A Single Person Could Swing an Election

http://www.washingtonpost.com/wp-dyn/co ... 01451.html

A Single Person Could Swing an Election
Electronic Systems' Weaknesses May Be Countered With Audits, Report Suggests

By Zachary A. Goldfarb
Special to The Washington Post
Wednesday, June 28, 2006; A07



To determine what it would take to hack a U.S. election, a team of cybersecurity experts turned to a fictional battleground state called Pennasota and a fictional gubernatorial race between Tom Jefferson and Johnny Adams. It's the year 2007, and the state uses electronic voting machines.

Jefferson was forecast to win the race by about 80,000 votes, or 2.3 percent of the vote. Adams's conspirators thought, "How easily can we manipulate the election results?"

The experts thought about all the ways to do it. And they concluded in a report issued yesterday that it would take only one person, with a sophisticated technical knowledge and timely access to the software that runs the voting machines, to change the outcome.

The report, which was unveiled at a Capitol Hill news conference by New York University's Brennan Center for Justice and billed as the most authoritative to date, tackles some of the most contentious questions about the security of electronic voting.

The report concluded that the three major electronic voting systems in use have significant security and reliability vulnerabilities. But it added that most of these vulnerabilities can be overcome by auditing printed voting records to spot irregularities. And while 26 states require paper records of votes, fewer than half of those require regular audits.

"With electronic voting systems, there are certain attacks that can reach enough voting machines . . . that you could affect the outcome of the statewide election," said Lawrence D. Norden, associate counsel of the Brennan Center.

With billions of dollars of support from the federal government, states have replaced outdated voting machines in recent years with optical scan ballot and touch-screen machines. Activists, including prominent computer scientists, have complained for years that these machines are not secure against tampering. But electronic voting machines are also much easier to use for disabled people and those who do not speak English.

Voting machine vendors have dismissed many of the concerns, saying they are theoretical and do not reflect the real-life experience of running elections, such as how machines are kept in a secure environment.

"It just isn't the piece of equipment," said David Bear, a spokesman for Diebold Election Systems, one of the country's largest vendors. "It's all the elements of an election environment that make for a secure election."

"This report is based on speculation rather than an examination of the record. To date, voting systems have not been successfully attacked in a live election," said Bob Cohen, a spokesman for the Election Technology Council, a voting machine vendors' trade group. "The purported vulnerabilities presented in this study, while interesting in theory, would be extremely difficult to exploit."

At yesterday's news conference, the push for more secure electronic voting machines, which has been popular largely on the left side of the political spectrum since the contested outcome of the 2000 presidential election in Florida, picked up some high-profile support from the other side.

Republican Reps. Tom Cole (Okla.) and Thomas M. Davis III (Va.), chairman of the House Government Reform Committee, joined Rep. Rush D. Holt (D-N.J.) in calling for a law that would set strict requirements for electronic voting machines. Howard Schmidt, former chief of security at Microsoft and President Bush's former cybersecurity adviser, also endorsed the Brennan report.

"It's not a question of 'if,' it's a question of 'when,' " Davis said of an attempt to manipulate election results.

http://votetrustusa.org/index.php?optio ... &Itemid=51

Too Much, Too Fast, More Than They Can Chew
By John Gideon, VotersUnite.org and VoteTrustUSA
June 09, 2006
Pottawattamie County, Iowa is a small, urban county in the southwest corner of the state, on the Nebraska border. I-29 nearly cuts the county in half. The county has only 60,536 registered voters who voted in 41 precincts in this year's primary election. They selected Election Systems and Software (ES&S) M-100 Precinct Optical Scan machines for their poll-site tabulators, and they use AutoMark machines as their accessible voting system at the polls. The county is a good example of a small county in the "Heartland" where people tend to know more about other people just because there are fewer people to know.

So it is really no surprise that County Auditor Marilyn Jo Drake knew there was a problem with the vote counting machine at the county, and the poll-site optical scan machines as well. She saw that a 19 year-old college student, Oscar Duran, was taking the lead on absentee ballots in the race to be the Republican candidate for County Recorder, and an incumbent with 23 years in office, John Sciortino, was falling behind. She noticed this when there were only 178 absentee ballots counted, and she watched as the divide between the two grew larger.

Once the totals were nearly complete, Drake had her workers do a hand count of the absentee ballots to see if there was a problem. The count on the tallying machine made and programmed by ES&S read Duran = 99 and Sciortino = 79. The hand count, however, told a different story – Duran = 25 and Sciortino = 153. There was definitely a problem and Drake knew this problem was probably evident on all ballot positions.

At this point Drake stopped the machine count, called the Secretary of State's office, and talked to her county Board of Supervisors requesting permission to hand-count all of the county's ballots. The state and county gave Drake permission to do the hand-count and it was carried out on Wednesday.

