In the 2004 election, 30 percent of US voters cast their ballots on electronic voting machines with software that does not produce a paper trail, allow a recount, or have the ability to retain any traces of possible tampering to final election results. Meanwhile, a growing number of experts have been warning that without any trace, today's leading e-voting systems are neither secure nor reliable.

Reports of irregularities in the 2004 elections poured in faster than they could be investigated in the days after November 2. Animosity at the polls, fraud, malfeasance, machine malfunctions, voter intimidation, vote suppression, and votes changed on screen were just some of the 26 problem types classified by Voters Unite, a nonpartisan group dedicated to accurate elections.

Voting anomalies were also recorded by the Verified Voting Foundation, a nonprofit group promoting election audit and verification. Another nonpartisan Web site recorded more than 37,000 incidents.

The elections were so chock-full of irregularities that on November 5, three members of the House Judiciary Committee submitted an urgent letter to the General Accountability Office requesting an investigation into electronic voting machines. They urged the GAO to improve election systems and administration.

"During the early voting period, my office received numerous accounts of difficulties with the touch-screen machines. Dozens of voters called to notify us that they attempted to select John Kerry for President or Betty Castor for Senate but instead noticed in the review screen that their votes were shown for Bush and Martinez," said Congressman Robert Wexler (D-FL), who had written the letter with Jerrold Nadler (D-NY), and John Conyers (D-MI).

"Based on these chilling accounts as well as many other serious reports of election irregularities in Florida and across the country, I joined together with Congressmen Conyers and Nadler to call for a thorough, nonpartisan investigation of these many reports," said Wexler.

On November 8, three Congressmen joined them—Robert C. Scott, ranking member of the Subcommittee on Crime, Terrorism and Homeland Security, Melvin Watt, the ranking member of the Subcommittee on Commercial and Administrative Law and Congressman Rush Holt (D-NJ).

The six Congressmen submitted a second letter to the GAO, stating: "There is substantial concern that much of the primary evidence needed to evaluate these allegations will not be preserved without immediate action. We would greatly appreciate any steps you could take to ensure the preservation of this important information."

On November 1, the day before the election, Senator Tom Daschle took the Republican Party to court for an incident involving alleged intimidation of American Indian voters in South Dakota. US District Judge Lawrence Piersol issued a restraining order to prevent aggressive Republicans from doing anything to "harass, intimidate, or discourage voters." David Jordon of Charlottesville, Virginia, a volunteer Daschle poll watcher, testified that Republican lawyers had followed American Indians out of the polls and were writing down their license plate numbers at Lake Andes, SD. At the time, Daschle was locked in a close race for his fourth term with his Republican opponent, John Thune, who won by a narrow margin.

Voters Unite also recorded malfeasance in Colorado on Election Day in Boulder, Denver, Jefferson, Douglas, and Weld counties, where election judges mislead voters about the voting process and sent many voters away without allowing them to cast ballots. Some judges misinformed voters by telling them that they could only vote for president on provisional ballots, causing considerable confusion in the voting process.

Black Box Voting is a nonpartisan, non-profit consumer protection group that is conducting fraud audits on the 2004 elections. Founder Beverly Harris is the co-author of Black Box Voting: Ballot Tampering in the 21st Century with David Allen. Harris says that she has amassed evidence of election fraud in the 2004 elections in several states involving voting machine vendors working with complicit election officials.

Black Box Voting's forensic auditing begins with indicators, like oddball statistics, mismatched records, or secretive, obstructive behavior. The group obtains diagnostic documents and pulls all ballots for hand recounts. "It's okay to say the 'F' word—Fraud," said Harris. "This is more than suspicious. There is more than ample evidence that there has been manipulation of the voting machines. We just don't know the scope of it yet." Harris claims the greatest disparity between the exit polls and the results are in places where e-voting machines were used.

"We want to audit these systems," said Kathleen Wynne, an investigator with Black Box Voting. "They have the Help America Vote Act. We want the Help America Audit Act. And this is not a partisan issue. It's an American issue. It gets right down to the core of our democracy. People we've been talking to are concerned that they see their democracy disappearing right before their very eyes."

When Harris and a group called Florida Fair Elections approached Volusia County, Florida on November 18 to perform a legal audit and hand recount of the votes, the Elections Office provided copies of poll tapes but not the originals. A poll tape is the printout from an optical scan voting machine used in official recounts. The crew found the original poll tapes in the garbage. "We have a lawyer who is expediting the process for us now," said Harris on November 19. "And we've expanded the number of precincts we will be recounting in this county."

Chuck Herrin, a computer security expert in Greensboro, North Carolina, thought Beverly Harris was a conspiracy theorist. He decided to disprove her by double-checking evidence that Diebold machines are too easy to break into. "I thought the whole thing was overblown," said Herrin. "I thought there was no way it could be that bad. But once I started evaluating the software, it became evident that the design was beyond poor, way beyond poor." It is not necessary to hack into a Diebold e-voting machine to get access to and tamper with election results. "Anybody using Microsoft Office can do this easily," said Herrin. "It's just a matter of opening the Access election database and changing the numbers." (Both Herrin and Harris's Web sites—www.chuckherrin.com/hackthevote.htm, www.blackboxvoting.com—contain instructions on how to hack a Diebold e-voting machine.)

