Bad Ideas For Voting Just Keep Coming by Charles E. Corry, Ph.D.
Bad Ideas For Voting Just Keep Coming by Charles E. Corry, Ph.D.| EJF Home | Where To Find Help | Join the EJF | Comments? | Get EJF newsletter |
| Vote Fraud and Election Issues Book | Table of Contents | Site Map | Index |
| Chapter 7 — Building Better Ballot Boxes |
| Next — To Register Doubts, Press Here by Sam Lubell |
The way it was when voting worked
Bad idea #1 — Electronic voting machines
So why should such secret and illegal methods of counting ballots be used?
Problems with optical scanners to count ballots
Electronic voting machines commonly invent or lose ballots
Direct recording electronic voting
Unassisted voting for the handicapped
Adding up the cost of electronic vote counting
How to hand count paper ballots
Bad idea #3 — Increasing voter turnout
Bad idea #4 — Early, absentee, or mail balloting
Why precinct voting stops fraud
Losing ballots by the thousands
Bad idea #6 — Make election day a holiday
Bad idea #7 — Conduct many elections
Bad idea #8 — Do away with secret ballots
Bad idea #9 — Avoiding detection of election fraud
Appoint committees of the unqualified
Sign non-disclosure agreements with your vendors
Use your own judges if it goes to court
Bad idea #10 — Internet voting
Prof. Douglas Jones of the University of Iowa computer department presents a very nice Brief Illustrated History of Voting. For those who would like additional background Alexander Keyssar's book The Right To Vote is recommended.
Not too long ago we had a nice, simple election system and, for the most part, it worked. Citizens went to the county clerk's office in person, proved they lived at a particular location and were citizens of the state and the United States, and registered to vote. The county clerk then duly entered their name in a precinct poll book.
On election day good citizens went to their local precinct, again in person. When their names and addresses matched what citizen election judges had marked in the poll book, they were given a paper ballot that contained only the races and candidates applicable to that precinct and only local issues that the citizen was eligible to vote on based on where they lived. For example, county residents did not get to vote on issues or candidates for a city election.
In a private booth the good citizen marked his or her paper ballot in secret, free from any intimidation or observation by anyone as thousands of years of experience had taught us was necessary for a democratic election.
When the citizen finished marking the ballot by hand any stub or other means of identifying the ballot needed by the county clerk for tracking ballots was removed by an election judge without looking at the ballot, which was usually covered by a secrecy sleeve or folded. Blind or infirm citizens were helped to vote by citizen election judges to the extent necessary.
The ballot was then placed in a locked and sealed ballot box. When the polling place closed at the end of the day, citizen election judges, at least one from each major party and not public employees, opened the ballot box and counted the ballots by hand under the gaze of poll watchers. As a typical precinct usually had about 1,000 to 4,000 registered voters it was extremely rare that more than 1,000 ballots had to be counted, and the task was usually done in a couple of hours.
Election judges posted the results of their count on the door or window of the polling place for public view and then at least two citizen election judges, one from each major party, took the paper ballots and the results of their tabulation (total ballots cast, total ballots remaining, totals for each candidate, and totals for each issue) to the county clerk's office.
The county clerk then manually and publicly added up all the results for each race and issue, checked that the number of ballots cast plus the number of ballots remaining equaled the number of ballots issued to that precinct, posted the results in public view, and notified the press. Typically this was completed by midnight of election day.
If the results for a race were very close, or something else went wrong, the hand-marked paper ballots could be recounted. Again this was done in view of the public and press by citizen election judges because history showed public officials should not be trusted. Again, election judges from each major party had to be present and participate in the counting because Republicans tended to win if Republicans did the recount, and Democrats tended to win if Democrats did the recount (see Chicago Rules of Election Fraud). Also, poll watchers from both parties and independent citizens observed the recount. Not an exciting job but hardly impossible.
Election fraud was still possible, as shown by the Chicago Rules of Election Fraud, but it was difficult and logistical problems tended to localize the scale on which fraud could be committed. That is, a local race or special district election was easier to manipulate than a county race, and state and national election fraud was difficult but not impossible. So, for the most part, citizens had confidence in the election process and the genius of the Founding Fathers in providing a method of ensuring the smooth transition of power was apparent to all.
There is an ancient adage to the effect: "If it ain't broke, don't fix it." But, if one wants to control power and public spending, elections are the place to do it. And all the gee-whiz technology, labor-saving machines, high-pressure developers, cheap politicians, and fast-talking, big-spending salesmen proved simply irresistible to country-bumpkin county clerks. Of course money and corrupt individuals have never been in short supply with regard to election fraud. As a result, county clerks, who often seem more mentally suitable for shoveling out feedlots, were unable to resist the temptation to tinker with elections and begin "improving" the voting system at the behest of politicians and big-money developers.
As the United States grew it became more laborious for county clerks to tabulate election results manually. These individuals, seldom the brightest bulbs in the district, were as error prone as anyone, and big numbers and long columns of addition tended to confuse and frighten them. Also, precinct results tended to come in late at night, and the thought of sleeping before they came to work when the polls closed just didn't occur to them. So the poor dears were tired when they tried to add up all those numbers. That led to more "mistakes" in which their favorite candidates tended to win.
Now most of us will use an adding machine in preference to balancing our checkbook by adding and subtracting by hand on pencil and paper. And Quicken on your computer is even better yet.
Well, for one, in most states it is illegal to count ballots secretly, and rightly so. Vote counts done in secret, with undisclosed methods, using "trusted and trained" people who work for the equipment vendor or the county clerk, tend to come out with a little different election results than exit polls, or other cross checks might suggest.
No surprise there and that has been known for centuries.
• How much do you know about how Quicken balances your checkbook on your home computer?
• How much you know about how that black box at your polling location counts your ballot?
Coming up with some zeros here? Worse, even when fraud, or gross incompetence are evident (hard to tell the two apart), the courts won't let even experts examine how those black-box-voting machines work. Adds up to a secret ballot count, doesn't it?
As Stalin said: "Those who vote determine nothing, those who count the vote determine everything." And you now know nothing about how your vote is counted, nor do election judges, poll watchers, or even the county clerk. Only the technician who programs the vote counting computer and the manufacturer know what that machine does with your ballot. Trust us is their motto! And doubters and sceptics are "ivory-tower academics who don't know how voting works" or radicals who wear tinfoil hats and are taking time off from reporting black helicopters to attack Diebold. Of course, when things go wrong, and they all too frequently do, the voting machine manufacturers blame everyone but themselves.
We've been using printed paper ballots for more than a hundred years. There's no big government grants, no being wined and dined by manufacturers if you use paper ballots, no big equipment shows, just blah paper ballots. Anybody can count paper ballots and get it right.
The technology is magic and writing a voting program is child's play for an experienced programmer. In fact, those voting programs that have leaked out look like a child wrote them.
There are millions of state and county, and billions of federal dollars to buy computer voting machines with. You aren't going to get on the gravy train if you stick with plain old paper ballots. No big conventions to go to with paper ballots and being a clerk is quite dull otherwise.
As a result it has proven quite irrelevant in most cases that virtually every computer expert, of whatever specialty, e.g. security, networking, software development, database, etc., who has had a chance to examine computers used for counting ballots has repeatedly and loudly decried their use and design. In fact, the only people who "trust" computer vote counting equipment are the manufacturers and the election officials they've conned or bribed into buying the machines.
There are three types of electronic vote counting machines currently in common use. Punch card readers that feed a mainframe computer (these are being phased out under HAVA ), direct recording electronic (DRE) touch screen machines, and optical scanners that read a paper ballot and convert the markings on the ballot to a tabulation of votes (maybe).
The typical assumption is that optical scanners are the most accurate of these methods. Unfortunately, that probably is not the case. Optical scanners are far more widely used in grocery stores for checkout with nationally-standardized product UPC bar codes and have been since the 1970s. Therefore, one expects the major bugs have been worked out of the system. Grocery store scanners also have the great advantage that the customers gets a printed receipt showing what the scanner read, unlike voting optical scanners where the voter has no idea how the scanner interpreted their ballot. Since bar code scanners are in such common use they provide a reasonable comparison for the error rate we might expect from ballot scanners. In a survey of 24,000 readers Consumer Reports (October 2006, p.37) stated that seven percent (7%) of the respondents found the store scanners got the total wrong. Are we then to accept election results that are likely no more accurate than ±7%?
Lets look a real simple idea. The county clerk prints up some paper ballots. She or he then distributes the right ballots, i.e., proper ballot styles, to the various precincts. Voters come to their individual precinct, identify themselves and verify they still live at the same old address. They are then given a paper ballot and a voting booth where they can mark the ballot in private any damn fool way they want to with most any writing instrument imaginable. Doesn't matter if they scribble on the ballot, draw outside the oval or make an > × < instead of a connecting line >—<, use pen, pencil, charcoal, or crayon because Mark One human eyeballs of both Republican and Democratic persuasion are going to look at the ballot. 9,999 times out of 10,000 these citizen election judges and poll watchers will have no problem determining that Joe Voter wants Bob Feedlotshoveler to be county clerk for the next four years and he sure doesn't want that big tax increase they've been pushing so hard.
The reality is a little different with the new voting machines and if we count paper ballots with optical scanners we can run into all sorts of difficulties.
With optical scanners the county clerk still has to print up and distribute a bunch of paper ballots. So that cost and bother are still there plus the cost of the scanners at each precinct. Now, in addition, technologically-challenged county clerks are required to hire Diebold, or some other company of equal repute, at great expense to program the hundreds or thousands of voting machines purchased at a cost of around $4,000 each from this "reputable" company. The cost for this programming and testing will amount to at least $1 per ballot cast so this is big money.
