Government Reform Coordinator of the
November 21, 2006
Thank you. My name is Neal Rosenstein. I am the Government Reform Coordinator for NYPIRG, the New York Public Interest Research Group. We’ve monitored election administration issues across the city and state for more than 30 years and have closely followed implementation of the Help America Vote Act (HAVA) and it’s mandates for change, including for new voting systems.
I’d not only like to thank the Board for the opportunity to testify today, but also to thank the Board for having this hearing in the first place. Such events are necessary to allow the public, along with advocates, the opportunity to have their voices heard before this important decision is made. We urge the Board to conduct additional hearings before and after the certification process is completed, and to provide complete access to the Board’s own usability studies, fiscal analyses of the costs of each system, scorecards and rating documents to ensure public confidence that the best system will be chosen. An open and transparent process is vital to ensure maximum public confidence that the best system will be chosen and not the system with the best lobbyist.
NYPIRG has long advocated for the modernization of our lever voting machines and have been following the effort to replace the city’s antiquated machines since the late 80’s. We recognize that many have a sentimental attachment to their physical presence and the ability to hear that satisfying clunk when pulling that big red lever to cast their votes. But there’s no denying their lack of accessibility features, undervote protections, ongoing breakdowns and limits to accommodating the multiple languages required in our increasingly diverse city. It’s time for improvement and change.
Optical Scan is the Clear Choice for New York
The best choice for a new voting system for New York City is the introduction of an optical scan voting system that gives each voter the choice of utilizing a dedicated ballot marker at their Election District (ED) table or filling out their ballot by hand.
There are a number of different factors that will determine the success of a voting system. These include:
ü Ensuring that votes are cast. Any system must have strong protections against undervotes (when a voter fails to vote for a candidate in a particular race), pro-actively encourage voters to vote on all races including those towards the bottom of the ballot, and eliminate the possibility for overvotes (votes cast for more than one candidate, thus voiding both selections);
ü Providing full access to voters with disabilities;
ü Being capable of clearly presenting multiple languages as required by the Voting Rights Act (VRA);
ü Causing minimal disruptions to voters due to system malfunctions/breakdowns;
ü Enjoying the confidence of the public that their votes are counted as cast and that the system is safe from tampering or shoddy programming;
ü Avoiding being too complex as to require the privatization of aspects of the election process and maintaining the running and counting of elections in public hands; and
ü Being convenient and easy to use by both poll workers and the voting public.
1) Ensuring that votes are not lost due to undervotes or overvotes. Undervotes and overvotes - also referred to collectively as “Residual Votes” - can occur for a number of reasons, including voter unfamiliarity with voting systems, a conscious decision not to cast a vote for any of the candidates running for a particular race and the type of voting technology used.
A recently completed study by the Brennan Center for Justice has looked at the issue of lost votes, with a particular focus on the types of “full-face” machines that would be purchased in New York. Optical Scan systems fared far better than their DRE counterparts. The Brennan Center Study found a significant advantage to using optical scan systems for reducing lost votes in the 2004 elections. Optical scan systems had 42% fewer residual votes in this comparison:
ü Precinct Based Optical Scan .7%
ü Full Face DREs 1.2%
Just today, the Brennan Center has released even more convincing numbers that show precinct based optical scan system faring far better than their "full-face" DRE counterparts in registering residual/lost votes on races further down the ballot. Optical scan systems fared better than the national average of a 9.3% residual vote rate, with an 8.8% rate. "Full-face" DRE touch screen machines however, registered a rate of 15.4%, far more than the national average. This alone should sound the death knell for DRE systems. (Lever machines came in at an abysmal 32.1%.)
Additional statistical evidence comes from the definitive Voting - What Is, What Could Be, a July 2001 Report of the Caltech-MIT Voting Technology Project. This study also found that optical scan systems had far lower rates of residual votes than DREs.
Undervotes: Ballot Markers and Scanners. At last Thursday’s voting system demonstration in Queens, I had the opportunity to try each of the machines present in an informal test of their undervote safeguards. It was clear from my experience that the presence of ballot markers at every poll site could help reduce the likelihood of both undervotes and questions of voter intent from voters using ballot markers.
For my test I purposely failed to completely fill in the ovals or connect the arrows on my optical scan ballot. Instead, I used check marks, partially filled in ovals, placed a dash or an X in the space provided or circled my candidate’s oval or “connect the arrow” space. These are all the types of markings that could be reasonably expected from voters. While the scanners recognized and counted many of these markings, they also failed to register a considerable number of my choices.
