Voting Machine Primer
What You Should Be Aware Of
1. Could Democracy be stolen from us in 2004?
— We've developed safeguards to prevent tampering with ballots. We've figured out ways to impede tampering with the voters themselves. But we DO NOT have adequate safeguards to prevent the most dangerous election-rigging of all: Tampering with programmers, vendors and technicians! Why has everything changed, and what are the dangers to Democracy?
How does the integrity of our voting system change when we use machines to count the votes?
- Machines are more inaccurate than hand counting. The machines lose some votes (up to twice as many as hand-counts) and count some votes for the wrong candidate. Human error compounds the mistakes. Election officials must be trained, and software programming errors have resulted in mistakes as high as 100 percent.
- The machines produce new tampering and vote-rigging vulnerabilities. We do not have adequate systems to protect against tampering with programmers, vendors and technicians.
- Voting machines create hidden changes in the way our voting system works. Using the machines, in effect, replaces sworn, elected officials with unsworn, unelected technicians. In most states, elected officials can no longer look at the voter-verified evidence. Many state laws prohibit officials from looking at the paper ballots, and only allow them to look at the counts coming out of the machines, even when there is a recount.
- A safe voting system is one that many eyes can view. Machines eliminate transparency in vote-counting. The newest machines eliminate the paper trail — the only voter-verified evidence of how votes really were cast, effectively saying "trust us" — voters and local election officials no longer have any way to verify that votes were counted accurately.
- It used to be that we knew who our elected officials were and the names of local election officials were a matter of public record. Manufacturers, who now count our votes, are not required to reveal the names of owners or key people. The codes counting our votes are considered "proprietary" and outside officials are not allowed to examine them.
- Some voting machine manufacturers are salted with vested interests. Among the owners of voting machine companies and testing labs: active politicians, corporate lobbyists, former CIA directors, and people who have been involved in prosecutions for bribery, kickbacks, and fraud. Our "watchdog" groups are also influenced by special interests. Voting machine companies are using lobbying and political influence to influence purchase of machines, specifications and regulations.
- Vote-rigging on computerized machines may be possible on a grand scale, not just a local scale. It's hard to stuff more than a dozen ballot boxes in your trunk, and it's nigh-on impossible to get 100,000 dead people to vote. But with these machines, we sometimes lose hundreds of thousands of votes in a single city!
>Why did we buy computerized vote-counting machines?
- Election officials were persuaded to buy them by lobbyists and voting companies — with premiums, kickbacks, party boats, and political contributions from vested interests.
- The media demanded to know results faster, in order to win ratings wars.
- Election officials usually failed to ask the right questions. In the rare cases where election officials asked the right questions, voting company salespeople failed to disclose known error rates and known security flaws.
- The media has been engaging in a pep rally instead of covering the real issues. Instead of examining whether the machines are accurate, tamper-proof or free of criminal or special interests, the press has focused on how easy and fun it is to vote on these machines.
What are the risks to Democracy?
- As long as we've had elections, we've had people trying to rig results. Now, vote-rigging is possible on a massive scale, by tampering with the computer programs that count millions of votes.
- Sooner or later someone will start stealing elections. If they haven't already.
- At some point, the balance of power in Congress may shift to the party that was not actually elected, "mandates" will not be mandates, and we may get a president whose votes were augmented by a handful of programmers, instead of an accurate vote "of the people, by the people and for the people."
2. The Distinguished History of Vote-Rigging
Let's face it: We're a flawed species. The best in us shows up in our desire to make our government "of the people, by the people and for the people." But the worst in us shows up every election when, no matter what the system, somebody figures out how to cheat. Here is a brief history of election-tampering. (But computerized vote-tampering alters the scale of things -- we're no longer talking about the machines rigging a city; Now, we're looking at the ability to commit massive election fraud.)
Paper ballots have been used for over 2,000 years. The first known use of paper ballots in an election in the U.S. was in 1629, to select a church pastor.
Early cheaters: Because at first there was little voter privacy, candidates tried to pay people to vote for them.