Each of the hundreds or thousands of optical scanners will need to be individually programmed. Once programmed each and every machine should undergo a "logic and accuracy" test to determine that it functions and counts the ballot style it is programmed for correctly. But that's a lot of work. So typically only one machine per precinct or ballot style, whichever is lesser, is tested after the programming is done. And what election officials mean by "logic and accuracy" testing is a far cry and a long, long way from what I would consider proper testing. For example, the voting machines are checked only in "Test" mode, never in "Election" mode. And the "Test" mode only uses 100 ballots to see if the machine gets the same count as the technician and election officials using nice, proper, carefully filled out ballots with only the correct marking device. And never, ever are test ballots marked by handicapped or vision-impaired voters.
Remember taking all those tests where you were told to use only a No. 2 pencil to mark your answer? Optical scan ballot counters have the same limitations. Some will read only black ink, some will read only pencil, and for all we know some read only what they want to read.
In a November 2003 mail-in election in Garfield County, Colorado, on the instructions it said to use a black pen, on the ballot it said to use a pencil. The optical scanner could basically only read pen markings, though sometimes a lead pencil marking would be read. In a March 2004 election in Napa County, California, the optical scanner was calibrated to detect carbon-based ink, but not dye-based ink commonly used in gel pens, and did not properly read some ballots. Also, some ES&S optical scanners use red light to scan the ballots. Anyone with a basic understanding of optics and physics will realize that a ballot marked with red ink or pencil cannot be read by a scanner using red light to illuminate the ballot.
Problems with a September 7 th primary election in Maricopa County, Arizona, led Prof. Doug Jones to review the problems with writing instruments used to mark ballots read by optical scanners. A whole lot can go wrong with absentee/mail in ballots.
It has also been found that when mail ballots are folded and mailed that, in the crush of the mailbag, some of the toner or ink used to print the ballots may transfer on to the other side of the ballot. If that transfer overlaps a ballot position not selected by the voter then an unintentional overvote may result, or a vote may be cast that the voter did not intend.
In a number of elections it has also been found that the scanner heads become dirty or scratched. The obstruction on the read head translates into a line, or blank on the ballot. Where that blank space overlaps a vote the scanner can't read it and no vote is recorded, nor is an error generated. One way to detect such errors is a review of undervotes but many counties don't report undervotes.
There is also the problem of extraneous marks on ballots as they really mess up an optical ballot scanner. So if the voter doesn't make the exact type of mark required on the ballot, within the limits indicated, and using the required marking device in the proper color, their ballot simply won't be read by the scanner. For absentee or mail ballots, the voter may set their coffee cup down on the ballot at home or office, or spill orange juice on it, or some other household incident may occur that stains the ballot. No problem for a Mark One eyeball but the scanner now can't read the ballot. In some jurisdictions, when the scanner rejects such ballots, election judges will mark up another ballot "supposedly" following the voter's intent. But it is anybody's guess whether such "spoiled" ballots are counted.
Optical scanners used to count ballots in a precinct will reject any ballot with an overvote or stray markings. The mismarked ballot is returned to the voter, who is given a replacement ballot to try again with. But when optical scanners are used to count mail in ballots, e.g. absentee, election officials usually turn off the feature that rejects ballots with overvotes or stray marks to facilitate a speedy count. Thus, any mismarked ballot goes unchecked through the machine and proper votes may or may not be counted. Greg Palast, in his book The Best Democracy Money Can Buy, documents (p. 18) how that feature may have been used to eliminate Black votes in the 2000 Florida presidential election. In predominately white Leon County (Tallahassee and the state capitol) voting machines automatically kicked back faulty ballots. However, in very Black Gadsden County, the same machines were programmed to eat mismarked ballots. In fact, throughout Florida he found a strong correlation between race and the percentage of ballots that were not counted for whatever reason as shown in Table 9.
So just like back in grade school, the device you use to mark your ballot with can determine whether your ballot is counted or not. A real problem when one third of ballots are now absentee ballots marked at home with whatever pen or pencil is handy. Or in Oregon's case, all ballots are marked at home and mailed in. And you have virtually no way to determine whether your mail in ballot was marked with a device the vote counting scanner could read and, thus, whether or not your ballot was counted.
There are also persistent rumors that optical scanners have been programmed to give undervotes and overvotes to default candidates of the programmer's choice. Since it is possible to do that, and there is no way to test for such a possibility, such rumors simply further erode citizen's confidence in electronic voting.
Aside from the potential for fraud, why not simply hand count ballots in the first place?
Repeatedly, election districts using electronic vote counting machines have come up with records of many more ballots cast than registered voters in the district. One of the more blatant examples of this occurred in a November 2003 election in Boone County, Indiana, where 5,352 voters somehow cast 144,000 ballots.
When these machines aren't adding ballots, they're losing them. For example, in the November 2004 election an electronic voting machine in Carteret County, North Carolina, lost 4,438 votes. The machine had mistakenly been set to keep roughly 3,000 votes in its memory, which was not enough. And in a spectacularly poor design decision, it was programmed to let people keep "voting" even when their votes were not being saved. There are no tests at any time or any place that might catch such mistakes. And this is only one of hundreds of such "glitches," or "technician errors," or "programming errors" reported with computer voting machines.
Election officials response to such problems: "There are problems with paper ballots as well."
Inventing or losing ballot problems occur whether the vote counting is done by punch card (uses a computer to count the ballot cards), optical scanners for hand-marked paper ballots, or direct recording electronic (DRE) touch screen machines. While such problems do happen with hand counted paper ballots the magnitude is much smaller and more easily recognized and corrected.
The problems with ballot counting are even worse with direct recording election (DRE), or touch screen, computers. These are basically a Game Boy toy where you get to pick your candidate from a computer display (requires Windows Media Player — 2 MB).
According to Dr. David Dill, professor of computer science at Stanford University, all elections conducted on DREs "are open to question." Challenging those who belittle the danger of fraud, Dill says that with trillions of dollars at stake in the battle for control of Congress and the presidency, potential attackers who might seek to fix elections include "hackers, candidates, zealots, foreign governments, and criminal organizations," and "local officials can't stop it," and might well foster or condone election fraud.
During a public talk on The Voting Machine War for advanced computer-science students at Stanford in the fall of 2003, Professor Dill asked,
"Why am I always being asked to prove these systems aren't secure? The burden of proof ought to be on the vendor. You ask about the hardware. 'Secret.' The software? 'Secret.' What's the cryptography? 'Can't tell you because that'll compromise the secrecy of the machines.' ... Federal testing procedures? 'Secret' ! Results of the tests? 'Secret'! Basically we are required to have blind faith."
As part of the MIT/Cal Tech Voting Technology Report Project, Prof. Ted Selker from MIT has pointed out that "A common practice for local election officials is to let election companies run their election — make up their ballot, set up their machines, and even count their tallies. This is a dangerous practice." In fact, under many contracts with voting equipment manufacturers using proprietary equipment and software, election officials may have little choice but to pay the vendor to do exactly that, particularly in small counties or where the election officials have little in-house technical expertise (very common).
For about five years I was actively involved with the Institute of Electrical and Electronic Engineers (IEEE) P1583 voting equipment standards committee, testified repeatedly as an expert witness on computer voting in court and before the Colorado legislature, worked as a precinct election judge, had numerous discussions with local election officials, and written extensively on this topic.
As a result, it is my opinion that if you have confidence in the accuracy and reliability of electronic vote counting machines you're way too friendly with Jack Daniels, or you've been smoking far too much of that Maui Wowee. And DREs are just the tip of iceberg elected(?) officials have steered us into.
Also, these machines supposedly make it "easier" to count the ballots (keeping the counting process secret sure makes it easier), and speeds up giving the press and candidates news they couldn't possibly wait a couple of hours to hear. And we constantly hear from election officials how much voters love those Game Boy DRE's. Of course I haven't heard that from many actual voters but then I don't mingle much with the politically-correct crowd.
A much-touted "advantage" to computer voting is that overvotes can be detected and prevented and that unintentional undervotes will be detected with DRE's. That "advantage" isn't borne out by experience in Miami-Dade County, Florida, however. In fact, in New Mexico it was found that DREs may cause unintentional undervoting.
Also optical scanner failures lost thousands of votes in two Arkansas counties in the November 2004 election due to apparent overvoting. In one county ES&S attributed the error to a scratch on a sensor in the optical scanner that probably occurred during the election.
Another problem is that DRE's may cause voters to inadvertently undervote. After the 2004 presidential election generated extraordinary numbers of undervotes in a deposition, Terry Rainey, CEO of Automated Election Services explained that "if you go to a [push button] DRE machine and you walk in, the first thing you' re presented with is a list of political parties, and if I...say, yes, I'm a Democrat, and I push the button for Democrat, then that activates vote choices for all the Democratic candidates." Fair enough, most states in the country allow for "straight party voting".
But Rainey went on to explain, "if I decided I wanted to vote for Senator Kerry, and I push that button again, I have deselected my vote. And if I'm not aware that that's the case and I push the cast vote button, then I have lost my vote." He went on to speculate "that's why I believe it's so much higher in DRE counties. You don't see that...high level of undervoting in primary elections where the straight party option doesn't exist."
So if you were an occasional voter, or a voter who had registered for the first time, or a voter with limited language or computer skills, or for any reason didn't understand how the straight party option functioned, you could easily erase your own vote.
In truth, many election officials don't even tabulate overvotes and undervotes so a "fix" for those problems, even if it worked, is of dubious value. And with DRE's it is all too common to find the number of ballots cast by the machines is quite different from the number of voters who signed the poll book.