Unfortunately, both the Diebold and ES&S scanners did little more than alert me to the first race I undervoted upon, displaying a notice on a relatively user-unfriendly screen atop the scanner. This is not a very effective way to have a voter take back and re-correct multiple mistakes on a ballot…there would be no way of knowing whether another race was undervoted for example until the voter re-deposited their ballot and repeated the process. This increases the likelihood of a voter whose ballot contained more than one undervote of simply growing frustrated enough to allow their ballot to be fed into the scanner, undervotes and all rather than repeatedly coming back to the scanner. It also increases the likelihood of ballots with unclear markings being accepted.
As previously mentioned, objective studies have determined that optical scan systems have low undervote rates. But we believe a common sense method of further reducing ballots with undervotes, stray or undecipherable marks and clearly establishing voter intent is through the use of ballot markers. Since ballot markers fill in the ballot for voters, all the problems associated with voter error should be dramatically reduced. Additionally, since voters are notified of undervote problems on a ballot marker before the ballot is printed, this should increase participation on down ticket races and reduce voter frustration with having to re-do their ballot once rejected by a scanner.
2) Providing full access to voters with disabilities. DREs and ballot markers both present a wide range of disability features to voters and in order to be certified for use in New York and would have to be compliant with HAVA and stronger state laws for disability access. Audio head phones, tactile control panels, sip and puff technology are standard on all the systems being considered here in New York.
We realize that many disability groups tend to favor DREs. They contend that optical scan systems would require some voters with disabilities to vote differently than other voters or require assistance in the handling and insertion of their ballot into either a ballot marker or optical scanner. However, we believe having a ballot marker present at every ED table along with proper poll worker training would help ensure that voters with disabilities are not treated differently at the polls.
One of the core reasons that NYPIRG supports ballot markers at each and every ED is to ensure that voters with disabilities or who wish to use alternative language features are not segregated onto separate lines at poll sites where an accessible system is located. Such a separation of voters raises questions of equal access and treatment at poll sites. A ballot marker at each ED would also reduce the time voters wanting or requiring their use would face on election day if forced to vote on a single ballot marker shared by three or four EDs.
To decrease the chances of voters being treated differently at poll sites, we urge the Board to develop and adopt new election day procedures. For example, there’s no reason that a poll worker shouldn’t be required to insert every voter's ballot into a ballot marker, and not just offer the service to those with disabilities. Similarly poll workers should be required to ask every voter using a ballot marker whether they wish assistance with transporting and inserting their ballot into the scanner. Protocols such as these will minimize the differences between the treatment of voters at the polls and could also help reduce delays or problems with the feeding of ballots.
3) Being capable of presenting multiple languages as required by the Voting Rights Act. Both DREs and ballot markers for optical scan systems offer many of the same language capabilities. An optical scan system that does not use ballot markers for language mandates under the Voting Rights Act would instead print the paper optical scan ballot (and instructions) in the languages required under law. It’s important to note that translated paper ballots would be required at many poll sites regardless of whether a DRE or optical scan system was chosen. For example, when voters are forced to use paper Affidavit Ballots at the polls because their name is not in the poll books or when paper Emergency Ballots are used when the voting “machine” breaks down. Paper Absentee Ballots are also currently required to be translated in the city under the VRA. NYPIRG believes that the presence of ballot markers at each ED will offer an alternative to translated hand completed ballots that offer benefits such as clear instructions in the voters native language, something not always available at poll sites. Close attention must be given to ensure that those pre-programmed instructions are accurate and we urge the Board to have language rights group given full access to the systems to test just that.
4) Causing minimal disruptions to voters due to system malfunctions/breakdowns. Optical scan systems enjoy the advantage that poll workers are already familiar with - and many voters will choose to use - a paper ballot filled out by hand. The increased familiarity, will likely result in less confusion and less disruption if a ballot marker were to break down as compared to a DRE breakdown. Additionally, many proponents of optical scan systems hold that since DREs are more complex than optical scan systems, and that they are subject to far higher breakdown rates. Perhaps because of the relative newness of these technologies, we have seen no studies comparing DRE and optical scan breakdown rates. However, there have been a plethora of problems reported with DREs across the country. The logic that the simpler system is less likely to break down makes sense particularly in New York, where poll worker training has been lacking.