People used to wander around town with their ballots, where the slips of paper got into all kinds of trouble. (This began happening in Oregon recently, when they converted to absentee balloting: Modern-day scoundrels stood on street corners with official-looking boxes to "collect" the ballots. Nobody knows what happened to them after that.)
- It was considered a great innovation to print standardized ballots at government expense, give them to the voters at the polling places, and require people to vote and return the ballots on the spot. No, this wasn't invented in America: The Australians came up with this procedure, which is now the most widely used voting system in the world.
Properly administered, this paper ballot system sets a very high standard: privacy, accuracy and impartiality. In fact, why are we willing to trust our votes to anything else?
It's difficult, but not impossible to rig the Australian paper balloting system. Here's how certain political machines in the USA achieved paper ballot hand-counted vote-rigging: First, create a fancy set of rules for which votes "count." Take control of the vote count. Train your team better than the other team. A really well-coached vote-counting team used to be able to exclude up to 40 percent of votes!
Lever machines made their debut around 1890 and became popular throughout the USA by the 1950s. They've been out of production since 1982, but are still in use. (Maybe, because they're too heavy to lift? These are truly gigantic metal contraptions.)
How to cheat with a lever machine: Like the new touch-screen DRE voting machines, lever machines left no audit trail. And like the touch-screens, they also have the highest error rates (M.I.T./CalTech study rates error rates for hand-counting at 3.3%, optical scan 3.5%, punch card 4.7%, touch screen machines around 6% and lever machines around 7%. Uh, we're spending $3.9 billion to switch from an error-prone punch card system to the only modern system with a HIGHER error rate?)
Technicians learned to rig the lever machines, in ways that were very hard to spot. Some of the rigs stayed in place for years! Yet, lever machines cannot be rigged on a national scale. At least their unauditable, not very accurate, riggable problems are confined to small geographic areas.
- With lever voting machines you have to put your trust in the technicians who maintain the machines. If you want to rig an election, you buy a few technicians.
Punch cards also date back to the 1890s, but really became stylish around 1964. But the early "Votomatic" machines had such problems that in 1988, the National Bureau of Standards published a report by Saltman recommending the immediate abandonment of this technology. Since punch cards were by then the most widely used system, it's taken awhile to let go of them.
One way to rig a punch card system is to consolidate ballot-counting in one area so that precincts are mish-mashed together; Then, the scoundrel team picks someone to quietly add punches to the votes for the other candidate. The double-punched cards become "overvotes" and are thrown out.
In Duval County, Florida in 2000, over 20,000 overvoted punch cards appeared, primarily in two black precincts. No one was allowed to look at them, much less do a statistical analysis on the probability of such a high overvote rate in a discrete area. Governor Jeb Bush has recently cut the funding for the library that safeguards the six million votes from Florida 2000, and when moving the collection, it is likely that the ballots will be destroyed, eliminating any further examination of Duval County.
Optical Scanning: You fill in the dot on paper ballots, and a computer reads them. Pioneers in optically scanned vote-counting were Todd and Bob Urosevich of Omaha, Nebraska, who founded Data Mark Systems, which became American Information Systems, and in 1997 changed its name to Election Systems and Software. Along the way Bob split away to form I-Mark Systems, which became Global Election Systems, which is now Diebold Election Systems.
Some optical scanners have trouble distinguishing faint deliberate marks from smudged erasures. The newer machines can read more accurately, and accuracy is second only to hand-counting.
People thought optical scan machines could not be rigged, but there are anecdotal reports of likely rigging of these machines as far back as 1980.
Touch screen "DRE" machines These are the latest fad, and they are sleek and fun and convenient. However, like the old lever machines, they haven't been associated with the best accuracy rates, and they don't have any paper trail so voters really have no idea if their vote was counted the way it was cast. The manufacturers surely knew their industry well enough to know that audit trails are a critical component of elections with integrity, but instead of creating a simple receipt so we could say, "Yup! I voted that way!" and drop it in a ballot box, the manufacturers simply took away the paper trail.
It is not difficult to develop touch-screen machines that spit out a voter-verified paper audit trail. It is not even expensive, so one wonders what could possibly have motivated anyone to do away with it. One answer lies in a statement made by Richard Jablonski, of ES&S, who groused about the expense of doing a hand recount when their machines made yet another mistake due to yet another of their own software programming errors. Jablonski said it would be much cheaper to just rerun the machines.
