Statement before the Government Operations Committee
And the Committee on Mental Health, Mental Retardation, Alcoholism, Drug Abuse & Disability Services
Urging detailed and continuous oversight of
the Board of Elections process of selecting new voting equipment
Thank you for the opportunity to appear before you today, and for conducting this hearing.
Over the weekend I received a fax of a letter that was sent by John Ravitz to Marjorie Gersten on June 21, 2006 (a copy is attached). The letter lists ways in which DREs and optical scanner voting systems are similar. It does not mention the ways in which they are different.
Our Board of Elections in the City of New York seems to want to treat these two kinds of voting systems as similar, if not equivalent. We have seen this also in their testimony before the Governmental Operations committee in past hearings, most notably on April 24, 2006.
Yet DRE voting systems are very different from paper ballot/optical scanner systems.
The continued efforts of the Board to ignore the differences does not inspire confidence in the informed public. In fact it creates an impression that there may be something to hide. For example, some people now speculate that the Board has not published cost estimates for the two types of systems so as to not call attention to the huge cost differences.
I believe there may be a problem here, and it signals the need for detailed and continuous oversight by the City Council Governmental Operations Committee.
It also signals the need for the City Council as a whole to deal with Resolutions 131 and 228, and to pass them.
I would like to conclude with a historical comment. Before our present lever voting machines were purchased, they were tested in public. It is especially critical with computer equipment that it be fully tested in public before it is purchased, via the methods listed in Resolution 228.
We know that federal and state certification testing is only "function" testing. That means, it only confirms that required functions are present. It does not test the system as a whole, nor guarantee that it contains no malicious back doors.
In the Information Technology industry, 72% of projects fail. Public testing before purchase is the only way to pressure vendors to produce and deliver voting systems that might work. The Board’s resistance to Resolution 228 signals, again, a need for the Governmental Operations Committee to maintain detailed and continuous oversight of the process of evaluating new voting systems, and to support passage of Resolution 228 by entire City Council
a. Voter’s intent. Paper ballots marked directly by voters (by hand or accessible ballot-marking-device) are very likely to contain the voters’ intent. With DREs, the invisible electronic ballot can differ from what is on the screen and paper printout.
b. Can/will voters verify DRE paper printouts? Some voters may be unable to accurately verify their paper printout -- estimates of how many try is “few” to 90%, depending on training.
c. Will the votes on paper be counted? All paper ballots will be scanned, and 3% of them will be manually recounted. Only 3% of DREs will be randomly selected, and the votes on their paper printouts will be manually counted. Thus for 97% of DREs, the only votes to be tallied will be the invisible electronic ones. (If 3% is “statistically significant” what is 97%?)
d. How much equipment is needed? One optical scanner handles 2000-5000 voters per election day. At the March 7, 2006 GovOps hearing John Ravitz said the State Board might set a limit of 200 voters per DRE.
e. Voter and poll worker training is minimal for paper ballots and optical scanners. Accessible ballot marking devices and DREs may require the same amount of training.
f. Security. Paper ballot security is understandable to all, technically easy to maintain, and requires only a political will. Computer security is difficult and few understand it.
g. Recounts. Paper ballots are easier to handle/read than voter-verified paper strips.
h. Wait time. Many voters can mark paper ballots at the same time without long lines at the scanner (scanning takes 0-2 seconds). Only one voter at a time can vote on a DRE.
i. Equipment failure. If a scanner breaks down, voters can continue to vote -- not with a DRE.
j. Participation. Simple, understandable, observable technology enables people to participate. Computers shut people out and force them to “trust” experts, thus undermining democracy.
72% of software projects in a typical year, 2000, were complete or partial failures, including 23% that were completely abandoned after huge expenditures (and waste) of time and money. Regarding partial failures, if a computer system “partially” doesn’t work, that means it doesn’t work. http://www.pbs.org/cringely/pulpit/pulpit20031204.html "Why the Current Touch Screen Voting Fiasco Was Pretty Much Inevitable" by Robert X. Cringely, December 4, 2003.
What does testing mean? If a computer works during testing, that means it might work during actual use. If a computer does not work during testing, that means that it probably won't work during actual use. But no testing, especially function testing as contemplated by our New York State Voting Systems Standards, can tell you whether or not a computer systems will actually work, nor whether it is secure. Critique of NYS VSS is at