds that the failure to provide accessible voting equipment violates the Americans with Disabilities Act of 1990.
Electronic voting machines can also help in reducing lost votes. 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.
The “Paper Trail” Debate
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. While there are legitimate security concerns surrounding electronic voting, it is important to separate fact from fiction. Much of the opposition to electronic voting rests on mistaken information about the present state of technology and a failure to understand the security checks that already exist. Some of it is undoubtedly influenced by understandable suspicion wrought by the 2000 election fiasco. And some of the concerns regarding electronic voting are fueled by doomsday scenarios and conspiracy theories widely circulating about the internet.
Contrary to what some media reports have suggested, the current debate is not about whether touchscreen voting machines should be required to have an audit trail, or even a paper audit trail. Contemporary touchscreen systems already have multiple levels of redundant backup and most jurisdictions have strenuous security procedures – including logic and accuracy testing – to ensure that the machines work as intended. The Help America Vote Act (HAVA), moreover, already requires touchscreen machines “produce a permanent paper record with a manual audit capacity ....”
Instead, the present debate is focuses on whether electronic voting machines should be required to print out a paper replica of the electronic ballot at the time of voting. This would require a special kind attached printer capable of generating a paper replica of the electronic ballot, in such a way that the voter could see but not touch. Though this paper replica is sometimes referred to as a “receipt,” it is more accurately described as a contemporaneous paper replica of the electronic ballot, for which I here use the term CPR. Under the so-called Mercuri method (named after computer scientist Rebecca Mercuri who first proposed it and is one of its chief advocates) the contemporaneous paper replica would print out behind a glass or plastic screen, so that the voter can see it but not touch it. This is necessary in order to prevent the voter from pocketing the paper ballot replica, taking it with him or her, and using it in vote-buying schemes.
Far from supporting the call to require a CPR, recent evidence and developments cast doubt on its proponents’ claims that it is a necessary or even effective means to promote secure elections. Before a requirement for a CPR or any other particular device is enacted, three questions should be answered in the affirmative:
1. Is it workable?
2. Is it necessary?
3. Will it solve the problem?
As set forth below, the answer to each of these questions is no. The evidence reveals that 1) the CPR would be much more difficult to implement than its proponents have assumed, 2) the CPR is not necessary to promote security, 3) it is not an effective means by which to promote security.
Is the CPR Workable?
As noted above, there is no touchscreen system currently certified that has the capacity to generate a CPR. While three prototypes (marketed by Avante, Accupoll, and Sequoia) are being developed, none of them has successfully been implemented in any jurisdiction. In the two places of which I am aware that attempted to use a CPR system on a limited basis, it proved problematic at best.
The experience of Sacramento County bears out the practical difficulties in making such a CPR system work. While voters in Sacramento reacted favorably to touchscreen voting, as have voters in other jurisdictions, “[a]dding a printer and paper to the voting process was a challenge.” As the Sacramento registrar reported
It was confusing for some [voters] because they thought they could take it [the paper replica] with them....
If the printed record jams, the machine is out of service until someone can take care of the problem .... A few times when the printed record stuck they had to be extracted with many creative tools that were on hand at the early voting site such as a windshield wiper or back scratcher .....
The voter viewed the printed record through a plastic shield in front of the machine. Voters complained that it was difficult to read because of the length of the ballot, size and darkness of the print and the location of the shield. Most voters wanted to remove the paper copy and check it out before it went back into the machine....
There was concern that the machines would have problems sorting the voter’s printed receipts. It was decided to empty the tray every ten voters. This procedure was stopped. The machines must not be opened during the day to empty the tray. While voters in Sacramento had rave reviews for the touchscreen machines, the printer attachments proved to create more problems than they resolved – requiring that officials use coathangers and other creative devices to dislodge ballots when the printers jammed.
A CPR system was also tested on a limited basis in Wilton, Connecticut’s November 2003 election. After that election, the deputy registrar commented that the “ease of use and human factors … are appalling.” The voting system created “numerous problems for voters and placed great stress on the poll workers ….”