Internet Voting -- Almost no one believes that Internet voting is ready for prime time, but that hasn't stopped upstart companies like Election.com and Votehere.net from trying to talk everyone into it. Problem is, VoteHere seems to be succeeding; a recent news report indicates that Washington State approved Internet voting for military overseas voters. Again, let's get rid of that pesky paper trail (You know, the one that lets us know for sure how people voted.)
Internet voting advocates, and they are difficult to find, even among techies, say that encryption techniques are the key. Well, they may be one key, but there are many ways around encryption. (Think: Send a worm to all the Democrats. Or, mess with Earthlink and AOL on Election Day. Or, pay a technician in the telephone company to create a disruption of phone lines.)
Although every method offers vote-rigging opportunities, only the optical scan, touch-screen and Internet systems enable high-volume vote-rigging on a national or international scale. And the optical scan, touch-screen and Internet systems give the smallest number of scoundrels opportunities to find ways to tamper with the greatest number of votes. It may take only ONE programmer to tamper with literally millions of votes at once.
3. Known errors
They knew about it. They admitted it was their own fault. But they never told us about it -- not when making sales presentations to local county commissions, not when answering questions from the press. Voting machine errors in the range of 20 to 100 percent are not uncommon. (You heard me). Who are "they?" The voting machine manufacturers, who had a duty to disclose, but didn't. Now, it's time for us to grab the reins!
We can't know how often the machines get it wrong, because: We aren't allowed to compare the actual vote with the machine-counted vote. Even when there are recounts, we often aren't allowed to compare the actual vote with the machine-counted vote.
But we do know that the machines have gotten it wrong. Manufacturers knew about the errors, but did not disclose them to buyers or regulators.
Following are just a few of the errors manufacturers knew about before the 2002 election cycle. Yet, in the minutes of dozens of recent county meetings, where sales reps made presentations to persuade officials to buy their machines, nowhere were these problems disclosed -- even when officials asked direct questions about error rates and problems!
McLennan County, Texas: Better than a pregnant chad -- these machines can actually give birth! In one precinct, about 800 votes were tallied, although only 500 ballots had been ordered. "It's a mystery," declared McLennan County Elections Administrator Linda Lewis. Like detectives on the Orient Express, officials pointed fingers at one suspected explanation after another. One particular machine may have been the problem, Ms Lewis said. That is, the miscounted votes were scattered throughout the precincts with no one area being miscounted more than another, Ms. Lewis also explained. Wait -- Some ballots may have been counted more than once, almost doubling the number of votes actually cast. Yes! That could explain it! (Er…excuse me, exactly which ballots were counted twice?)
"We don't think it's serious enough to throw out the election," said county Republican Party Chairman M.A. Taylor. Size of error: 60 percent. Honolulu, Hawaii: Tom Eschberger, a vice president of ES&S, said a test conducted soon after the election on the software and the machine that malfunctioned in a Waianae precinct showed the machine worked normally. He said the company did not know that the machine wasn't functioning properly until the Supreme Court-ordered a recount, when a second test on the same machine detected that it wasn't counting properly. "But again, in all fairness, there were 7,000 machines in Venezuela and 500 machines in Dallas that did not have problems," he said.
Dallas, Texas: Over 41,000 votes were not counted due to software programming errors. A recount was done and ES&S took the blame. A recount showed the error size was about 10 percent. Democrats picked up over 1,000 votes, not quite enough to overturn the election.
Caracas, Venezuela: Venezuela's highest court suspended elections because of problems with the vote tabulation. Venezuela sent an air force jet to Omaha to fetch computers and experts in a last-ditch effort to fix the problem before the delay was ordered. Dozens of protesters chanted "Gringos get out!" at ES&S technicians. Venezuelan President Hugo Chavez, who U.S. officials would very much like to see unseated, accused ES&S of trying to destabilize the country's electoral process. Chavez also asked for help from the U.S. government because, he said, the United States recommended ES&S.