Voter confidence in voting system is paramount, without it cynicism grows and participation suffers. Our position has been that if any DRE breaks down or malfunctions, the results are suspect and the machine should be removed from voting for the remainder of the day. If no replacement machine is available, voters would change to using paper ballots. Even if this standard was not incorporated by local Boards of Elections, when a DRE breaks down, voters would be forced to either wait for the machines to be repaired or replaced, return later in the day in the hope the machine was functioning, or switch and vote on a paper ballot. That’s the process right now, voters fill in an optical scan ballot when the lever machines are out of use. Even if DREs are chosen, poll workers and voters will have to be trained and proficient in paper ballot systems.
5) Enjoying the confidence of the public that their votes are counted as cast and that the system is safe from tampering or shoddy programming. No matter what safeguards are introduced, it is unlikely that DREs will satisfy the concerns of security experts, conspiracy theorists and the general public. Recent news reports for example, have demonstrated severe security flaws in at least one prominent DRE vendor, Diebold Inc. It is not difficult to imagine increasing and ongoing public suspicion and cynicism if DREs replace our current lever machines. And all the vendors marketing DREs in New York have done little to bolster public confidence, refusing to make their source code public so it could be scrutinized for security weaknesses.
While the mandate for a VVPAT reduces these concerns, it is not foolproof, especially given the shortcomings of federal and state certification requirements. NYPIRG also believes the audit provisions of state law are inadequate and should have allowed for targeted audits by candidates suspicious of the results in particular EDs. Although many of the same concerns arise for ballot markers and scanners, in an optical scan system the paper ballots are the official vote record and can be recounted.
While intentional tampering with electronic vote totals is a concern, many consider shoddy programming a more likely scenario that could lead to improperly functioning machines on Election Day. Poorly funded Boards of Election may not possess the technical know-how to ensure that computerized DREs are displaying and tabulating votes correctly. Such problems will only increase voter cynicism and undermine our elections.
6) Avoiding being too complex as to require the privatization of aspects of the election process and maintaining the running and counting of elections in public hands. Elections are a fundamental foundation of our democracy. The preparation of ballots, casting of votes and tabulation of results must remain an open and transparent process free from the possibility of influence from private, non-public sources lest public confidence in the integrity of the vote be questioned. Highly complex DRE machines often require ongoing service and maintenance from vendors and this is one of the lucrative incentives for companies to push the technology. The patronage staffed and poorly funded Boards of Elections across the state are ill prepared to take on this burden meaning an ongoing role for private vendors in the running of our elections, we find that unacceptable. It should be noted that optical scan Absentee and Affidavit Ballots are already used by the New York City Board of Elections and are scanned centrally at Board offices. It’s a technology ready to be incorporated for all voters across the city.
7) Being convenient and easy to use by both poll workers and the voting public. Both DREs and ballot markers offer a tantalizing array of user-friendly interfaces to assist voters in casting their ballots. However, some voters will undoubtedly be unfamiliar and not feel comfortable using an electronic interface. For these voters a hand filled paper ballot provides a simple and reassuring method of casting their votes. Optical scan systems will also have the benefit of potentially eliminating long lines on Election Day. If voters face a line for use of a ballot marker for example, they can elect to fill out their ballot by-hand at a low cost privacy booth. The number of voters who can be filling out ballots simultaneously is primarily limited only by the space available for low cost privacy booths.
The successful utilization of any voting system is largely dependent on the quality of the poll workers setting up and running the polls on Election Day. Both DREs and ballot markers will necessitate an extraordinary effort at educating poll workers on their use. Since poll workers will already need to be trained on procedures for filling out paper Affidavit and Emergency Ballots, the use of hand filled out ballots envisioned under optical scan systems will not represent an additional burden. It should be noted that nationally, there has been increasing unease from election officials over the use of DRE systems.
Finally, because of the likelihood that up to 2 DREs might be needed to replace a single lever machine, our crowded poll sites will be even more chaotic. Any move to a new technology should factor in such considerations into the decision making process.
Fiscal Impact. It's become increasingly clear that the State will not have the money to pay for a full replacement of voting machines with DREs. The full-face machines currently being marketed in New York State may cost up to $10,000 each. With about 20,000 current lever machines statewide, that would mean $200 million for a 1:1 replacement ratio. However, it's unlikely that the city or other counties would be able to replace their lever machines at a 1:1 ratio since it will probably take more than three minutes for voters to use a DRE. At three minutes a voter, only 20 voters an hour will be able to use a DRE. That's 300 voters evenly spread out over a 15-hour, 6am to 9pm election day. At four minutes, only 225 voters and at 5 minutes only 180 voters could be accommodated. With EDs of 800 voters and turnouts of well more than 50% in many EDs during presidential election years, it’s clear that the city would need significantly more DREs than lever machines.