A recent paper by Ted Selker and Jon Goler of MIT assesses the practical problems with the CPR. Their paper "Security Problems and Vulnerabilities with VVPT” (available at http://www.vote.caltech.edu/Reports/vtp_wp13.pdf ) finds that:
[The CPR] complicates two of the top three problems that have compromised more than one percent of American votes in 2000: equipment problems and polling place operations. It complicates the setup, teardown, and operation of the ballot place. It complicates polling place procedures during the vote. It gives extra and difficult tasks for a person to do and increases the problems with user experience and user interface. It also increases the length of time of voting, which makes it, with more steps, easier to make mistakes.
It is unfortunate that so little public attention has, to this point, been focused on the practical problems that Selker and Goler identify. In their haste to propose a solution to the electronic voting security problem that will appease the public, some advocates have failed to consider the real-life problems that have resulted in lost votes in past elections.
Experience and research thus demonstrate that implementation of the CPR is considerably more difficult than some advocates’ public statements might lead one to believe. Touchscreen machines are not desktop computers to which ordinary printers can easily be attached. As noted above, the CPR would require that they print out ballot replicas behind a screen so that the voter can see but cannot touch them.
The difficulties in implementing such a system relate not only to costs – for which no reliable estimates are available – but also to the practical imperatives of election administration. While touchscreens are sometimes compared to ATM machines, ATM’s need not be lugged to and from hundreds of precincts each election day. Voting machines must be transportable. When they break, longer lines at the polling place will result. As numerous reports since the 2000 election have documented, the nation’s polling places are dramatically understaffed, often by elderly poll workers. Requiring the transportation and implementation of an add-on device would considerably complicate an already complicated process, and further tax scarce poll worker resources. Because of the practical difficulties and uncertain costs of implementing touchscreens with a CPR, the likely result of imposing such a requirement is to force counties to stick with inferior paper-based systems such as the discredited “hanging chad” punch card.
While further experimentation is warranted, the CPR has not yet proven workable, and it would be grave mistake to require this device at this time – and certainly to require implementation by the November 2004 election. If a CPR is required, it is likely that counties will forego electronic voting systems entirely, given that the CPR has yet to be proven workable in any real election. Counties that have already moved to DRE voting systems would be placed in a particularly difficult position, given that there is inadequate time between now and November to implement a CPR, given the nascent state of this technology. A CPR requirement would therefore have the effect of decertifying systems that many jurisdictions already using – leaving them to scramble to implement new equipment in the seven months between now and the November election. That is far too short a time in which to implement an experimental device, especially if numerous jurisdictions attempt to do so at once thus putting a strain on vendor resources. In short, attempting to impose a CPR requirement at this time is a recipe for disaster.
Is the CPR Necessary?
In the past year, three states have commissioned thorough examinations of the security of existing touchscreen systems: California, Maryland, and Ohio. These studies have found security vulnerabilities in the manner in which touchscreen systems have been implemented – although to this date these security concerns remain hypothetical. Significantly, none of these three studies have recommended that a CPR be implemented. As all of these reports have concluded, there are other more effective means by which to improve touchscreen security, including better procedures.
For example, the California Secretary of State’s office commissioned a report 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.
There is also no evidence to support the conclusion that the CPR is necessary to ensure touchscreen security. The one judicial decision to have addressed the matter, Weber v. Shelley, 347 F.3d 1101 (9th Cir. 2003), rejected the claim that voting rights are violated by touchscreen machines that lack a paper audit trail. 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.
A recent paper by Michael Shamos, Distinguished Professor of Computer Science at Carnegie-Mellon (http://euro.ecom.cmu.edu/people/faculty/mshamos/paper.htm), confirms that the CPR is not the best means by which to promote election integrity. Professor Shamos’ paper, entitled “Paper v. Electronic Voting Records – An Assessment,” finds that the CPR is not an effective remedy for the security risks that exist with DRE’s. He also suggests other measures -- such as better procedures, open source, and parallel testing -- that are much more likely to enhance election integrity than the CPR.