• We had a perfectly good voting system: Voters were given a paper ballot, verifiable by the voter as he/she marked on it using any available writing instrument, and all identification was removed before the ballot was dropped in a locked ballot box;
• After the polls closed election judges from at least both major parties, plus poll watchers, and other observers, counted the paper ballots;
• Election results were available within a few hours;
• Recounts were no problem: In the case of any dispute the paper ballots, which required no special storage, were hauled back out and counted again
Now under direction from ill-informed election officials and a critter with the head of an elephant and the rear end of a donkey, we have replaced this system with toy machines where:
• Voters poke buttons on a screen;
• The machine does whatever it wants to do and neither the voter, the election judges, or the county clerk know what the machine does;
• The machine spits out a count for (hopefully) each candidate and issue on the ballot, or maybe a few more or a few less issues and candidates if the ballot style is wrong;
• Everyone is told to believe the numbers the machine spits out because computers and programmers are never wrong or make mistakes;
• In case of questions a recount is impossible.
Then there is the problem of being able to do a recount if Murphy steps into the election arena. Due to overwhelming citizen demand as a result of numerous elections screwed up by DRE's, manufacturers are being required to add a voter-verifiable paper ballot printer to the touch-screen voting machines.
If I understand the situation, elections started out with a cheap, printed paper ballot. Now the government is forcing voters to use a very expensive (DRE's cost around $4,000 each) that counts their vote in secret.
In response to citizen's very real concerns about election integrity using DRE voting machines, taxpayers are now being required to foot the bill for adding a printer (~$1,000 more per machine) to verify that the machine is casting votes the way the citizen wants. And this is aside from the fact that most such printers will use a take-up reel that will make it quite easy to determine who voted for whom and for what, thus effectively eliminating a secret ballot. Further, at least with Diebold TSx machines, only about 75 votes can be recorded before the paper spool must be changed by a trained technician rather than an election judge.
Then there are the issues and costs of programming and testing each DRE for each election, which most counties cannot do without outside technical help. And that says nothing of the costs for the repeated upgrades and replacements these delicate computers will require at least every 4-5 years.
It seems to me we've simply gone full circle back to paper ballots at a cost of billions while losing our secret ballot.
Currently, if Joe Voter has some problems seeing or writing he can bring in a relative to help him or a couple of the election judges, his neighbors, help him as much or little as he wants. If he's in a wheelchair the election judges clear off a space for him to vote in private and maybe help him get around if needed. Right neighborly and friendly with a lot of greetings and howdies exchanged along the way as everyone gets together at their precinct to vote.
Now lets take a look at how this is done with a modern computer voting machine that solves all the nonexistent problems for handicapped voting (turns out Diebold apparently paid handicap advocates to promote this idea).
DRE's are being mandated on the premise that they will allow handicapped voters to vote unassisted. In practice it hasn't quite worked that way. Blind voters have been severely critical of these machines in an actual election, as distinct from staged demonstrations by vendors and handicap organizations the manufacturers are financially supporting.
There is no evidence to suggest that current DRE's will make it any easier for people with other handicaps to vote that is better than what they can currently do with a paper ballot. Indeed, small knobs, switches, and small areas on a DRE screen may make it harder for those with prosthetic devices, arthritis, palsy, and other disabilities. And many people are simply intimidated and afraid of computers.
The best of the current machines allow vision-impaired voters to cast their votes using a screen with large print and high contrast, headphones, a Braille keypad or foot pedal. For physically disabled voters such as paraplegics, a sip-puff tube can be used to assist in marking the ballot. As of August 2006 few of the DREs available have all these features.
Cost factors that manufacturers and election officials somehow always forget to mention include upgrades, service contracts and maintenance that by the terms of the non-disclosure agreement can only be done by the manufacturer.
Additionally, election computers are supposed to last for ten years or twenty elections. Typical life span of a computer today is 3-4 years. More tax dollars in the vendor's pockets for needed upgrades and "critical security updates." And I'm ignoring the hidden costs of kickbacks, bribes, and wining and dining election officials in order to get them to buy these machines in the first place.
Voting computers must also be kept in expensive, secure, climate-controlled storage between elections.
Conversely, paper ballots are cheap and easy. No money in them for upgrades, replacements, service contracts, storage, maintenance, or kickbacks.
• Electronic voting machines are not independently tested;
• Have a meaningless certification;
• Are subject to malicious intrusion (hacking) and computer viruses;
• Are often mis-programmed for a given election;
• Logic and accuracy testing is a farce;
• Ballots are counted by a secret process known only to the vendor;
• Are known to lose ballots by the thousands;
• Repeatedly have reported many more ballots cast than there are registered voters; and
• Can only be used in a recount when problems occur if an expensive ballot printer is added to each DRE that will, for all intents and purposes, eliminate a secret ballot and require a hand count.
Additionally, electronic vote counting machines:
• Have a high initial cost (~$4,000 each, $5,000 with printer);
• Require expensive maintenance or replacement every few years;
• Require reprogramming and testing by skilled technicians before each election;
• Demand an operational skill level often far beyond citizen election judges;
• Require expensive environmentally-controlled secure storage, and;
• Are subject to tampering and errors on a scale unobtainable before.
Counting ballots must also be done at the same location where they are cast and the polling location should be at a precinct independent of the county clerks office because:
• Ballots in transit tend to become "lost" and data cartridges are easier to "lose" than ballot boxes;
• The number of ballots in the "box" changes, i.e., any 16-year-old hacker can rig a data cartridge or memory card, or
• The vote counts for candidtate(s) may "change" during transport or at the county clerk's office. See How To Hack The Vote for a step-by-step demonstration of how this is done with Diebold election systems. Other voting systems are no better protected.
Not all election officials are crooks but, history shows, enough are that all possible protections against election fraud must be taken in every election if integrity is to be maintained. Thus, election officials must be considered guilty until proven innocent.
Another problem unique to computers, especially those based on a Microsoft Windows operating system, is the virtual certainty that at some point the machines will become infected with computer viruses, trojan horses, worms, hack tools, spyware, back doors, and many other forms of malicious code. Such infections of computers using a Microsoft operating system are a virtual certainty today, p > 0.9, and most voting machines use Microsoft.
Such malicious code may come built in with the computer in the boot sector of the hard drive, on a separate partition, in firmware on a chip, or hidden in many other locations on the machine. An intruder may simply slip in a CD-ROM, floppy disk, smart card, etc., that installs malicious code. But the malicious code may well come hidden on a CD-ROM used for an upgrade or to install new software. And many programmers include back doors in their code to allow troubleshooting.
Or there may simply be unintentional bugs in the operating system or voting software that show up during elections. Many examples of such equipment problems have been encountered. Or the technician who programs the election may "accidentally" give one candidate's votes to someone else and vice versa.
There are also serious, and unresolved problems with the security of any network the voting machine may be connected to. Such connections include, but are not limited to, telephone, Internet or private networks, and even powerline communications. Of, in the case of wireless communications, no connection is necessary.
Even the simplest threat model, e.g., a teenage hacker, assumes a much higher skill level than the county clerk's office possesses. And more sophisticated, well-funded attacks must be assumed given the critical nature of elections.
Needless to say, hand counted paper ballots avoid these extremely serious and largely unaddressed problems.
Voting machine manufacturers and election officials make a great play of claiming their
voting machines are thoroughly tested
by an Independent Testing Authority (ITA). In fact, the ITA's are paid by the manufacturer and only test what their employer wants them to test. For example, security is
not
tested. No standards currently exist as to what must be tested or how. Elsewhere Dr. Avi Rubin has looked at these
Dirty Little Secrets
of ITAs.
After testing, the machines are supposedly certified by the National Association of State Election Directors (NASED). As best I can determine, no NASED member has any technical or engineering training or experience. Most of them are lawyers. So the NASED certification is a meaningless stamp. In Colorado for many years the Diebold certification dated from 1995-1996 when the AccuVote system ran on a UNIX operating system. Today, Diebold Accu-Vote runs on a Microsoft Windows NT system but, until recently, had not been recertified in Colorado.
Diebold has been caught repeatedly slipping in untested and uncertified software on their voting machines. But if these tests and certifications are worthless, what does it matter? Well, for one thing, in California that is against the law. For another, in a March 2004 election in San Diego County, California, more than 40% of the Diebold machines would not boot up. Numerous other problems with Diebold equipment were encountered in many other California counties in that election. Somewhat irritated, the California Secretary of State decertified Diebold voting machines after that fiasco.
There is also a persistent and disturbing rumor that some computer voting machines have been set so that an undervote or an overvote gives a default vote to the programmed choice.
I repeatedly hear that the numbers of ballots in today's elections makes it impossible to hand count them. Canada and European countries such as Switzerland, however, seem to be able to do hand counts quite quickly and efficiently. So what we are seeing is ignorance in action with regard to using such methods in the United States.
Therefore, and herewith I present a simple, cheap, fast, effective method of hand counting paper ballots that protects the honesty and integrity of elections, which electronic vote counting demonstrably does not. No doubt this simple outline can be improved and streamlined to the point even a county clerk might implement it.
• Voting is done primarily in precincts on Election Day, i.e., absentee voting is extremely limited.
• A typical precinct has 1,000-4,000 registered voters.
• All counting is done at the precinct.
• Issue separate, color-coded ballots for candidates in national, state, and local races.
• Issue separate, color-coded ballots for state and local issues.
• Optionally, use separate ballot boxes for each ballot color, or simply separate into separate color piles before counting.
• Each precinct will have a maximum of somewhere between 500 and 2,000 ballots to count when polls close.
• Count total number of each color ballot and compare with number of voters who signed poll book.
• Tabulate votes, undervotes, and overvotes for each candidate in each race or on each issue separately for each color ballot.
Takes less than two hours to complete and post and requires no special training of election judges.
With the high cost of voting machines, that we don't need, it is proposed in the name of "government efficiency" (an oxymoron) that we do away with the security of precinct voting, citizen election judges, printed poll books, and consolidate into a few dozen "voting centers" where "trained" county workers will "unfailingly" provide the correct ballot style (from among dozens to hundreds to chose from) to anyone who walks in the door with any half-baked identification. The idea is that anyone in the county can vote at any "voting center."