David Bear, spokesperson for Diebold Elections Systems, contested these claims. "They've put together something they want you to see. People are making these allegations on something they know nothing about." Bear contends that the software Harris and Herrin are evaluating is outdated and has never been used in an election.

"This is wrong," said Dan Wallach, a Rice University professor of computer science who specializes in security. Wallach tested the Diebold systems for the Hopkins Rice Report, which he co-authored with Avi Rubin, a computer scientist at Johns Hopkins University. The Hopkins Rice Report documents the deficiencies of e-voting systems. "We found the quality of their software engineering to be far below standard," said Wallach. "The complexity of modern programs of this size means that the version you ship in 2002 is similar to the version you ship in 2004. It bears a strong resemblance to what they are using today because if it didn't, it would be even worse."

Diebold's Bear said that the software is reviewed and certified both at the national and the state levels by independent testing authorities. But Wallach countered that Federal Election Commission standards in the 1990s and in 2002 did not include specifications for electronic voting software, so the software released in 2002 has not undergone a review or certification process by the FEC.

Diebold pointed out that its systems are password protected but Wallach responded that the password is written in the manual and is the same for all machines. "Imagine if every car that was manufactured used the same key. It wouldn't be long before someone figured out that one key worked for all the others. The design of the system makes it vulnerable to fraud. And furthermore, it's undetectable fraud. If such fraud occurred, you would never know."

There were also numerous complaints about votes that changed from Kerry to Bush right on the screen as well as machine malfunction of leading e-vote vendor Election Systems and Software's e-voting machines. But ESS denies this. "There has never been a claim of a security breach or a hacking of the system," said Meghan McCormick, an ESS spokesperson.

By November 14, VoteProtect.org had recorded 4,313 incidents in the state of Florida, with the highest numbers in three predominantly Democratic counties: Miami-Dade, Broward, and Palm Beach County.

In many Florida precincts, the number of new Democratic registrations just didn't match up with election results. A certain percentage was explained away by what is known as the "Dixiecrat" phenomenon, when a registered Democrat from the Old South votes for a Republican, but it could not explain the high frequency with which this occurred, nor the number of new Democratic registrations that were known to have registered specifically to vote for Kerry. "We know of 15 precincts where votes flipped and Democratic votes went to Bush that should have gone to Kerry," said Al Rogers, communications director for Jeff Fisher, a Democratic congressional candidate in the 16th district who lost the election.

Rogers and Fisher claim to have evidence of e-voting fraud that involves manufacturers and complicit election officials. They are working with Ralph Nader, seeking a recount in Florida. By November 15, Nader had secured a recount in New Hampshire and was working to secure one in Ohio. The recounts must be completed in time for the Electoral College vote on December 13.

According to New Hampshire senior assistant attorney general Bud Fitch, New Hampshire law only allows for optical scanning machines that count the vote on paper. Voters get paper, mark their vote, then feed the ballot into an optical scanning machine. The state prefers column-style ballots, set up to list offices on the left of the ballot with a column for each party. But some optical scanning machines cannot be programmed for the column ballot. This caused confusion in the 2004 election.

"The column-style ballot is visually more intuitive for those people who vote a straight ticket. We have some machines we will not use in the next election. They will be replaced but we don't know with what yet," said Fitch. New Hampshire uses optical scanning machines from both Diebold and Election Systems and Software, but will stop using the Diebold machines.

The Ralph Nader hand recounts in 11 wards in New Hampshire began on November 18. Nader requested recounts where the results seemed anomalous in their support for President Bush and the votes were counted on optical scan machines, particularly the Diebold AccuVote machine.

Also on November 18, the Ohio Democratic Party launched a federal court battle regarding 155,000 provisional ballots. At that time, Ohio votes were still being counted. But for the first time, provisional ballots became a hot topic. "Elections are always close in Ohio, but provisional ballots weren't big until this year. So it's uncharted territory. Both Democrats and Republicans want to set precedent in court," said Bill Sloat, the reporter at the Cleveland Plain Dealer who wrote the article covering the Democratic federal court battle.

Odd numeric correlations in the 2004 election did not occur solely in New Hampshire and Florida. They occurred in many other states, most saliently Ohio and Pennsylvania. The pattern prompted a University of Pennsylvania professor, Steven F. Freeman, to issue a paper on November 10 entitled, "The Unexplained Exit Poll Discrepancy."

Freeman points out that while nearly every exit poll showed Kerry ahead by sizable percentages, vote counts showed "very different numbers than the polls predicted and the differentials were all in the same direction." They represented an advantage for Bush.

His study shows that in three critical states—Florida, Ohio, and Pennsylvania—which determined the outcome of the election, the exit polls differed significantly from recorded tallies. He then analyzes the statistical likelihood that the exit polls and results would differ as widely as they did and the chance of it happening in all three battleground states.