Supporters of the CPR requirement are thus left to argue that, even though it is not needed to ensure security, the CPR should be required in order to enhance “voter confidence.” It was on this ground that California Secretary of State Shelley issued a directive requiring that all touchscreen voting machines have a CPR, notwithstanding the recommendation of the California task force and the opposition of civil rights groups. In so doing, Shelley acted contrary to an Attorney General opinion that requiring a CPR would violate the ADA. But even Secretary Shelley, when stating his intention to require a CPR, recognized that it was not necessary to improve security. His asserted basis for taking this position was “so voters can feel more confident.” As a result, a federal lawsuit has recently been filed, American Association of People with Disabilities v. Shelley, alleging that Shelley’s action violates the ADA and other disability rights laws. This commission should avoid imposing a requirement that would conflict with such civil right protections.
To be sure, the idea of a “paper trail” carries considerable visceral appeal for some citizens. And were the CPR a costless, effective, and workable solution, promotion of “voter confidence” might be a goal worth proceeding. But requiring a CPR is neither costless, effective, nor simple to implement. Indeed, the very spectre of a CPR being required has already frozen in place many counties plans to upgrade their voting systems. Those hurt the most by this freeze are our most vulnerable voters, who stand to benefit most from the implementation of touchscreen technology.
Would the CPR Really Enhance Security?
The argument that the CPR would serve as an effective means of policing fraud and error depends upon the electronically voted ballots being checked against the paper replicas. In the event of a discrepancy, CPR advocates urge that the paper replicas should be used. This argument thus rests on three assumptions: 1) that recounts will allow comparison of electronically voted ballots against their paper replicas, 2) that voters will actually have checked and thereby “verified” the paper replicas, and 3) that the paper replicas will be more reliable than their electronic counterparts. None of these assumptions is supported by the evidence or by experience.
Contrary to what some voters may believe after the 2000 Florida elections, full manual recounts of all voted ballots in a jurisdiction are uncommon. What is more commonly done to verify the accuracy of election results is to recount a smaller sample of recounts. California, for example, has a law requiring a manual recount of 1% of voted precincts. A recent study shows that such a manual recount would not in fact provide an effective means of detecting fraud or error with touchscreen voting systems. C. Andrew Neff, Election Confidence (2003). Only by conducting a full recount of at least 250 precincts would a recount verify accuracy at a 90% confidence level – and even then, it would only do so with a 1.2% error margin. Thus, unless we are prepared to conduct a full manual recount of all voted ballots (something that, to my knowledge, no one recommends), the CPR will not provide an effective check on electronic voting results in close elections. And even with a full manual recount, arguments for a CPR requirement assume that the printer-generated ballot replicas will be more accurate than the electronically voted ballots, an assumption that is dubious at best in view of the likely prospect of printers sometimes failing and the vulnerability of paper ballot replicas to fraud and error.
In support of their argument for a CPR, electronic voting critics have pointed to problems that have occurred in the implementation of electronic voting in some jurisdictions. To be sure, no system is invulnerable to error and it is inevitable that there will be some problems in the implementation of any new technology. For example in the March 2004 election, three counties (Alameda, Orange and San Diego) experienced problems in the implementation of new electronic systems. This has led to proposed legislation that would forbid the use of electronic voting systems altogether in the November election. Most recently, California Secretary of State Kevin Shelley has decertified one touchscreen system (the Diebold TSx), and provisionally decertified all the others used in California, contingent on their satisfying specified conditions. However well-intentioned such proposals may be, they ignore the fact that electronic voting systems appear to have worked well in 11 of 14 counties.
Moreover, the CPR would not have done anything to prevent the sort of problems that occurred in these three counties. 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. A CPR requirement, however, would not solve these problems.
None of this should be understood as minimizing the problems that have occurred in some counties that have used electronic voting. 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. Moreover, it must not be forgotten that paper-based systems – including both punch cards and optical scans – are susceptible to fraud and error.
In fact, in one other California 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.
This is not to deny that there are things that can be done to improve voting security. 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.
This position is actually consistent with the “Resolution on Electronic Voting,” which appears on the voterverified.org website. This resolution, signed by many computer scientists is often touted by CPR supporters as demonstrating the need for a contemporaneous paper trail. But while the resolution calls for a voter verifiable audit trail, it does not insist that this audit trail be paper. Even David Dill, perhaps the most prominent critic of contemporary touchscreen systems, has acknowledged publicly that the CPR is not the only way of ensuring “voter verified” auditability. Requiring a CPR thus goes well beyond what even the “Resolution on Electronic Voting” demands. It would stifle innovation by locking us into a false fix for the legitimate security concerns that exist.
