Count Crisis?

Daily Business Review (2004-05-13) Matthew Haggman

This Page: https://copswiki.org/Common/M695
Media Link: http://www.law.com/jsp/article.jsp%3fid%3d1084316008117
More Info: Election Integrity

By Matthew Haggman
05-13-2004


A scathing internal review of the iVotronic touch-screen voting machines used in Miami-Dade and Broward, Fla., counties, written by a Miami-Dade County elections official, has raised fresh doubts about how accurately the electronic machines count the vote.

The review, contained in a June 6, 2003, memo that came to light last month, concludes there is a "serious bug" in the voting machine software that results in votes potentially being lost and voting machines not being accounted for in the voting system's self-generated post-election audit.

The Miami-Dade County Commission's elections subcommittee has scheduled meetings today and Friday to discuss the issues raised in the memo.

The memo could cast a new shadow over the credibility of electronic voting as the November presidential election approaches. Electronic voting machines are coming under increasing criticism for being glitch-prone, not providing an adequate way to perform a recount in close races, and being vulnerable to computer hacking and fraud.

In the e-mail memo, Orlando Suarez, division manager of the county's Enterprise Technology Services Department, wrote that the system is "unusable" for auditing, recounting, or certifying an election. Suarez came to his conclusion after analyzing one precinct in a North Miami Beach municipal runoff election held May 21, 2003.

"Unfortunately, if my observations are correct, we cannot use these reports in their present state for any of these purposes," he wrote. The e-mail memo was sent to a Miami-Dade elections official named Jimmy Carmenate, who is a director of administrative services in the Miami-Dade courts. Suarez declined comment for this article.

The memo was brought to the attention of the County Commission's elections subcommittee by the Miami-Dade Election Reform Coalition at the subcommittee's April 19 meeting. The coalition, a group of civic activists, obtained the memo through a public records request.

At the subcommittee meeting, the new county elections supervisor, Constance Kaplan, acknowledged the problem's existence and said that it also arose in the March 2004 elections. She said she's been aware of the problem since December.

The reform coalition -- which until now had focused on pressing the counties and the state to install a paper backup system to allow recounts -- claims the Suarez memo demonstrates a fundamental problem that supersedes all other voting machine issues.

"We thought the system had basic integrity in terms of how it recorded and tabulated votes," said Lida Rodriguez-Taseff, a partner at Duane Morris in Miami who chairs the coalition. "But we are dealing with a problem that is even more dramatic. We have a system with systemic problems that does not accurately report or tabulate votes."

Rodriguez-Taseff also alleged that the county may have illegally used software that wasn't certified by the state in the May 2003 election analyzed by Suarez.

In an interview this week, Kaplan acknowledged there is a glitch in how the machines "audit" elections. The audit log records the voting machines every activity from the moment it is started until it shut down.

But Kaplan said her office has come up with an "work-around" solution, and that it has ordered new software to solve the problem. That software, however, has not yet been certified, as required, by the state of Florida.

Kaplan denied that the iVotronic system, manufactured by Omaha, Neb.-based Electronic Systems & Software, has shown any problems tabulating votes. She said that system accuracy tests have been conducted regularly and "there has not been any instance where tabulation is in error."

Kaplan accused the election reform coalition of undermining public confidence in the integrity of the election system by presenting misleading information. "I am concerned that making this connection [between the acknowledged auditing failure and the alleged failure to accurately count votes] is not accurate and is certainly deluding voter confidence in our system," she said. "One of my priorities is getting voters back to having voter confidence."

But one computer science expert says the Suarez memo damages the credibility of the iVotronic voting system, and that blaming the critics is unproductive. "The problem is that if the audit records are corrupted, how do you know the voting records are not also corrupted?" said Douglas Jones, a University of Iowa computer sciences professor who serves on the Iowa Board of Examiners for voting machines.

Jones, who was brought in by the reform coalition to serve as election expert, said the glitch requires a "careful public demonstration" showing that the flaw does not affect how votes are tallied in elections.

Jenny Nash, a spokesman for Florida Secretary of State Glenda Hood, said she was not familiar with the issues raised by the Suarez memo. But "if there is a problem to be fixed, we will fix it quickly," she said. "The vendors act on these things very quickly and we work closely with the vendors."