Crozet, Virginia: "When I pushed the button beside 'No' the machine registered my vote as a 'Yes.' I tried this a couple of more times and got the same result. Finally, I poked my head outside the curtain and asked the attendant what I should do. He suggested then that I should intentionally push the wrong button..."
Clay County, Kansas: A squeaker -- no, a landslide -- oops, we reversed the totals -- and about those absentee votes, make that 72-19, not 44-47. Software programming errors, sorry. Oh, and reverse that election, we announced the wrong winner -- The machines said Jerry Mayo ran a close race but lost, garnering 48 percent of the vote, but a hand recount revealed Mayo won by a landslide, earning 76 percent of the vote.
Memphis, Tennessee: Computer problems halted the voting process at all 19 of Shelby County's early voting sites during the 2000 Republican presidential primary, forcing officials to use paper ballots (supposed to be provided by the voting machine company as a backup, but for some inexplicable reason, they were unavailable when they were needed). Election officials had to make voters wait in line or tell them to come back later. Because early voting turnout in this election was six times normal, this snafu affected about 13,000 voters. If there was a beneficiary of the problem, it was George W. Bush, who needed to defeat John McCain in Tennessee - Shelby County, which contains the urban Memphis population, usually votes less conservatively than the rest of the state. Chicago, Illinois: Size of error: 25 percent -- One hundred and eight of 403 precincts were not counted. A pin from the cable connecting the ballot reader to the counting computer apparently got bent, after three-fourths of the precincts were counted correctly. No one could explain how a pin inside a cable became bent during the middle of the count. Democrats requested a full recount; a judge disallowed it.
Newport Beach, California: The voting machine programmer reversed the "yes" and "no" answers in the software used to count the votes, discovered only because California had a law requiring a random sampling of votes by hand. (The new touch-screen machines, however, have no voter-verified audit trail.) Size of error: 100 percent.
Polk County, Florida: A machine count said that Republican Bruce L. Parker won the election, but after a hand recount, Democrat Marlene Duffy Young turned out to be the winner. The manufacturer (ES&S) denied that its machines were responsible for the miscount.
4. Top 10 Ways to Rig a Voting Machine
Computer and accounting experts tell us that there are not 10, not 20, but literally hundreds of ways to rig the system. Faster than the best programmer can create computerized safeguards, new scoundrels will pop up with ways to beat the system. To conceptualize this, think virus: every time we get protected, a new one is created. With hundreds of millions at stake in getting the "right" candidate into office, the motive is already in place. (The candidate need not even know that his election was rigged!) Here are just 10 of the methods that our sources have identified:
1.) Optical scan machine: Create a dummy ballot using a special configuration of "votes" that launches a program when put through the machine. This is called a "back door," it takes just ONE programmer to insert this, it uses very short code and is almost undetectable even if certifiers actually look for it, though indications are the the software examination is not rigorous during certification, and even if it was, nothing guarantees that the software that's certified is the same as what's in the actual machines at every precinct.
2) Help the machine pass testing by creating a program to tamper with the day and time. Make sure it can't be triggered to execute the program until the actual election.
3) Replace certified files with new ones during a "technical servicing." Most of these machines carry service contracts, and the technician goes through a set of diagnostic steps. Technicians are often given software patches or "upgrades" to install that do NOT got through any official approval policy. And again, even if they did, there is no guarantee that the software patch used in the machine is the one that was examined.
4) Include a layer of software that is insulated from certification testing. There are two ways this can be done: First, by incorporating Microsoft Windows into the system. Bypass the testing on the Microsoft portion (and embed malicious programs in the Microsoft operating system instead of the voting software). Another way is to patent a layer of the software and sell it to other voting machine manufacturers. This way, the mischief lies with a vendor, not the manufacturer. A single vendor could compromise many companies at once.
5) Build in a "diagnostic" tool that lets technicians add or change programs just before, during, or after Election Day. Some of the new DRE machines have a panel that, when removed, allowed access to the very bowels of the machine, potentially erasing all the votes and replacing them with a brand new cartridge.
6) Have your technicians obtain their files from an internet site. Tell them how to troubleshoot using a batch of replacement files and patches on a server. Anyone who gains access to the server can replace one with another, for example, replacing the central counting program with a file of the same name which contains a variation of the program.