As a result of the need for more time for voters to use DREs the replacement ratio to lever machines is more likely to be at least two DREs per ED. So, the likely cost of replacing lever machines with DREs would total $400 million. As previously mentioned,. New York has about $140 million of HAVA funds in the bank for the replacement of our lever machines. An additional $50 million has been asked to be returned because it was earmarked as "early lever machine buyout money." It's unclear whether NY will be able to hold on to this money. However, even the full $190 million will be insufficient if DREs are chosen. Currently, New York is required to increase federal HAVA money with only a 5% match.
The optical scan system NYPIRG prefers will be less expensive in acquisition costs than DREs. We support having one ballot marker per ED. These run about $6,000 along with a couple of privacy booths. Optical scan systems also require an actual scanner at the site which also cost about $6,000. However there are two cost mitigating factors with optical scan systems: 1) Only one ballot marker per ED is needed since many voters will choose to vote by hand or if there are lines for a ballot markers, decide to switch to completing ballots by hand; and 2) While we call for one ballot marker per ED, only one scanner per every 3 or 4 EDs is required.
At a one ED poll site, the cost for our ballot marker per ED proposal would be approximately $12,000, far less than the $20,000 dollars needed for 2 DREs (as previously noted more than one DRE will likely be needed per ED.) At a poll site with 3 EDs, the cost would be $18,000 for 3 ballot markers and $6,000 for one scanner to service them for a total of $24,000. That’s 20% less than the $30,000 that three DREs would cost. And if six DREs were needed, the cost would be $60,000 or two and a half times the cost.
One of the drawbacks to optical scan systems is the cost associated with the storage of paper ballots. We understand the hesitation of many election commissioners at the thought of having to supply and then store large quantities of paper ballots after Election Day. But after the initial set-up of a record retention system, we trust that the Board will be able to aptly manage it and that it would not prove burdensome.
In conclusion, NYPIRG believes the selection of an optical scan voting systems that gives each voter the choice of filling out their ballot by hand or utilizing a ballot marker at their own Election District table is the best option for the voters of New York City and State. Undervote protection, accessibility, language access, security, ease of use and a higher level of public confidence are only some of the advantages offered by ballot markers and optical scan. We hope the Board agrees.
NEW YORK STATE MUST RELY ON OPTICAL SCANNING TECHNOLOGY FOR ITS VOTING MACHINES
New York is Replacing its Voting Machines
Throughout the state, New Yorkers vote on lever machines that are more than 40 years old and no longer manufactured. They cause unacceptable levels of uncounted votes, long lines at poll sites and are not accessible to many voters with disabilities. Each county in New York will soon be making a vital decision on how to replace them.
ü Enjoying greater public confidence. Optical scan systems are inherently safer from tampering or shoddy programming. It is unlikely that Direct Recording Electronic computerized machines (DREs) will ever enjoy the same level of public confidence as optical scan systems. Public cynicism about the integrity of elections undermines participation.
ü Reducing the number of lost votes where voters leave without casting a ballot. DREs register undervotes at a rate that’s 40% higher.
ü Providing greater access for voters. Ballot markers at every voter’s table would offer dramatically increased access for voters with disabilities and provide ballots in alternative languages such as those required by the Voting Rights Act. Care must be given to ensure all voters have maximum privacy and independence when casting their ballots.
ü Elections won’t be privatized. The preparation of ballots, casting of votes and tabulation of results must remain an open and transparent process lest public confidence in the integrity of the vote be questioned. DREs use of secret computer code and the necessity for private vendors to maintain and update complex machines means private vendors will be running and counting elections.
ü Optical scan systems are easier to use by voters and easier to oversee by poll workers. Both DREs and ballot markers offer a tantalizing array of user-friendly interfaces to assist voters in casting their ballots. However, some voters will undoubtedly be unfamiliar and not feel comfortable using an electronic interface. For these voters a hand filled paper ballot provides a simple and reassuring method of casting their votes.
 Nordon, Creelan, Kimball, Quesenbery, Brennan Center for Justice, August, 2006, The Machinery of Democracy – Usability of Voting Systems.
 New York Times, September 24, 2006, The Big Gamble on Electronic Voting
 New York Times, September 25, 2006, Officials Wary of Electronic Voting Machines
 Nordon, Creelan, Kimball, Quesenbery, Brennan Center for Justice, August, 2006, The Machinery of Democracy – Usability of Voting Systems.