While no statewide race results were changed by the new tallies at least one local race was changed. Machine results were indicating that Pottawattamie County Board of Supervisors Chairman Loren Knauss was trailing in a field of ten candidates in the race for the county board. The hand-count revealed that, in fact, Knauss was the top vote getter, by a wide margin, in a race for three candidates on the Republican ticket to face three Democrats in November.

Of course the county realized they had a problem somewhere and that it was probably with the ballot programming that was provided by their vendor, ES&S. The same ES&S that has failed in Texas, West Virginia, Indiana, Arkansas, and Pennsylvania in this year's primaries. Investigation found that, in fact, the ballot programming was the problem.

It is normal practice, in races where there is more than one candidate, that the names be rotated on the ballots in different precincts. So, one candidate may be first on the ballot in one precinct but he may be second or tenth on the list in another precinct. The paper ballots are printed with a code that is supposed to tell the vote tally machines which precinct the ballot is from so it knows who is in the top spot and who is in the second or tenth spot, and the votes are then given to the correct person.

ES&S failed to set up Pottawattamie County's software to manage the name rotation, so it counted every ballot as if they were all printed exactly alike. Hence, the incorrect totals that they were receiving.

Kudos to Drake and the Secretary of State's office and the county supervisors for recognizing the problem and making the decision to cut the machines out of the vote tallying process and for allowing the hand-count of the ballots. But………

Why did the Logic and Accuracy (L&A) test not find the software error? The L&A test is supposed to be done with a representative test deck that should have had the ballot positions rotated. If this had been done the error should have been apparent because the results of the tallying of the test deck would not be accurate. A call to Charles Krogmeier of the Secretary of State's staff found that they have the same question about why the L&A test did not find the error. Was it even accomplished? They are investigating.

Who made the test deck? The deck was probably provided along with the ballot programming to the county. If so, then ES&S provided a test deck that was just as flawed as their software. If this was the case, and I suspect it is, then the county needs to use this as a 'Lesson Learned' and inspect their test deck from now on. This error should have been readily observable. The Krogmeier told me that they suggest that each county use two different test decks when they do their L&A testing. One deck from the ballot programmer and one made up by political party observers on the day of the testing. In fact state law gives any voter the right to add ten cards to their counties test deck. That probably did not happen.

Iowa has 23 other counties using the same equipment as Pottawattamie County, and seven counties use the ES&S touch screen machines. Who at ES&S did the ballot programming for Pottawattamie, and for how many other counties did that person provide ballot programming? Is there any reason to believe that this is an isolated incident? Is there any reason to think that the results in other counties programmed by ES&S are correct? The state of Iowa should be asking ES&S these questions, and maybe Texas, Pennsylvania, West Virginia, Indiana and Arkansas also. Remember that it was only because of an observant county auditor that this problem even came to light. Every ballot program done by that person should be brought into question and tested to ensure every voter's wishes were accurately counted.

Krogmeier told me that they would be investigating who the programmer was and how many other counties may have been affected. But will they do hand-counts anywhere else in the state? And how would they check the accuracy of the touch screen machines? I suggested that Krogmeier call Arkansas and ask about the ballot programming errors that are keeping eight counties there from using their voting machines in their run-off election next week.

Krogmeier also told me that there was a problem in Dallas Co., a suburban county just to the south of Des Moines, where a professor from Drake University asked to use the AutoMark machine when he voted. He went through the ballot, marking his choices, and when he was through he checked the ballot to find that one race had been swapped. His ballot was "wasted" and he voted again with the same results. He then agreed to allow a poll worker to sit and watch while he voted. The same thing happened and the machine was taken out of service at that point. The problem was the same as in Pottawattamie County except that we only know of one ballot that was affected. How many more were marked incorrectly on this and the AutoMark systems used in 20 other Iowa counties.

Ballot programming errors are a new threat at every new election, and it is time election officials realized it. Problems are cropping up all over the country, and there is no indication that they will abate. Elections officials cannot just accept that what they get from ES&S, Diebold, or any other vendor is 'good to go'. They must test the software and they must be heard when there are problems with it. If it hadn't been for an observant county clerk in Pottawattamie County, the machines' choices for County Board of Supervisors* not the voters' choices * would have been declared the winners.

Either the vendor does the programming, or the county does it. The vendors want to do that work because they make a lot of money for the service. ES&S requires that they do the programming for the AutoMark machines in many of their contracts. Counties could do the work themselves but the cost of the programming software is so expensive that it is hard for some counties to justify the expense.

This is just the primary season with primaries stretched from March until September. The vendors are failing to provide good service to counties now. What will happen when they have to provide service to every county in the country on the same day in November? We can't simply sit by and watch. Something needs to be done to prevent the train wreck that is a certainty otherwise.

Can't say I am in the least bit surprised.