There can be no question that safeguarding the integrity of our voting system should be among the EAC’s highest priorities. While we should obviously take seriously the security concerns that have been raised in recent months, the Commission should reject calls to require a CPR or to decertify existing paperless systems. Instead, the EAC 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.
Rather than adopting a hasty and drastic “fix” – such as a CPR requirement -- that is likely to cause more problems than it solves, the EAC should focus on making sure that adequate procedures are in place for the coming election. This includes not only those used to implement electronic voting, but also those used to implement paper-based systems. By so doing, the EAC could enhance the integrity of the state’s voting systems without sacrificing the voting rights of the millions of voters who stand to benefit from electronic voting.
Please do not hesitate to contact me if there is further information that I can provide.
Daniel P. Tokaji
Riverside County to Sue Secretary of State Over Electronic Voting
Just when you thought the California electronic voting controversy couldn't get any hotter . . .
The Riverside County Board of Supervisors voted today to sue California Secretary of State Kevin Shelley, over his decision on Friday to decertify the electronic voting system used by the county. See here for discussion of Shelley's decision. According to this story in the Riverside Press-Enterprise:
After meeting in closed session, the supervisors announced in angry terms they will take Shelley to court to stop his "assault on the touch-screen voting system pioneered by Riverside County," said Roy Wilson, chairman of the board of supervisors. . . .
Riverside County has not decided yet when and where to file the lawsuit, said Bill Katzenstein, county counsel. The county may join forces with other counties and may consult with attorneys who already have sued the state on behalf of disabled people who oppose Shelley's efforts.
A press release issued by the County Board of Supervisors states:
Today, the Board of Supervisors met in closed session to discuss the California Secretary of State's assault on the touch-screen voting system pioneered by Riverside County. In recent orders, Mr. Shelley decided to decertify all touch-screen voting systems in California unless counties complied with certain conditions.
This order was made without any evidence of errors and problems with Riverside County's voting system. That system has been certified for use by state and federal officials and has accurately tallied results in 29 elections.
If Riverside County cannot comply with the conditions of Mr. Shelley's whim before the Presidential election in November, it could cost Riverside County more than $3 million to put a back-up system in place - a system that would be less accurate, more costly, less accessible to disabled voters and that might take weeks to tally Riverside County residents' votes.
This all comes at a time when problems with the state budget already have siphoned off millions of dollars in local funds that should instead be spent on services for Riverside County residents.
As a result, with respect to Board Item A9, with all members present, the Board of Supervisors unanimously voted to initiate litigation to challenge the Secretary of State's orders decertifying direct recording electronic voting systems.
My take: Riverside County would seem to have a strong case against Secretary Shelley. To decertify a voting system, it must "be defective or prove unacceptable," under California Elections Code 19222. In addition: "Six months' notice shall be given before withdrawing approval unless the Secretary of State for good cause shown makes a determination that a shorter notice period is necessary."
While the Secretary of State's office had before it evidence of improper activity on the part of Diebold, there was no such evidence regarding the Sequoia system used by Riverside County. Nor did the Secretary of State have evidence that would suffice to constitute "good cause" for decertifying the Sequoia system on less than six months notice. And all this is aside from the adverse impact on disability rights -- and possible Americans with Disabilities Act violation, discussed here -- that can be expected to result if Shelley's directive sticks.
Michael Shamos Responds to N.Y. Times Editorial
Michael Shamos, Distinguished Professor of Computer Science at Carnegie-Mellon, has submitted a letter responding to the N.Y. Times' recent editorial on electronic voting, discussed here. Unfortunately, the NYT has not published Prof. Shamos' letter. Therefore, with his permission, I reprint it in full below.