Election officials also cite the difficulty of training all the required election judges on how to use the new electronic voting machines. So, ignoring history and the "many eyes" principle, they want to replace citizen election judges with their own employees and often their own computer programs and programmers. Or worse yet, they will let the voting machine companies run the elections. Stupid doesn't even begin to describe these election clowns.
In the past, even if a "voter" registered to vote in many places and under many names they were discouraged by the fact that they had to show up in person at the local precinct polling place on election day. Worse, the election judges and poll watchers knew a significant number of their friends and neighbors who came in to vote, and might recognize a stranger. Logistics and risk then limited the number of times a "voter" might cast a ballot on election day. Thus, in-person precinct voting has proven one of the best guarantees against election fraud ever developed.
Also, in precinct voting there are almost always only one or two ballot styles so there is very little chance of confusion and a voter being given the wrong ballot.
Nonetheless, voting centers have begun to spread throughout Colorado and other locations in conjunction with the widespread use of DREs, which allow voters to vote any ballot style in the county. In Colorado's August 2006 primary a pattern of voters being given the wrong ballot style to vote on the DRE at voting centers emerged. In Denver City and County the Colorado Speaker of the House and another state representative were both given ballots that did not include their names. In Chaffee County a columnist for the Denver Post, a Democrat, was given a Republican ballot. How widespread this problem is is not yet known because in most cases voters are unlikely to know what ballot style they should get and the error almost certainly goes unnoticed and unreported in most instances.
HAVA requires at least one voting machine that allows the handicapped to vote unassisted in each polling place. That amounts to hundreds or thousands of such machines per election district, depending on size.
As a result it has become obvious that the DRE's that many election officials falsely claim are mandated under HAVA are expensive and that current voting machines, purchased at great public expense and profit to the manufacturers, do not meet any reasonable standards for handicap access. Therefore, current voting machines will have to be either replaced or radically upgraded.
One of the underlying reasons for proposing vote centers is this HAVA requirement that every polling place have at least one handicap-accessible machine for voters to cast their ballot unassisted. Voting centers would presumably lower the number of such machines. Instead of hundreds of precinct polling places in a county there would only be tens of voting centers. In practice it hasn't worked that way.
The whole idea of voting centers for handicapped voters is quite ridiculous in the first place. Many, if not most handicapped voters request an absentee ballot and don't go to their precinct to vote anyway under current practices.
Here is a better idea: Since handicapped voters will have to be transported to their polling place in any event, equip a few polling places in neighborhoods with nursing homes, or a high percentage of elderly, with handicap voting machines and let all handicapped voters vote in the few polling places that are ADA compliant. Or have the handicapped who want to vote in person go to the county clerk's offices. Those offices must be ADA compliant in any case.
The voting machines would then only be used for handicapped voters. Also, such machines are supposed(?) to be capable of handling all the various ballot styles. That would preserve the protections provided by precincts and citizen election judges for the great majority of voters. As is always good practice, such methods would also greatly limit the potential for fraud.
One of the most common errors encountered with electronic voting machines is that the number of votes registered on the machines does not match the number of voters who have signed the poll book. With paper ballots a careful inventory is a basic function of election judges but that is impossible with touchscreen DRE machines and the number of ballots cast on such machines may differ from tens to thousands from the number of voters logged in the printed poll book.
Conveniently, it is a fundamental requirement for voting centers that an electronic poll book be used. Ostensibly the reason is that election judges must be able to determine if the voter has voted elsewhere, has requested or returned an absentee ballot, has already voted, or, more basically, is registered to vote in the election district. But if we make the poll book electronic, as well as the voting machines, embarrassing situations where the number of votes counted do not match the number of voters logged in the poll book can, and will be easily avoided.
The requirement for an electronic poll book at voting centers introduces a myriad of additional issues as well. Problems include the requisite connectivity to a central voter registration database and associated consequences of reliability (no one can vote if the connection goes down or is overloaded), hacking, and the loss of a paper record of who voted (one of the greatest safeguards presently available against election fraud, incompetence, and simple errors).
As election officials move more and more toward electronic voting, costs have become prohibitive, particularly for maintenance and upgrades. Also, as of January 2007 there are very limited choices for vendor software to run voting centers with. And Sequoia disgraced itself in the November 2006 election in Denver, Colorado, with vote center/electronic poll book software it sold the election commission.
Due to all these factors, election officials are developing software in-house that is not subject to any outside testing or review, let alone meeting any standards. Such homegrown software is obviously a prime target for election fraud even if well written. The possibilities for bugs and viruses are endless, and of course our technologically-challenged election officials have never heard of Murphy's Law.
As of January 2007 existing vote centers are unable to report election results by precinct even though that is required by law and custom. Colorado county clerks using vote centers state that it will require an extensive effort, and an unspecified amount of money, to be able to report results by precinct.
Frankly, given the performance problems evident in many Colorado counties in November 2006, the smarter move would be to abandon the vote center concept entirely and go back to precincts.
Voting centers require from 5,000-10,000 square feet of open space to set up the election equipment and such facilities have to be ADA compliant to facilitate handicap access. That generally precludes use of homes, garages, apartment recreation rooms, schools, and other locations currently used free for precinct voting.
Clearly, such large facilities must be rented. No estimate of the facility costs for voting centers has been published but our election officials are commonly affiliated with the local Chamber of Commerce. So it is likely they have friends who gain financially from implementing voting centers.
A rough estimate suggests voting centers must move an average of ten voters per minute (10 voters/minute) through the polling place in order to keep lines at minimum length and waits to vote less than an hour.
Generally that has not proven possible. In the November 2006 election in Colorado many voters in counties using voting centers stood in line until after 1 AM the next day in order to cast their votes. It is obvious that throughput rates had been inadequately calculated. Thus, where voting centers have been implemented election officials have encouraged, if not driven voters to vote absentee or early.
There are also other logistical problems such as parking spaces for the now hundreds of voters showing up at each voting center. With hours-long lines complaints about parking, restroom facilities, seating, weather, food, childcare, time off from work, problems with health and handicaps, etc., are numerous. Obviously people are deterred from voting by such insane practices as "vote centers."
Many of the "improvements" made, voter registration drives, absentee ballots, early voting, computer voting, etc. are made in the name of "improving" voter turnout. One would expect then that there is a standard measure of "voter turnout."
In the November 2004 election in Colorado (requires Excel spreadsheet) what was done was take the number of voters who voted in that election and divide it by the number of "active" voters (those who voted in the last election) to show there was an 85% turnout in the latest election due to all the "improvements" that were made. Think about that for a minute.
The usual measure of voter turnout is to divide the number of ballots cast by the number of registered voters in the district. But there are problems with that approach as well.
We've all seen the "Get Out The Vote" drives:
• Been mailed the voter registration forms to send in;
• Motor Voter when we went to renew our driver's license;
• Had someone come to the door to ask if we're registered to vote;
• Been handed a voter registration form at the mall.
One would think that voter registration was somewhere near zero from these campaigns.
The real problems in most localities is that the poll books contain far too many registered or phantom voters. Alaska and Montana come to mind, but there are many other states and counties where the number of registered "voters" exceeds the number of eligible citizens in the county or state.
In reality, rarely are more than 75% of the eligible citizens registered to vote in an election district. Certainly if the voter registration roles contain names for more than 85% of the population tabulated by the census in your neck of the woods then there are big problems in your county clerk's office. For example, "registered voters:"
• May have been dead for many years. In Boulder, Colorado, some "active" voters were 119-years old;
• Moved out of the precinct, county, state, or country. Over 40 million Americans change addresses every year;
• May be in prison or have been convicted of a crime that revokes their right to vote;
• May not be citizens of the Untied States;
• May be registered two or more times under different names, e.g., married and single names;
• May be in a nursing home suffering from dementia or Alzheimers;
• May be registered in two or more counties or states;
• May not exist, i.e., someone filled out and mailed in a voter registration form with a phony name but valid address. This is but one reason voter registration needs to be done in person.
• Address does not exist. About 10-15% of requested absentee ballots are returned by the post office as undeliverable.
What is found in practice is that voter registration drives increase voter fraud. Good citizens, who are interested in the issues, candidates, and their country, register and vote without being prompted to do so.
Under the Motor Voter Law, anyone getting a driver's license or applying for welfare must be given a chance to register to vote. It is illegal to ask for proof of citizenship, which of course, makes it very difficult to keep illegal aliens from voting. Several states, such as South Dakota, allow absentee voting with no identification needed to fill out the absentee ballot.
As John Fund points out in his book Stealing Elections, there is "accumulating evidence that Motor Voter has been registering illegal aliens, since anyone who receives a government benefit may also register to vote with no questions asked."
Moreover, says Fund, an INS investigation in 1996 into alleged Motor Voter fraud in California's 46 th congressional district revealed that "4,023 illegal voters possibly cast ballots in the disputed election between Republican Robert Dornan and Democrat Loretta Sanchez." Representative Dornan lost that election by fewer than 1,000 votes.
The reauthorization of the Voting Rights Act in 2006 locks in the abuses that are becoming prevalent under the Motor Voter Law. It does this by prohibiting "any qualification or prerequisite to voting" that could have the effect of keeping just one law-abiding citizen from voting.
What are "qualifications" or "prerequisites to voting"? Consider that many states or localities require photo IDs to protect against voter fraud. But supporters of Motor Voter insist that requiring a photo ID to vote or execute an absentee ballot is a design to reduce voting. Similarly, some supporters of Motor Voter do not want dead people's names removed from the voter lists, because doing so might accidentally remove legitimate voters. Both measures — requiring photo IDs and list clean ups — are different ways to ensure against illegal vote fraud.
Problems with current voter registration practices aren't bad enough that they can't be made even worse. Always wanting to increase voter turnout by increasing the number of registered voters, a number of groups and individuals have proposed that people be allowed to register and vote on Election Day. Such procedures are now permitted in Minnesota, Maine, Wisconsin, New Hampshire, Wyoming, and Idaho.