The County Commission's elections subcommittee consists of Commissioners Dennis T. Moss, Jimmy L. Morales and Betty T. Ferguson, who chairs the panel. None of the members returned calls for comment.

Electronic Systems & Software said it is "absolutely confident that every vote cast was counted accurately." It said it would work with Miami-Dade officials to address the issue raised about the audit log.

Broward County Supervisor of Elections Brenda Snipes also did not return calls for comment.

HEADLONG RUSH

In recent months, electronic voting machines have come under increasing scrutiny across the country as the November presidential election approaches. The iVotronic system is one of the main systems in use nationally. Other major manufacturers are North Canton, Ohio-based Diebold and Oakland, Calif.-based Sequoia Voting Systems. The Sequoia system is used in Palm Beach County.

In the wake of the 2000 presidential election debacle, many counties and states invested heavily in the new, largely untested voting technology. Miami-Dade spent roughly $24.5 million to buy touch-screen voting machines from Elections Systems & Software. Broward paid $17.2 million to also buy machines from ES&S.

It's estimated that about 30 percent of the registered voters across the country will use touch-screen voting machines this fall. But critics contend that the headlong rush into electronic voting was a mistake because the systems are unproven. Indeed, there have been problems and unexplained anomalies in Florida and across the country since the electronic machines came into wide use in 2002.

The American Civil Liberties Union did a study of 31 Miami-Dade randomly selected precincts in the September 2002 election and found that 18,752 voters signed in at the polling place to vote but that only 17,208 of them were recorded as having cast votes. The ACLU reported that 1,544 votes, or 8.2 percent, were potentially lost.

More recently, in a January 2004 runoff election for a state House seat straddling Palm Beach and Broward counties -- a special election in which voters were asked to select someone for just one office -- of the roughly 10,000 people who signed in at the polls, 134 supposedly failed to vote. The winner, Ellyn Bogdanoff, won by just 12 votes.

Critics questioned whether 134 people would go to the trouble of showing up to vote on that state legislative race, then not vote. They said it indicated that the iVotronic system may have failed to record people's votes, a problem known as the undervote.

Some election officials explain that there are voters who go to the polls and decide not to cast a vote, while others choose candidates on the touch-screen machine but then fail to press the button that actually records their vote.

"When the coalition is talking about lost votes, that is a misnomer," Kaplan said. "That is voter falloff, when voters choose not to vote."

Snipes, Broward's head of elections, has said that she believes the biggest problem is voters' failure to press the blinking vote button to record their vote.

One of the biggest complaints about touch-screen voting systems is that it is impossible to do a true recount of each vote using touch-screen machines unless the electronic machines are specially equipped with a voter-verified paper trail system.

But Florida Secretary of State Glenda Hood, who was appointed by Gov. Jeb Bush, has opposed the printers as unnecessary, and the GOP-dominated Legislature resisted Democratic efforts this past session to authorize them.

Earlier this year, U.S. Rep. Robert Wexler, D-Boca Raton, filed suits in Palm Beach Circuit Court and U.S. District Court alleging that the touch-screen machines do not comply with Florida election law because they are not equipped to allow manual recounts in close elections.

Critics, including prominent computer scientists across the county, also argue that electronic voting systems are vulnerable to tampering or malfunctions. Last month, California Secretary of State Kevin Shelley decertified electronic voting on Diebold machines in four counties after voting machines malfunctioned.

INACCURATE AUDIT

While the issues of security and lack of a paper record have been widely debated, the question raised by the Suarez memo -- whether electronic voting machines accurately tabulate the vote -- has not been widely discussed.

To understand the issues raised in the Suarez memo, it's necessary to first understand something about how the iVotronic machine works.

When someone votes on an iVotronic machine, that vote is recorded in two places. First, the vote is recorded in the individual machine's electronic ballot box, which is technically called the vote image report. It compiles how many people voted each candidate.

Second, the vote is also recorded in the audit log. That log does not specify which candidate received votes. Instead, it lists everything that happened on the machine -- the time when the machine was turned on, the number of times votes were taken, the time when votes were tallied, and the time when the machine was shut down. The audit log can be used to reconcile the total number of votes the machine recorded with the total votes cast in the voter image report.