Freeman's report concludes: "The likelihood of any two of these statistical anomalies occurring together is on the order of one in a million. The odds against all three occurring together are 250 million to one. As much as we can say in social science that something is impossible, it is impossible that the discrepancies between predicted and accurate vote counts in the three critical battleground states of the 2004 election could have been due to chance or random error."

A paper refuting Freeman's conclusion about Florida was issued the same day (November 10) by the CalTech-MIT Voting Technology Project. It asserts that partisan voter registration statistics and the Bush/Kerry vote in certain Florida counties are due to political changes, noting that counties with the greatest disparity are closely aligned with the "Dixiecrat" South.

However, the CalTech-MIT paper is not an apples-to-apples comparison with Freeman's paper. While Freeman looks at three states, the CalTech-MIT study addresses Florida counties and did not analyze every county and precinct in Florida.

On November 18, a Quantitative Methods research team at the University of California Berkeley released the results of a statistical study of voting irregularities where electronic voting machines were used in Florida. The study found that electronic voting machines probably awarded 130,000—260,000 or more excessive votes to President Bush in the 2004 presidential election.

The study found a discrepancy between the counties where e-voting machines were used compared with counties using traditional voting methods. The three counties with the highest number of voting anomalies were Democratic: Broward, Palm Beach, and Miami-Dade.

"For the sake of all future elections involving electronic voting, someone must investigate and explain the statistical anomalies in Florida," said Professor Michael Hout, who is an expert on statistical methods, a member of the National Academy of Sciences and the UC Berkeley Research Center.

The highly partisan positions of Diebold and ESS top executives, who are avowed Bush supporters, have only fueled the firestorm. Senator Chuck Hagel, Republican of Nebraska, is a former CEO of ESS, which is headquartered in Omaha, Nebraska.

Walden O'Dell, CEO of Diebold, is an avowed Republican; in fact, he is a member of President Bush's "Rangers and Pioneers," which raised $100,000 per participant for the 2004 race. In November 2003, O'Dell issued a letter stating: "I am committed to helping Ohio deliver its electoral votes to the president next year."

Diebold is one of the largest worldwide providers of bank ATMs, issuing billions of paper receipts for financial transactions on a daily basis around the world. The banking industry demands reliable, secure ATMs from Diebold and that's what the company delivers. So many observers wonder why Diebold's e-voting system is so lacking in security and reliability.

Herbert Thompson, a security advisor to the National Security Agency and the Defense Advanced Research Project Agency, among others, points out that anyone could tamper with e-voting election results, from the touch screen, the optical scanners, or the modems. "Anyone with physical access to the machine can tamper with vote totals. Once you connect the modems, it gets a lot scarier because someone anonymous can get in remotely," said Thompson.

Diebold denies the charge. "There's never been a security issue with touch screens," said spokesperson David Bear. "There's never been a vote lost due to a security issue with a vote screen."

Diebold's claim is difficult to prove, partly because the company and leading e-voting vendor Election Systems and Software refuse to release their software source code for inspection by independent third parties. They claim it would violate their right to copyright secrecy.

However, in early November, Diebold Election Systems's mode of operation cost Diebold $2.6 million to settle a lawsuit with California for making false claims about the security and certification of its equipment to state leaders in order to get payments from the state.

There is strong resistance by some, but not all, Republicans in Congress to passing legislation that would mandate a voter-verifiable paper trail in e-voting systems. Representative Rush Holt (D-NJ) has tried for 16 months to amend the ironically named Help America Vote Act (HAVA), which has paved the way for the proliferation of faulty e-voting machines. Yet Holt's office says that Representative Robert Ney (R-OH) has blocked Holt's Voter Confidence and Increased Accessibility Act (HR 2239) from moving forward.

"The bottom line is that Representative Ney has elected not to take any action on this legislation, even though we are up to 157 co-sponsors. HR 2239 would allow voters to visually inspect the print-out of the receipt to verify whether the vote was what the voter intended or not. And this is a bi-partisan bill. We've got Republicans such as Tom Davis of Virginia who is a member of the Republican leadership, co-sponsoring this bill," said Pat Eddington, Rep. Holt's communications director.

Representative Ney's office did not return calls for comment by press time.

New York is the only state in the country which has flatly refused to buy e-voting machines. "We won't purchase these machines in New York unless there is a way to track the vote. Any machine we buy in New York will have to be auditable and have a paper trail," said Lee Daghlian, director of public information at the New York Board of Elections.

But e-voting machines aren't the only thing wrong with elections in this country, with its patchwork of election policies and procedures that differ from state to state. Some have observed Board of Election officials who seem too partisan for a job that requires objectivity or who have relationships with Diebold or ESS that seem far too cozy.

Many agree that there is a strong need for election reform. "Electronic voting is only one of a number of problems we need to address," said Wallach. "We really need to re-examine the entire election franchise from soup to nuts."