These procedures allow anyone to show up at the polls and present a driver's license or some other form of identification, e.g., a student ID or utility bill with their name and address on it, register on the spot and vote. Obviously there is no concern in this method for election fraud.
Even Colorado's secretary of state, Donetta Davidson, one of the most incompetent election officials we are aware of, recognized (or perhaps she was coached) that there would be no possibility that county election officials could check for fraud with same-day voter registration and that extremely expensive computer equipment, procedures, and training would be required to implement the electronic poll books needed to implement same day voter registration.
There is also the problem that the ballots of people (one hesitates to use the term "citizens" as they may or may not be) who register and vote on the same day are counted with every other ballot at the polling place. Conversely, current law requires that people whose eligibility to vote is in dispute, for whatever reason, be given "provisional ballots" that are separated and verified later before being counted. Thus, same-day registration creates a privileged class of voters whose bona fides cannot be checked by normal processes and whose ballots are counted regardless.
In 2002, when Proposition 52 to enable same-day registration was on the ballot in California, an aide to then Secretary of State Bill Jones warned that, if the proposition passed, busloads of people could "move into" a close, targeted district for a day and vote, then leave town without technically breaking the law.
And for those who believe there aren't any problems with same-day voter registration consider the following examples:
• In 1986 Oregon's voters overwhelmingly rejected same-day voter registration after a cult tried to register hundreds of supporters on Election Day in order to take over a town government.
• In 2000 a New York socialite working for the Gore campaign was caught on television bribing street vagrants with packs of cigarettes to register and vote in Wisconsin.
• After the 2000 presidential election in Wisconsin the U.S. Postal Service returned at least 3,500 Election Day registration confirmations as undeliverable. The inescapable conclusion is that several thousand people had registered and cast invalid ballots by means of same-day registration.
In the long term such measures as same-day voter registration only weaken confidence in the election process. Disillusioned by fraud and corrupt officials, fewer and fewer citizens even bother to vote.
In his book Stealing Elections, John Fund (p. 141) candidly states that:
"It's about time someone stepped forward and admitted that the root cause of low turnout isn't restrictive voting laws, but voter apathy. Some people are fed up with mediocre candidates, gerrymandered districts, and uncompetitive (and possibly illegitimate) elections."
Good county clerks (there are many) carefully check death notices, compare their poll books with the US Post Office's National Change of Address (NCOA) database, exchange address information with county assessors, check who is in the state prisons, and so on. However, election officials are often hampered by state and federal laws that mandate how long a "registered voter" must be retained in the poll book, typically two federal elections after the last time they voted in that precinct.
To "fix" the problems, enter those technological wizards, to wit Members of the Congress of the United States. Ignoring the fact that most current election problems are the result of previous laws they've passed, e.g. Motor Voter, they passed another real winner in 2002 called the Help America Vote Act (HAVA).
One of the provisions of HAVA requires that by January 1, 2006, all states will implement a statewide voter registration database. Never mind that a relational database requires considerable technological sophistication and years of planning, as usual Congress passes a law and the "problem" is considered "solved." However, it is likely the development of a secure yet readily and rapidly accessible relational database with adequate referential integrity will be one of the most difficult and time consuming HAVA requirements.
From the perspective of April, 2005, with less than a year to go, lets see how statewide voter registration was playing out in Colorado:
• First, attorneys with no discernible technological training or database experience are in charge of the development;
• Secondly, development has been contracted out to a discredited company, Accenture, that is based in Bermuda. Florida has already fired Accenture for screwing up a voter registration database there;
• Third, the declared intent is security by obscurity and outside review is not invited, wanted, or allowed;
• Fourth, the Colorado Secretary of State, Donetta Davidson, was in the habit of appointing Blue Ribbon oversight committees whose members have no technical background and the only technical input sought or received is from vendors.
In my professional experience this is a recipe for failure. The present process for implementing such a database in Colorado is a classic example of how not to do this. After spending over $200 million dollars to develop a statewide welfare recipient database for roughly a half-million clients, Colorado has not been able to make that work in years. Yet by next January we are expected to believe the Colorado Secretary of State is going to develop a statewide voter registration database for five (5) times that many citizens who are registered to vote, at a projected cost of only $12 million using a company Florida has fired for screwing up the problem in that state.
On November 30, 2005, the following notice appeared in the Denver Post:
Colorado pulled the plug today on its problem-plagued voter registration computer system and will miss a January 1, 2006, federal deadline for having it up and running.
Dana Williams, a spokeswoman for Colorado's secretary of state, said the system had trouble registering voters and other problems.
She said a letter was sent to the data-processing company Accenture canceling the $10.5 million contract.
The state has already spent $1.5 million on the system, she said.
"When we saw these problems, we decided we were not going to throw good money after bad. We' re going to get this fixed as soon as we can," Williams said.
Accenture spokesman James McAvoy blamed state officials for delays in the project, and added: "If the state had not canceled the contract, there would have been sufficient time to test and pilot the Accenture system and train elections officials in its use." Adams County Clerk Carol Snyder said the state tried to do too much at one time, setting up a system that would update computers automatically from all agencies involved in elections, from county clerks to the Corrections Department.
Williams said the state may have to settle for a smaller system that would require county clerks and state agencies to submit information on voter registration once a day.
And the woman behind this failure, Donetta Davidson, was on the National Institute of Standards and Technology Technical Guidelines Development Committee developing voting standards and now serves as one of the commissioners on the federal Election Assistance Commission, having resigned as Colorado Secretary of State to take that position. That certainly doesn't build my trust and confidence in voting systems. Particularly since the man she mentored to replace her as Arapahoe (Colorado) County Clerk was described as a Doofus Too Dumb To Quit and forcibly removed from office.
We might hope that other states are doing better but initial reports don't support a great deal of optimism. For more on problems with statewide voter databases see the section on Accenture — Epitome Of Incompetence.
The National Commission on Federal Election Reform (PDF) co-chaired by former presidents Carter and Ford issued a comprehensive report in August 2001 on the trends toward early voting, mail in elections, and easy absentee voting. That report argues that such voting methods do not satisfy the following five essential criteria for sound and honest elections:
1. Assure privacy and a secret ballot while protecting against coerced voting.
2. Verification that only duly registered citizens can cast ballots.
3. Adequate safeguards for ballots against loss or alteration.
4. Assurance that the ballots will be counted and counted in a timely manner.
5. Fostering the communal sense of citizens voting together.
However, the Motor Voter Act of 1993 had made it illegal to check an individuals identification prior to allowing them to register to vote and required states to allow mail-in registration. Thus, it became trivial for individuals to register under numerous false names with virtually no chance of detection unless they did something really stupid. Once registered by mail, under current "no excuse" policies these nefarious individuals can then request absentee ballots for each "phantom" voter they've registered and cast as many ballots as they care to. The Help America Vote Act (HAVA) of 2002 attempted to remedy the worst of the problem by requiring identification to register and vote. However, the forms of identification are so lax, e.g., a utility bill or student ID, that there are no effective checks on voter registration or casting absentee ballots. And nowhere is there any check to ensure the voter is a citizen of the United States.
About early voting, Jeff Jacoby, a columnist for the Boston Globe, says that having so many people vote weeks early "is rather like having a judge announce that any juror may vote and go home" when the juror decides he or she has heard enough evidence.
In his book Stealing Elections, John Fund (p. 44) cites the following examples of how early and absentee voting permits voters to cast their ballots before obtaining useful and often telling insights into candidates personalities and records:
• Ross Perot suffered his famous meltdown on 60 Minutes, in which he accused Republicans of disrupting his daughter's wedding, only nine days before election day in 1992.
• Also in 1992, independent counsel Lawrence Walsh indicted then Secretary of Defense Caspar Weinberger and other figures in the Iran-Contra scandal only four days before Election Day.
• In 1996 the John Huang campaign fund raising scandal accelerated in the days just prior to Election Day. Bill Clinton is quoted as admitting the Huang issue prevented Democrats from regaining control of the House.
• In the 2000 presidential race news of George W. Bush's 1976 arrest for drunken driving hit five days before Election Day.
Maybe we should wait until the electors have heard all the evidence before they vote?
Before we go into elections with mailed ballots, it is necessary to understand that a fundamental requirement of classical election fraud is to first obtain a ballot. If you have one, or more ballots in your cheating little hands the rest becomes fairly mundane. Election officials, even elected former feedlot-shoveling county clerks, used to understand that a meticulous inventory of all ballots was fundamental to an honest election. They tracked every ballot printed, the forms and files used to print the ballots, every ballot style by number, the total number of ballots issued to each precinct. And after an election the number of ballots cast plus the number of ballots remaining had better equal the number of ballots issued and printed. That is every ballot printed was carefully itemized and there were no unaccounted ballots floating around.
Again, tedious work when hundreds of thousands of different ballots are printed in dozens of ballot styles, but it was done. After all, isn't that why we employ clerks?
Things are never so bad that it isn't possible to make them worse. Current voter registration policies clearly make it possible, and even encourage the son, the father, and the holy ghost to register to vote many times, in many places, and under many names. And lets not leave out mom and the kids, who can vote from home, at college, from the nursing home, or even from the grave.
In his book Stealing Elections, John Fund notes (p. 44) that:
"In 1998, former congressman Austin Murphy of Pennsylvania, a Democrat, was convicted of absentee ballot fraud in a nursing home, where residents who were not very much aware of their surroundings proved an easy mark. 'In this area there's a pattern of nursing-home administrators frequently forging ballots under residents' names,' notes Sean Cavanaugh, a former Democratic county supervisor. He says that many nursing home owners rely on regular 'bounties' from candidates whom they allow to enter their facilities and harvest votes."
It is also impossible to ensure a secret ballot with absentee/mail in ballots. Rather than the isolation of a voting booth where no other can witness what the voter marks, with a ballot in hand outside a polling place the voter is subject to numerous pressures and influences on how they vote. In many cases with absentee ballots voters are coerced or "helped" by employers, labor unions, social or political clubs, churches, nursing home administrators, relatives, or others to vote for particular parties or candidates.
