Statement before the Government Operations Committee
New York City Council
Good morning. My name is Teresa Hommel. Thank you for conducting this hearing so that the public has a chance to express our opinions on HAVA compliance, especially on our new voting technology.
I have been a full-time activist against electronic voting for more than two years. I am Chairwoman of the Task Force on Election Integrity of Community Church of New York.
My professional background is that I have worked with computers for 38 years, since 1967. I have worked for many of the Fortune 500 companies, the Department of Defense, and various state and local governments.
I have already briefed several Council Members on the problems of electronic voting and the reasons why paper ballots and optical scanners are the more responsible choice for our city.
This morning I am presenting a smoking gun -- a new report that reveals the disenfranchisement of minority voters in New Mexico in the presidential election last November. This report is especially significant to New York City because it concerns the Sequoia Advantage, which I believe is the machine of choice that our Board of Elections would like to get to replace our lever machines.
I have handed in copies of the report for the Council Members here, stapled to my testimony.. The report is called "New Mexico Canvass Data Shows Higher Undervote Rates in Minority Precincts where Pushbutton DREs Were Used." It was prepared by Ellen Theisen of VotersUnite.org and is available on the internet at www.votersunite.org/info/NM_UVbyMachineandEthnicity.pdf
Ms Theisen cross-referenced three kinds of data:
--ethnicity of voters in each jurisdiction, obtained from the 2000 Census
--election results from November, 2004, obtained from the New Mexico Secretary of State
--voting equipment type, obtained from the New Mexico Secretary of State and confirmed by telephone with each County Clerk's office in the state of New Mexico.
She ended up with 4188 separate sets of data, and then selected voting precincts in which 75% or more of the voting population was of a specific ethnicity. The ethnic groups were Hispanic, Native American, and Anglo. The selected data represented 246,290 ballots cast. She looked at undervote rates in the presidential race, meaning ballots that registered no vote for president.
Her analysis discovered:
-- Paper ballots tabulated by optical scan systems had virtually identical undervote rates for all ethnicities.
-- Lumping together all voting technologies, Hispanic precincts averaged more than 3% higher undervote rates than Anglo precincts. Native American precincts averaged more than 5.5% higher undervote rates than Anglo precincts.
-- The statewide disparity in undervote rates for different ethnicities occurred entirely on Danaher Shouptronic and Sequoia Advantage pushbutton paperless electronic voting machines.
-- Down-ticket undervote rates were consistently higher in minority precincts, suggesting that entire ballots may have been uncounted in minority precincts, particularly on Danaher Shouptronic and Sequoia Advantage machines.
Independent confirmation of her results is currently in progress. Anyone who is willing to do the same massive amount of work can duplicate her results because the data she used is all publicly available, at least at this time.
I would like to make some comments.
1. Regardless of what advantages computerized voting supposedly provides, computers open the door to unobservable tampering/errors and that is why we should not use computers for voting.
2. As this new study shows, computers also open the door to FOCUSED or TARGETTED tampering/errors that can disenfranchise specific voter populations. Now, people always have anecdotes about the fact that the lever machines in certain neighborhoods tend to be broken – but when you know the machine is broken, voters can use emergency ballots and vote on paper. Here, the computer can appear to work just fine, while in fact it is erasing or switching a given percentage of votes.
3. We have known for a long time that electronic voting equipment uses at least partially separate programming for ballots displayed in different languages. We know that separate programming is used to handle ballots cast by each disability group (such as blind/visually-impaired voters who use the audio, versus voters with manual dexterity disabilities who use the hand-held or foot-pedal devices, versus paralyzed voters who use the sip-puff device). Separate hardware and programming provide one kind of opportunity for focused tampering/errors that can disenfranchise specific populations of voters.
4. This smoking gun report is news, but tampering/errors by language group is not. It was reported more than a year ago by Kim Zetter of Wired.com (Aug. 12, 2004, "Wrong Time for an E-Vote Glitch," www.wired.com/news/evote/0,2645,64569,00.html).
She reported that "During the demonstration of the Sequoia machine last week, the machine worked fine when the company tested votes using an English-language ballot. But when the testers switched to a Spanish-language ballot, the paper trail showed no votes cast for two propositions."
5. Failures of electronic voting machines are typically blamed on voters or poll workers who are too old/dumb/poorly trained/inexperienced, etc. However, the Federal Election Commission 2002 voting system certification standards allow a maximum error rate of 1 in 500,000 ballot positions, thus giving vendors and Boards of Elections enormous motivation to blame errors on people so that their machines are not decertified. This study warns us not to accept the easy shrug of the shoulder and the contemptuous “Oh, those people, they just don’t know how to vote.”
6. Any tampering that can be done with lesser technologies (such as paper ballots) can be accomplished with electronic voting also.
The main difference is that electronic voting
· makes tampering faster and easier
· does not require anyone to be physically present before, during, or after the election
· requires only one or a tiny number of people, and less need to organize
· leaves no trace of evidence
· provides a nearly infinite number of different ways to achieve a given result so that tamperers never have to use the same approach more than once. You can have randomized selection of machines in which to switch votes; randomized percentages of votes switched; randomized selection within targeted populations or jurisdictions; undervotes last year, switched votes next year, rotated votes in the election after that; pushbuttons last year, touch screens next year, etc. If you can imagine what you want, it can be accomplished with a few minutes of programming, and with a second of communications contact with each machine.
7. Nobody wants to ask--or answer--certain questions. I’m asking: is it ok to disenfranchise individual voters, or specific communities of voters, if the errors are innocent? Or is electronic voting just too prone to errors, whether they are innocent or malicious? Several speakers will mention VotersUnite.org's list of documented failures of electronic voting systems, which is over 120 pages, www.votersunite.org/info/messupsbyvendor.asp
8. Will we ever get to the point where we say, "Enough is enough!" and demand a reliable, observable, LESSER technology that ordinary people can manage and keep secure? Our new state law has banned lever machines as of 9/1/07. We should switch to paper ballots, optical scanners, and ballot marking devices for voters with disabilities and minority languages.