In his memo, Suarez analyzed a precinct where just nine electronic voting machines were used. He first examined the audit logs for all nine machines, which was compiled onto one combined audit log. He found that the audit log made no mention of two of the machines used in the precinct.

In addition, he found that the audit log reported the serial number of a machine that was not used in that precinct. The phantom machine that appeared on the audit showed a count of ballots cast that equaled the count of the two missing machines.

Then he looked at the vote image report that was an aggregate of all nine voting machines. He discovered that three of the machines were not reported in the vote image report. But a serial number for a machine not used in the precinct appeared on the vote image report. That phantom machine showed a vote count equal to the vote count on the two missing machines. The other missing machine showed no activity.

"I find this unacceptable from an auditing and a certification perspective," he wrote in the memo.

Puzzled by what he found in his review of the audit report and vote image report, Suarez reviewed each report a second time on a separate computer. This time he made an even more disturbing finding.

Unlike in his first review of the audit log, he discovered that 38 votes cast went unreported in the audit log but not in the vote image report. The 38 votes was the exact total number of ballots cast on the two machines not reported in the audit log.

In his second review of the vote image report, he found that the report showed two "made-up" machines which were not actually used at the precinct. The number of votes cast in the phantom machines matched the number of number of votes in the actual machines in the precinct.

"In my humble opinion (and based on my over 30 years of experience in the information technology field)," Suarez wrote, "I believe that there is/are a serious 'bug' in the program(s) that generate these reports making these reports unusable for the purpose that we were considering (audit an election, recount an election and if necessary, use these reports to certify an election)."

Adding to the mystery, Rodriguez-Taseff said her group just discovered that one of the apparent phantom voting machines actually exists -- based on its serial number -- in the county's touch-screen machine inventory. But it was not used in the North Miami Beach precinct examined by Suarez for that May 2003 election.

In an interview, Kaplan acknowledged the problems identified by Suarez. She said the county's temporary solution to the iVotronic audit problem is that county officials in coming elections will perform all audits by directly inspecting each individual machine when there is a discrepancy between the vote image report and audit log.

Previously, the county had done audits by saving the audit log onto a device known as a flashcard. The flashcard then would be inserted into the machine and effectively download the audit information.

But the reform coalition's Rodriguez-Taseff said that in May 2003, at the time of the election analyzed by Suarez, the state of Florida had not certified the software to allow the audit logs to be saved onto the flashcards. That software was not certified until June 12, 2003.

If that's true, it would not be the first time that iVotronic's manufacturer was caught using uncertified software. Election officials in Marion County, Ind., discovered last month that ES&S used uncertified software in an election last November. The voting machine manufacturer acknowledged that it used uncertified software but defended the vote results as accurate.

VOTE ACCURACY IN QUESTION?

Nevertheless, Kaplan insisted that Suarez's analysis did not demonstrate any basic problems with the accuracy of the vote counts produced by the county's iVotronic system. "The Suarez memo has nothing to do with the tabulation process," she said. "It is very annoying that the coalition keeps equating the tabulation function with the audit function."

University of Miami law professor Martha Mahoney, a member of the reform coalition, said that argument doesn't fly.

"Kaplan repeatedly says this is a problem with auditing and not vote tabulation," Mahoney said. "But she does not address the fact that in the [Suarez] memo there are problems found in the vote image report."

The University of Iowa's Jones said the Suarez memo, and Kaplan's subsequent statements, raise broader questions about the integrity of the county's voting system. "The discussion that has grown from this memo has led to an admission by the county that [the problem cited in the Suarez memo] is not an isolated occurrence," he said.

Some experts say the Suarez memo proves that the state of Florida's process for certifying the iVotronic system is flawed, and that the system was certified without rigorous research, standards and review. "Certification did not protect us against this problem coming through," Mahoney said. "They tested the software and it did not detect whatever process is causing these anomalies to appear."

Rebecca Mercuri, a research fellow in computer science at the John F. Kennedy School of Government at Harvard University and a prominent critic of electronic voting systems, said the Suarez memo has "very serious connotations" for the November presidential election.

"Now we have evidence that at least some component does not work correctly," she said. "This is very bad."