Absentee ballots also make buying votes quite easy. With polling place voting the buyer has no guarantee the voter actually voted as they promised when they sold their vote. But with an absentee/mail in ballot the buyer can be sure the votes stay bought.
With mail in ballots campaigns also engage in practices such as collecting ballots or marking envelopes with party affiliation. Only those ballots from the "correct" party, interest group, gender, or address actually make it into the ballot box. Absentee ballots are also commonly stolen from mailboxes and registered voters who don't usually vote are unlikely to alert election officials if they don't get a ballot. Or the people have moved or died and their home or apartment stands empty. The unscrupulous will request an absentee ballot in the recently departed's name(s).
The Florida Department of Law Enforcement states it bluntly in their extensive 1998 report on voter fraud issues:
• Absentee Ballot Fraud: — The desire to facilitate the opportunity for each person to vote has resulted in increased opportunity to use absentee ballots improperly. (Once one has registered fraudulently, he or she can obtain an absentee ballot for every election thereafter if he or she wishes. The lack of "in-person, at-the-polls" accountability makes absentee ballots the "tool of choice" for those inclined to commit voter fraud.)
And their's is the voice of experience after the most sordid and brazen absentee voter fraud in U.S. history during the 1997 mayoral election in Miami, Florida. But Miami is hardly alone in such infamous practices with mailed ballots.
Thus, by registering as many "phantom" voters as necessary or convenient, and then requesting absentee or mail ballots for these non-corporeal voters, one can easily vote early and vote often without ever once appearing before an election official. And safeguards such as signature verification are useless because the same individual who submitted and signed the registration form will sign all the secrecy envelopes on the absentee/mail ballots. The chances of being caught and prosecuted for such election fraud is miniscule.
From personal experience at our house, lets look at how several people were registered and could vote in Colorado Springs.
In the spring of 2001 Frances, who suffered from Alzheimers, was moved here from South Carolina and placed in an assisted living facility. Her daughter was given power of attorney as Frances isn't with us anymore. A couple of months later the "activities director" filled out, or "helped" Frances fill out, a voter registration form unknown to her daughter. A "witness" also signed the form but her daughter doesn't recognize any of the "witnesses." When election time came around in November 2001 an absentee ballot was requested for Frances, and in subsequent elections as well. However, in May of 2004 it was necessary to move Frances to a nursing home as her condition worsened. Her daughter also filed a change of address notice with the post office. One result was that Frances didn't "vote" in the November 2004 presidential election and a routine postcard from the county clerk was forwarded to her daughter in January 2005, which was the first time her daughter found Frances had been registered to vote in Colorado Springs.
Kids grow up, live at home for awhile, register to vote, then marry and move away. In our case the daughter moved away over eight years ago but we still get notices and absentee ballot request forms for her. She could still be voting here long after we're dead.
I've been divorced from my ex-wife since June 1996. However, on October 9, 2003, a jury summons for her was sent to my address in Colorado Springs. With the help of the deputy district attorney, the jury commissioner, and the election director for El Paso County we found that sometime in 1999-2000 she, or someone else using her name and my address, had registered to vote here even though she has never lived in this county. If she hadn't been summoned for jury duty we would never have known about this.
In the past, even if a "voter" registered to vote in many places and under many names they were discouraged by the fact that they had to show up in person at the local precinct polling place on election day. Worse, the election judges and poll watchers knew a significant number of their friends and neighbors who came in to vote and might recognize a stranger. Logistics and potential recognition then limited the number of times a "voter" might cast a ballot on election day.
While some ingenious methods were invented to circumvent the difficulties posed by precinct voting, casting a fraudulent ballot was difficult, time-consuming, and expensive, and election fraud was thereby limited. Precinct voting also makes it very hard to sell your vote.
If a voter wanted to obtain an absentee ballot they generally had to go to the county clerk's office and give a very good reason they couldn't show up to vote on election day. The secrecy envelope also had to be notarized when the ballot was completed and before it was mailed. Military personnel could request an absentee ballot but generally their commanding officer had to sign both the application and the ballot secrecy envelope verifying the individual was who they claimed to be. It wasn't easy or perfect but it usually worked.
But instead of trying to make election fraud more difficult, election officials decided to make "voting" easier. Again, the basis for this inane proposal was to "increase voter turnout."
One way to make voting easier, it was decided, was to allow "voters" to mail in their ballots. In fact, Oregon has gone entirely to mail in elections with a resultant comedy of errors. In an all-mail-in election, ballots are usually mailed to all "active" voters (typically an "active" voter is one who voted in the last election).
An inactive, or first-time voter usually has to request a ballot within a specified time frame. In practice what that does is disenfranchise almost everyone who didn't vote in the last election as they commonly don't realize they need to ask for a ballot. So the unintended consequence of a mail in election is often to deprive thousands who never get a ballot of a chance to vote.
To try and hide their folly, election officials then recalculated voter turnout by only including active voters. Number of ballots received divided by number of active voters is certain to be a much larger fraction than dividing the number of ballots received by the total number of registered voters. But you never liked fractions in the first place, so what does it matter?
Of course those intent on election fraud are very careful to make sure they receive all their absentee ballots, and equally careful to return all of them precisely filled out with proper postage.
Tens of thousands of mailed ballots will not be delivered, or delivered to the wrong person in a mail-in election or with liberal absentee balloting allowed. Some apartment dwellers may receive a half-dozen ballots. Fraternity and sorority houses will often have piles of ballots. Nursing homes as well.
Now ask yourself: What is to prevent someone from mailing in these ballots? The answer: Almost nothing. Election officials claim that the signatures on the secrecy envelopes are checked against the signature on file, which may even be true in some counties. But many people's signature is an indecipherable scrawl. So scrawl the name in. If the signature doesn't match the worst that is likely to happen is the ballot isn't counted. There are no standards for matching signatures on mailed ballots nor any tested and certified equipment used in elections for this comparison.
Of course those intent on election fraud will have mailed in their voter registration forms, often by the hundreds, and they will be sure to request ballots for every "voter" they registered. And you can bet their signatures will match the one on record because they're the ones who signed the registration form against which the signature will be compared.
Want to rig an election? Mail in voter registration forms for as many "voters" as you can find a useful address for, e.g., nursing homes, a fraternity where your son goes to college, empty homes (especially if you're a realtor or developer), business addresses you use for shady deals, empty apartments, and wherever else your imagination suggests. Usually half-a-dozen working addresses will suffice with multiple "voters" at each address. Be sure to collect mail from all these locations regularly and keep copies of all the registration forms you turn in. When election time nears you then mail in requests for an absentee ballot for each "voter" and away you go.
But what about the ballot inventory? Forget it in any election where ballots are mailed. Even where the election clerk has a signed request for an absentee ballot with an address to send it to, 10-15% of these ballots are returned by the post office as undeliverable. One might conservatively estimate that another 10-15% of such mailed ballots don't reach a valid voter.
No one really knows how many ballots just disappear in the mail in such elections but it is typically in the tens of thousands for a medium-to-large election district. So one more basic protection for election integrity down the drain while extending an open invitation for election fraud in the name of "increasing voter turnout" and convenience.
After the mail ballots are received they are then counted in secret by election officials typically using optical scanners with the inherent problems of those machines (see above) with the additional concern that the counting is being done in a back room of the county clerk's office by "trusted" county employees. Further, it is common practice to use an optical scanner to count the ballots with the overvote rejection feature turned off. So any extraneous mark on your ballot, that would be rejected in a precinct polling place, is simply passed through the backroom scanner and your vote not registered.
Our grandparents knew that absentee and mail balloting was a bad idea, and central counting a mistake. But our election officials have "forgotten" that hard-learned lesson citing "voter convenience" as a reason for tolerating election fraud with the added bonus for them of secret ballot counting by computer voting equipment.
Benjamin Disraeli, English Prime Minister, 1868 and 1874-1880
Popularized in United States by Mark Twain
Statistical sampling of voters at carefully selected precincts as voters leave the polling place has been successfully used in many countries to monitor elections for roughly 50 years. It is a standard measure of election integrity. However, like any statistical sampling it contains an error margin that is often larger than the difference in candidates final vote counts.
The accuracy of an exit poll also depends strongly on how well the sampling is done, i.e., enough voters are sampled, precincts sampled are representative, and the questions asked are answered by a sufficient number of voters to be a random sample. That is there is no over- or under-sampling of women versus men, old versus young, wealthy versus poor, black versus white versus Hispanic, and etc.
The types and numbers of questions asked, and how such questions are asked, also affects the accuracy of the poll. Training, experience, and age of the pollsters has also been shown to be an important factor in obtaining statistically-valid results. Also, if the pollsters personally favor one candidate over another, their poll results will tend to favor that candidate due to well known sampling bias.
Exit polls also contain the implicit assumption that basically all voters vote on election day at their polling place.
As electronic voting, early voting, and absentee voting have increased, the validity of exit polls in elections within the United States has decreased. Around one-third of voters now vote absentee or early, so the underlying assumption of exit polls that voters only vote on election day at their polling place is no longer valid. Thus, the error estimates on exit polls are greatly increased and the utility of such polls decreased.
The widespread claim that exit polls in the November 2004 presidential race accurately predicted John Kerry as the winner, and there was some massive election fraud that gave the victory to George W. Bush is without foundation or merit. The issue was examined exhaustively in a report by the Caltech/MIT Voting Technology Project released November 11, 2004, titled Voting Machines And The Underestimate of the Bush Vote (PDF). The error between the exit polls and the final election results was only 1.3%, well within the margin of error for such surveys. However, for the reasons given above, if absentee and early voting continue to increase, sampling errors will likely increase proportionately.
