Recurring Themes in Electronic Voting System Failures

This is an analysis of documented failures of electronic voting and vote tabulating systems compiled by Ellen Theisen, published at http://www.votersunite.org/info/messupsbyvendor.asp

1. Board of Election dependence on vendors (whenever there's a problem, vendor technicians have to fix it and Board of Election folks are helpless).

2. Voters saw their votes transferred on the screen to a different candidate, but there was no remedy for the voter or the candidate. Questions about whether the will of the people was expressed by the election outcome cannot be resolved, permanently tainting the legitimacy of both the election and the elected government. "Plausible" becomes the standard for election outcomes. (Printing voter-verified paper ballots ("VVPB") and performing a complete count of the VVPB, followed by an investigation and audit to reconcile discrepancies between computer and paper tallies would reveal vote-switching errors and allow for their correction.)

3. The number of lost votes was greater than the margin of victory.

4. When results are suspicious, there is no remedy. Most suspicious cases involve large margins of victory, so a recount would not be triggered by the "close election" standard.

5. Mysterious recurring unfixable errors.

6. Shoddy quality of merchandise, which shows failure to test the machines properly or adequately.

7. Failure to deliver on promises of accessibility.

8. Pre-programmed "back doors" to enable future fraud.

9. Gullible, passive, dependent Boards of Election don't throw the vendor out and demand their money back when the product doesn't work.

10. Vendors show contempt for our country, our elections, and our elections professionals:

--take advantage of Board of Elections ignorance, dependence, trust, and good will,

--minimize and excuse their own errors,

--blame system failures on Board of Election people, poll workers, and voters.

Broader Overview

1. Election professionals are inexperienced in dealing with secure computer systems. Computerized election systems are not subject to professional Information Technology ("IT") standards because elections professionals have never previously participated in the IT industry.

2. Election professionals see things that make them uncomfortable but they don't speak up because they don't know enough about computers and don't want to embarrass themselves.

3. With paper systems, everyone knows how to protect and falsify the ballots and tallies, so Board of Elections and crooks are on the same (level) playing field. Election integrity can be achieved with appropriate multipartisan public oversight of all election procedures.

With computers, technicians know how to falsify the ballots and tallies, but election professionals don't know how to protect them - they are out of their league, and cannot exercise oversight.

4. The "close election" legal basis for obtaining a 100% recount has come from the world of paper ballots and mechanical lever voting machines, where it takes many people in many polling places to falsify individual ballots, or lose ballot boxes on the way to the central location, etc. With computers, one person in a remote location can connect to the voting machines or central tabulator and falsify all precincts and the entire election in less than a minute. (That is why Georgia 2002 is a likely example of fraud.) With computerized elections, a "close election" is not a useful concept since the exact percentage of the outcome can be predetermined. Also, the degree of difference from pre-election or exit polls are not useful triggers, since 2004 saw vast manipulation of polling data (for example, see "Gallup defends results against MoveOn.org attack," USA Today, 9/29/04).

5. The legal basis for recounts (when they are done, how many precincts are counted) should reflect the vulnerabilities of the election technology. All computerized vote-recording and vote-counting is highly vulnerable, and thus 100% needs to be audited with 100% accuracy required. This means counting all voter-verified paper ballots AND investigation and audit to reconcile computer and paper tally discrepancies This process would require knowledge of the election software, and legal mandate for vendors to provide it via training (because vendors claim the software is "trade secret").

Intermittent errors

1. Computer systems of this size (50,000 to 1,000,000 lines of programming) typically produce intermittent and inconsistent errors. That is one reason corporations perform complete audits. In elections this means that one machine may work correctly, and the next one won't. A spot check of 2-3% will not find all errors, and may not find any of the errors that have occurred. Spot checks of the voting machines won't find the errors in the central tabulator.

2. If spot checks could find all the errors, no company would audit.

Conclusion

The use of computers in elections (1) means privatization because no Board of Elections is computer-savvy, and (2) removes election procedures from multipartisan public oversight and thus invites undetectable errors and fraud.

Computerized elections can be made observable by requiring electronic voting systems to produce VVPB, and requiring 100% audits with 100% accuracy. This doubles our work, however: each voter must enter votes on the computer and then examine the VVPB; election staffs need training, time, and funding to count the VVPB and investigate, audit, and reconcile discrepancies.

Hand-marked hand-counted paper ballots are intuitively easy to mark and count, and have the lowest rate of errors (CalTech/MIT Study). America should require paper ballots marked by hand (or accessible ballot-marking or -printing machines for voters with disabilities and non-English languages), and ban the use of computers in elections except for accessible devices, and optical scanners each of which must be fully audited after each use.