To the Editor:
Your April 24 editorial "A Compromised Voting System" contains two serious errors of fact that invalidate its conclusion that paper audit trails should be required for electronic voting machines. First is the unsupported statement that "It is not hard to program a computer to steal an election." It is in fact virtually impossible, not "easy," to do so, and not only has it never been done, no one has even outlined a remotely plausible way in which it might even theoretically be done.
The second statement, referring to very real problems with direct-recording electronic (DRE) machines, that "The answer to all of these problems is a 'voter verified paper trail'." is likewise incorrect. The proposed paper trail does one thing, and one thing only -- it assures the voter that her choices were initially captured correctly by the machine. It provides no guarantee whatsoever that those choices were counted, will ever be counted, or will even be available in the event a recount is demanded. The reason is that the paper trail, unlike an electronic record, must be touched by humans to be processed and can be modified, stolen, lost, damaged, defaced, or augmented by anyone who may come in contact with it.
The U.S. has a sorry history of election manipulation accomplished through tampering with paper ballots. Even though DRE systems have been used here for more than 22 years, there has never been a single verified case of successful manipulation of any such system. That is not to say that it is impossible, but it provides powerful evidence that it is not "easy" to program machines to steal elections, and I question what support the Times has for such an assertion.
Michael I. Shamos
School of Computer Science
Carnegie Mellon University
The writer was statutory examiner of voting systems for Pennsylvania from 1980-2000 and for Texas from 1987-2000.
Too bad the NYT is unwilling to air views that differ from -- and expose the flaws in -- its editorial position.
PC World Article on Electronic Voting
The April 28 issue of PC World Magazine features this article on the electronic voting controversy. There are some points on which they too readily accept the speculative claims of electronic voting critics at face value.
For example, the article states: "The biggest danger of touch-screen machines is that if votes are lost or changed, no voter-verified audit trail is available for a recount, and the evidence of tampering could also be erased." But in the real world, that hasn't been the "biggest" problem with electronic voting, and it's far from treat that the contemporaneous paper replica is the best way of dealing withis problem. In fact, as Ted Selker & Jon Goler have pointed out, the CPR complicates the voter interface in a way that may cause more problems than it solves. See here.
The article also cites the CalTech/MIT reports finding that electronic machines did only slightly better than punch cards, in elections from 1988 to 2000. While this is true, it overlooks the fact that more recent data -- particularly that of Henry Brady, available here and here -- finds much better results for the present generation of electronic voting machines.
Despite these complaints, the PC World piece is one of the most thorough and balanced stories yet on the electronic voting controversy, and well worth reading.
Selker and Goler on Problems with the CPR
Ted Selker and Jon Goler of MIT have posted this paper on the CalTech/MIT Voting Technology Project's website. The report is entitled "Security Problems and Vulnerabilities with VVPT." As the title suggests, the paper considers the practical problems in implementing a contemporaneous paper replica (CPR) or "voter verified paper trail" with electronic voting machines. The paper finds that:
Experiments and elections have yet to establish that people can in fact verify their ballot using a paper receipt. Effective approaches for accurately counting the paper receipts for auditing purposes have not been established either.They conclude that:
The Voter-Verified Paper Trail discussion has diverted attention from the main sources of lost votes in past elections . . . .[The] VVPT complicates two of the top three problems that have compromised more than one percent of American votes in 2000: equipment problems and polling place operations. It complicates the setup, teardown, and operation of the ballot place. It complicates polling place procedures during the vote. It gives extra and difficult tasks for a person to do and increases the problems with user experience and user interface. It also increases the length of time of voting, which makes it, with more steps, easier to make mistakes.Selker & Goler's paper stands alongside that of Michael Shamos, discussed here, in raising questions about the efficacy and workability of the CPR that have yet to be answered by its supporters.
California Electronic Voting Decertification
On Friday, Secretary of State Kevin Shelley issued an order decertifying the Diebold TSx machine for use in the November 2004 election. Shelley also provisionally decertified electronic voting machines used in 10 other counties, although these machines may be recertified unless they 1) implement a contemporaneous paper replica or 2) implement 23 specified security measures are implemented. See here for Rick Hasen's blog entry, which includes links to news stories on Shelley's decision. My thoughts on the controversy are quoted in this story from IDG News Service, and in this post. I'll have more shortly.