There has always been a problem with the probability that broadcasting exit poll results before the polls are closed tends to influence the outcome of an election. England, for example, prohibits by law the broadcasting of any exit poll results until all polling locations are closed. With election districts in at least twenty time zone around the globe, such restrictions are impracticable in the United States, and probably unenforceable in any case.
As a result, exit polls have likely become another bad idea for general elections and too expensive for local elections. One might reasonably ask if it is so essential to hear a guess a few hours in advance of the actual vote count?
For those who care to entertain speculation, Lynn Landes has put forth the idea that exit polls are simply faked, and that organizations such as the Associated Press are simply accessing voting computers. Whether that claim has any validity or not, the possibility is simply one more argument against the use of exit polls in American elections unless and until we go back to personal voting in precincts using hand-marked, hand-counted paper ballots.
First, we might consider that the only people likely to actually be given a holiday for elections would be public employees. Businesses would be very unlikely to close voluntarily. And since between existing holidays, vacation days, sick leave, personal days, and etc., public employees are currently only required to be at their work site about 30 days a year, I can't get real enthusiastic about giving them even more holidays.
Second, are we to make every election day a holiday? Currently there are primary elections, general elections, city elections, county elections, special elections, district elections, and election elections. Often a voter is called on to vote three to five times a year. Generally, greedy developers and others with nefarious goals and schemes have learned to take advantage of holding a special election where voter turnout is very low and it is easier to swing the vote their way. Thus, it would seem to be of more public value to make the local elections holidays if we were to pursue this idea.
However, I think we should simply discard this bad idea.
There are currently primary elections, general elections, city elections, special elections, and district elections. It isn't unusual for a citizen to be called upon to vote three to five times a year for obscure issues or candidates. After awhile citizens become weary of the hoopla, the attack advertisements, the endless posters and signs, and their primary interest is peace and quiet without some politician in their face and on their TV screen. If they do bother to vote it is more likely to be against some obnoxious feedlot shoveler than for an issue or candidate they support.
It seems obvious that having all these elections is a primary cause of voter disenchantment and dwindling turnout. However, the trend has been to have ever more elections and even Direct Democracy is touted. Some dimwits missed their civics courses where it was explained that the United States is a representative republic. We pay our representatives to handle the details for us so we have some peace and quiet and time to tend to our own business.
• Limit elections to one primary and one general election held during every even year.
Nothing in our cities, counties, states, or nation would collapse if the issue had to wait a year for citizens to vote on it. We might even benefit if some politician died and his position remained vacant for a year or two. Or, as is often done now, if the position was absolutely essential a temporary replacement could be appointed. But voter interest and turnout is clearly inversely proportional to the number of elections.
One of the really dumb ideas that has been floating around recently suggests we do away with secret ballots.
Since voting began in ancient Athens, and probably before, it has been found essential that citizens vote in privacy and without coercion. It is fundamental that no one else be able to see or influence how a citizen votes if honest elections are to result.
The secret ballot is the simplest and most widespread measure to ensure that an individual's political views are not known to anyone other than the citizen. Secret ballots in an election are nearly universal in modern democracies and considered a basic right of citizenship. In fact even where other rights of privacy do not exist, this type of privacy very often does.
Without a secret ballot there is scarcely any reason to have an election as the most powerful party or group will simply bully and intimidate others into voting as they wish. Voting intimidation has occurred from employers, political parties, unions, organized crime, and virtually every other human organization when the opportunity presents itself.
As noted in Wikipedia:
"Because politics is fundamentally about settling disputes of when to apply violence and violent restraint, and appointing officials to hold a monopoly on violence, it puts winners and losers in unique positions to revenge themselves on each other. In a democracy, the winner is vulnerable as s/he often appears in public to explain policies and gain additional support. The loser is also vulnerable as s/he is subject to the interpretation of criminal justice by winners — law enforcement is often very prejudicial, and an independent judiciary is not always available to ensure fairness in how winners of a political conflict deal with losers. Uncertainty about who supported what measure, and the right to keep one's opinion to oneself and not be required to reveal it except voluntarily (by joining a political party or answering opinion polls), aren't generally challenged even by the most strident national security advocates. In this sense, supporters and detractors of the state have a common position: when operating within the existing legal bounds of a political system, opinions should be measured only in aggregate."
But this ancient wisdom is little constraint on current election officials and they are seemingly doing everything in their power to erode the fundamental security of a secret ballot.
If you use the Hart Intercivic voting system like Boulder County, Colorado, they have already done away with a secret ballot despite a clear mandate in the State Constitution requiring the cast ballot have no marking or other identifying features from which the voter can be identified.
Adding printers with take-up reels for ballots on DREs will make it quite easy to determine who voted for whom and what.
One major objection to absentee ballots and mail in elections is that there is no way a secret ballot can be maintained except to have the voter complete it in the precinct. There are two primary ways an election is corrupted if the voter has the ballot in their possession outside of their polling place:
• One is the voter can now be subjected to intimidation by their employer, union, political parties, organized crime, neighbors, and other organizations and individuals, and whether or not he obeys their dictates is observable with a mail-in ballot.
• In the alternative the voter can sell his vote to the highest bidder as the buyer is then able to verify how the ballot is voted. Selling one's vote in a precinct polling place is stymied by the fact that the voter only gets a ballot after qualifying and must surrender the completed ballot before leaving. Thus, the buyer generally has no way to determine how the voter actually voted.
One reason cell phones are, or should be banned from polling places is that camera phones would allow a voter to photograph their ballot while in the voting booth and transmit the picture to a vote buyer. Many other methods of controlling and buying and selling votes exist but most are possible only outside of a precinct polling place.
It really looks bad on one's résumé if you're an election official, county clerk, Secretary of State, e.g. Arkansas, and you get caught in election fraud, theft, kickbacks, being bribed, or your gross incompetence publicly demonstrated. Of course, in the event that does happen you can always go to work for Diebold Election Systems. Probably step right into management with the other felons. But lets assume you don't want to pursue the same career path so many other election officials have followed and you like it fine just where you are. And all of us hate to get caught with our hand in the computer, oops, cookie jar.
So how can you use your position as an election official to avoid getting caught in an election fraud, or at least revealing the sheer incompetence law school taught you to be so proud of?
Here are just a few tips from some pros.
The fundamental challenge for an election canvass board is to ensure that the number of ballots counted does not exceed number of ballots cast. One might think these numbers should be equal but that isn't true for a variety of reasons, all too often related to election fraud. See rule 10 in Chicago Rules of Election Fraud:
10. Recounts: In the unlikely event our candidates don't win the first count, then demand a recount. Fill the recount room with loyal supporters, and tow away the cars belonging to the enemy. If you can't win a recount, then you are not a Chicago Democrat.
The way this is usually done is that you enforce the rules for disqualifying ballots to suit your taste and political persuasion. Done well, Republicans can openly disqualify any number of Democratic ballots with no one the wiser. Some experience playing poker helps. This game is also much easier if there are lots of provisional ballots, absentee ballots, and early ballots, many of which tend to get lost or somehow go uncounted if not of the correct political persuasion. This fact is simply another reminder that ballots need to be counted at the time and place they are cast in a precinct polling place under the observation of both major party election judges and poll watchers. Always and forever there are little "glitches" in the counting when it is done in a back room at a central location by "trained and trusted" personnel.
Of course if vote counting machines are added into the mix we add plenty of variations for the sophisticate to play to ensure only the truly anointed are elected. For example, using an optical scanner one can claim the machine didn't read the ballot correctly (and often they don't but you can't check that) so election officials pull the ballots and rerun them. Tens of thousands of ballots can be "rescanned" as in a Colorado Springs election in 2001. In poker they call this "stacking the deck." I don't know what to call it in an election but it sure amounts to a fast shuffle of the ballots. And it ain't no surprise that the arithmetic just never adds up. But there is no question the "right candidates" have won the election by an indisputable margin when the shuffle is finished.
Even the simple problem of finding the number of ballots that were counted after all the games are played and then comparing that number with the number that were cast in the election is beyond the capabilities of many election officials. And then for months after the election uncounted ballots will turn up in strange places. Also, it is quite rare now for the number of ballots printed to equal the number of ballots cast plus the number of ballots remaining.
The public always likes to know there is an oversight committee presumably reviewing the actions and decisions you're making as an election official. Obviously, however, if the people on this committee knew the first thing about the subject in question they could be dangerous to your career. As a first step then you need to be sure you appoint the members from a select group.
Other attorneys are always good on a technical committee. Their arrogance won't allow them to admit they don't know Polish from polish about voting, computers, security, networks, hacking, databases, operating systems, flip-flops, gates, buffer overflow, or other arcane subjects of basic interest to the use of computers or voting. Attorneys are especially attractive for your committee if they already hold a public office that keeps them occupied full time, for example a recently elected district attorney is perfect.
For election issues who could be better on your committee than other county clerks or election officials. They likely don't know even as much as you about computers and etc. but they are "voting experts" by definition. An IT technician from a county clerk's office who has never had any computer education or experience is a perfect candidate.
Round up some out-of-office, or even in-office politicians. Present or former state representatives, county commissioners, former mayors are good. You get the idea.
Provide a built-in scapegoat if things go wrong by hiring a known racist for your election staff to head up Help America Vote issues. That will detract attention from real election problems if and when needed.
Whatever you do, don't let anyone with an engineering or science degree on your committee.
Publicize that your meetings are open to public participation but make it extremely difficult for busy engineers or scientists to attend. For example, hold meetings at 9 AM on Tuesday mornings in an inconvenient location that is only announced the night before the meeting. Change the meeting location frequently and unpredictably as well.
In her version of this chimera, the Colorado Secretary of State also states that she: "will try to afford a window of time for public comment. Those wishing to speak must submit their topic or opinion in writing prior to the day of the meeting." So open discussion is ruled out and you have to make a formal presentation probably with a five-minute time limit. But she only promises to try and afford a window for you to speak. And that window, if it exists, will almost certainly be at the tail end of the meeting when everyone is anxious to leave.