7) On the way to tabulating the votes, substitute one memory cartridge with another. In Georgia during Eletion 2002, dozens of memory cartridges were "misplaced," representing tens of thousands of votes. There was no documented chain of custody during the time they were missing.
8) Tell county commissioners that they don't need to see you demonstrate or test an "upgraded" system, because they saw the demonstration before with the previous version.
9) Get hold of targeted e-mail lists and send a fast-spreading worm to military voters of your opponent's party. When they try to vote on the Internet, they'll experience problems.
10) Buy a tech and plant him as a poll worker. Have him go through the training (this works best when the precinct uses your competitor's machine) and then have him flub the election, by preventing machines from booting up on time and then blaming it on the manufacturer. If things really get messed up, have him call the press and grant interviews. 5. The international problem
5.The International Problem
This is not a United States problem. It isn't a third world problem. It is a global problem. By planting clandestine code in the vote-counting machines, governments could win a war without using any bullets. Can these machines be used to manipulate international politics?
When safeguards aren't in place, clandestine computer code can be inserted into voting machines to affect election outcomes. Among those with current ties to election companies: A member of the Senate Foreign Relations Committee and a former CIA director.
- Latin American countries are pressured to buy machines made by corporations in the USA
- There are reports that some vendors who supply components for U.S. machines are from the Phillippines, or Russia.
- Diebold is working hard to get its machines into Canada. ES&S is spreading across South America and Europe.
- Votehere is in Sweden and Britain.
- Some countries are holding fast to the safe, accurate old Australian paper ballot system. Canada counts the whole country in four hours. France in six. It can be done, but the number of questions on national elections must be limited to a manageable quantity.
6. Easy, inexpensive ways to correct the problem
Safeguards made easy and practical activism: By mixing low-tech, people-oriented solutions with high tech systems, we can solve this problem. It will be fun (yes, really) and it won't cost taxpayers any money. And it starts with YOU!
What can we do about it? Get a paper trail, LOOK AT the paper trail, require disclosure
1. Require voter verified paper trails. Require manufacturers to retrofit their machines at their own expense. Do not give them a choice. The reasons: 1) They did not disclose known errors when they sold the system and 2) Like the rest of us, voting machine manufacturers are accountable for understanding their own industry. Shoup and Sequoia have been in business for over 100 years; obviously they knew vote-rigging was an issue. If your bank, for example, waltzed around making addition errors and pretending it was unaware that people might try to steal money, it would be held liable. Our vote is even more precious than our money, so make them accountable!
2. Allow voluntary comparisons of machine counts with hand counts by election officials in any precinct, for any reason. Because most precincts are small, at most 3,000 votes, this can be done on election night. Require spot checks, comparing hand counts with machine counts. Also, allow any citizen to get a hand count, if he pays for it. If this reveals a significant error when compared with machine counts, refund his money because he has done a public service. This can provide revenue for public universities, and by the way should be made available at very reasonable cost.
4. Require full disclosure of errors in future sales presentations and to the news media. Many industries are held to this standard. If you have to tell your next customer that your machines lost 103,000 votes in Florida, 41,000 in Texas, you had to recount the whole state of Hawaii, and Venezuela had people marching in the streets, you become much more careful to make sure the machines are accurate the FIRST time around.
Also, require disclosure of the names of owners and key people at voting machine companies. We have a right to know if people have criminal backgrounds, or if owners are running for office with their own machines counting the vote.
And, require more thorough inspection of the code. Currently, the manufacturers do not permit even the testing labs to do a thorough job of testing the software programming itself. Require excellent documentation of each step to be kept on file at the manufacturer. Allow spot checks by government or citizen auditing groups.
5. Require appropriate remedies when the machines miscount. This may sound obvious, but many of the errors uncovered in Black Box Voting were never corrected, even when machines elected the wrong candidate (which they have done several times, even when the election is not particularly close).
Are the above five safeguards in place?
In the new Help America Vote Act? (no) In any state statutes (no)
(In accordance with Title 17 U.S.C. Section 107, this material is distributed without profit to those who have expressed a prior interest in receiving the included information for research and educational purposes.)