Be sure the people you allow to make presentations on the topic of computer voting are strongly in favor of it. Diebold representatives are always available and will buy everyone on the committee lunch, dinner, and drinks as well.
Call them Blue Ribbon committees or task force groups. What fool citizen could possibly doubt that a Blue Ribbon committee is composed of anything but the finest experts available on the subject? Should citizens have the temerity to challenge you despite all these precautions, belittle them at every opportunity and always refer to them in derogatory terms, e.g., "uninformed academics who have no experience with actual elections."
Publish minutes so it looks like everyone is working real hard. A final report probably won't be needed, however, as it might show what a waste of taxpayer's money this was. After a few months of this charade announce the Blue Ribbon committee has agreed that the state (or county) must immediately purchase X-millions of dollars of Diebold voting machines. Or a cost-plus contract must be granted to Accenture to develop a statewide voter registration database, or any other damn fool project you want.
If you are a member of the public you shouldn't look too closely at how well some of the committee members are dressing or comment on how well they seem to be eating lately on a public servant's salary.
Look, if anyone with half a brain, or the least familiarity with computers and security were ever to get a close look at the voting equipment you've been buying and using, qui tam would take on a whole new meaning to you. Trust me, you really don't want that to happen.
Fortunately, there is an easy way to prevent such unpleasantness. Working with Diebold, or any other voting equipment manufacturer, hereinafter known as the Company, you sign a non-disclosure agreement that prohibits you, or anyone even remotely associated with you, from revealing any information whatsoever about the hardware, software, programs, procedures, processes, and so on. Just a few pages of fine print is all it takes.
The Company claims, in fine print, that all such hardware, software, programs, procedures, processes, and so on are proprietary and contain trade secrets. Of course it isn't and doesn't but if the contract says so who would dare doubt it?
As a result, under U.S. law, no one not expressly authorized by the Company will be allowed to review, examine, test, or even look hard at the outside of your voting equipment. Therefore, any fraud becomes impossible to prove, and any little "glitches" are either "technician error" if you and the Company absolutely cannot avoid accepting some blame, or the result of untrained, elderly, feeble election judges who just haven't kept up with the technology and didn't listen when they went to the training classes.
But sometimes even the best feedlot shovelers let things slip. Take El Paso County, Colorado, for example. Already "famous" for buying vaporware from Diebold, a Colorado Springs election being used as the type example of how to hack Diebold GEMS software, it just wasn't enough. In the November 2004 election (PDF-3 MB) we find that in precinct 5000021378 there were 71 registered voters but 175 ballots were cast. The Diebold report generator faithfully calculates that this was a voter turnout of 246.48%. What kind of ridiculous, amateur programming is this anyway? Then there is the curious case of precinct 5111821267 with 183 registered voters but only 28 ballots were counted, a turnout of only 15.30%. Strangely, virtually all the other precincts had a turnout of 60-80% and the county average turnout was 68.6% with 354,059 registered voters and, presumably, 242,888 ballots cast. Maybe some misplaced voters ended up in the wrong precinct?
My guess is no! The computer spits out the numbers and they have to be good. After all Diebold did the programming and who wouldn't trust them?
After enough of these incidents even the mildest citizen's sense of outrage reaches a boiling point and they take that old American remedy of filing a lawsuit. Fortunately, your official position often allows you to appoint your own administrative law judge to hear any court challenges of election results. Since its obvious no election official could ever do anything wrong or utterly stupid (like buy Diebold voting machines), the first, and often final step in your defense is to deny all attempts at discovery. Nope, those pesky citizens should not be allowed to look at election records and must not be permitted to examine the equipment used in the election (see non-disclosure agreements above).
But sometimes the judge doesn't cooperate. In those rare cases it is extremely important that you don't allow experts to look at your voting machines. That is why those non-disclosure agreements are so important to you.
Other creative steps that election officials have taken when the noose began to tighten were to lose ballot boxes or memory cards, and be unable to find election details by the time the issue gets to court. But these latter steps are much more difficult than simply keeping nosy engineers and programmers out of black boxes. And, of course, everyone knows that computer memories and hard drives are easily erased, by accident or misfortune obviously.
As though the ideas above aren't foolhardy enough, the proposal for Internet voting or Direct Democracy keeps sticking its head up. One feels as though they are combatting a hydra-headed monster. We might have hoped that with the end of the SERVE program this beast would have been slain but the life of a bad idea for voting is seemingly never ending.
I have tabulated below some more general concerns about electronic balloting. One or more of these problems will exist with any network election system implemented.
• The first, and currently insurmountable problem is there is no known way to ensure that the ballots entered into the election system via the Internet or other network are cast only by a valid citizen elector who continues to reside at their registered address.
• Who can vote for what and whom depends on where the voter resided when they registered. That may not coincide with where they live now.
This is a major problem with Internet voting that would allow ballot box stuffing from China, or anywhere else on the planet.
• In an Internet, or other network or computer balloting system, the voter marks no ballot that is preserved outside the computer.
• Computer counts of ballots may not be accurate due to either deliberate manipulation or flaws in the software, hardware, or network. There are numerous documented examples of this even without the added complexity of network voting.
• A "recount" of computer ballots is meaningless.
• Tabulation and audit trails are contained solely within the computer. An audit trail outside and independent of the computer, and manual or other independent means of recounting or verifying the balloting is impossible with network balloting.
• Communication links may be accidentally or deliberately broken or manipulated to control balloting. One may require transmitting information over a secure network but, in fact, no such network has yet been proven to exist (including Sipranet).
For example, a dummy system could easily be put in place indicating a vote was properly sent and recorded that, in fact, never reached the election computer. At the same time the dummy system could send a vote of its own to the computer doing the counting.
Or the communication link can be rigged to modify the vote in transit. Internet connections typically go through 15 to 20 separate machines en route to their destination.
• For Internet voting, vote modification may easily be accomplished by someone in another country not subject to laws of the United States.
• Adequate network testing of the myriad of ballots present in general elections, and even many local elections, is impossible as a practical matter.
• Computer or communication malfunctions can cause the inadvertent or deliberate loss of electronic ballots. Mandating uninterrupted availability while the polls are open looks fine on paper but ignores reality. Networks, no matter how much redundancy they have, do break down and Murphy dictates these breakdowns will occur at the worst possible time.
• Poll watchers and other independent observers cannot verify the accuracy of the balloting even if source code for the voting program is available to them.
• It is quite easy to present one listing of the computer code for verification while the computer is actually running a different version of the same program.
That may be done either by accident or deliberately.
• A substantial percentage of the electorate, estimated at approximately 10%, will not be able to use the computer either due to computer anxiety/phobia or other handicaps.
Efforts to ease those problems are certain to increase computer bugs or provide opportunities for deliberate manipulation of the vote totals.
• Inadvertent or deliberate "errors" or "bugs" in the computer code are virtually impossible to detect unless they cause gross mistakes.
A computer byword states that all nontrivial programs contain bugs. If there are no bugs in the program it is, by definition, trivial.
Voting algorithms with complex candidacies, initiatives, precincts, special districts, etc. are not a trivial programming problem.
By the same logic it is a certainty that I have not thought of all the potential ways a network election can be accidentally or deliberately corrupted.
• Transferring all vote counting to a networked central computer makes it impossible to determine local balloting errors and enhances the opportunities to manipulate the count.
• Programmers commonly leave themselves "back (or trap) doors" while developing computer code to facilitate testing and debugging. Such features facilitate later manipulation of the code, either authorized or unauthorized, and Internet voting simply makes such intrusions easier.
Possibly at some point in the future computers and networks will be so securely intertwined and their location so readily determined that network voting will be technically possible. But then issues of personal privacy will become of critical importance. Before proceeding down the road of Internet voting it is suggested that one familiarize themselves with George Orwell's 1984.
It is safe to assume that fraud will occur in any election unless active measures are taken to forestall it. Those interested in election fraud can be assumed to be well financed, technologically sophisticated, and highly motivated.
Election fraud is more likely in small, local, and special elections, particularly where developers are involved, than in large, national elections. The logistics are simpler and the number of votes that need to be manipulated to "win" are much smaller.
The chances of discovery for those committing election fraud are very low and prosecution, if discovered, is quite unlikely.
The best-equipped people to commit election fraud are election officials. And if computer voting equipment is used the IT technician is ideally placed. Also, it is extremely difficult to distinguish sheer incompetence from actual fraud in an election.
These statements reveal nothing new and not so long ago measures were taken to forestall the most egregious methods of election fraud. However, not content with tried and true methods, election officials, often with the support of an uninformed public, have been introducing and proposing many "new" methods of voting, ostensibly for the sake of "increasing voter turnout," "voter convenience," "helping the handicapped to vote," "government efficiency," or other high-sounding reasons and principles. All of which, in effect, validate the Chicago Rules of Election Fraud, and introduce some new and more efficient methods for fraud in the process, i.e., computer vote counting.
History clearly demonstrates that election fraud is minimized when:
• Voters are required to register in person with the county election office;
• Voters vote in person at a local or neighborhood precinct staffed by citizen election judges who are not public employees;
• Absentee ballots are allowed in only extremely rare cases, e.g., military personnel. And the soldier or sailor's commanding officer is required to sign the secrecy envelope to verify the individual was the actual voter claimed;
• Voters hand mark a paper ballot (note that there are many methods and devices available to help handicapped voters mark a paper ballot besides a DRE);
• Ballots are counted by hand at the precinct by citizen election judges of at least both major parties under public scrutiny by poll watchers.
| EJF Home | Where To Find Help | Join the EJF | Comments? | Get EJF newsletter |
| Vote Fraud and Election Issues Book | Table of Contents | Site Map | Index |
| Chapter 7 — Building Better Ballot Boxes |
| Next — To Register Doubts, Press Here by Sam Lubell |