Reliability of US Voting Systems:
An assessment in the light of recent changes
Prepared Remarks for the
Congressional Black Caucus
Hearing on election preparedness
October 7, 2004
[Notes in square brackets added October 8]
I am an Associate Professor of Computer Science at the University of Iowa, and from 1994 to 2004, I served on the Iowa Board of Examiners for Voting Machines and Electronic Voting Systems. I have consulted on election related issues with the Illinois chapter of the ACLU and with Miami Dade County, and I am a member of the National Committee for Voting Integrity.
I would like to thank the Congressional Black Caucus for their invitation to address this question:
What is your best assessment of the reliability of voting systems nationwide given the changes that have been made to voting equipment as a result of the "Help America Vote Act"?
With 50 states, 3,142 counties, and an election turnout of 110 million (in 2000), it should be no surprise when things go wrong. It is a fair guess that there are about 1 million pollworkers nationwide [EAC estimates put this closer to 2 million]. In any human endeavor involving this many people, it is extremely likely that not all of them are honest, and even more likely that many of them will make honest mistakes.
When the margins in an election are wide these problems are inconsequential, but when the margins are narrow, as was the case in Florida in 2000, these problems dominate the news. The upcoming election also promises to be close, so to find similar problems coming to light again would not be a surprise.
Public attention also plays a role. Both error and fraud thrive where there is no close scrutiny. If the public are alert and watching closely, as they are now, alerted by the events of 2000, many normally careless election workers will take extra care, and many who might have been tempted to cheat will not do so.
Because of this increased scrutiny, I believe that the upcoming election may well be among the most accurate elections in recent history. On the other hand, we may never know this because increased scrutiny in a close election will certainly uncover problems that would normally go unnoticed.
In moving to touch-screen voting systems, we have traded a small increase in ease of use, for the voter, for a horrendous increase in difficulty for the pollworkers. Paul Craft of the Florida Division of Elections said this to the Technical Guidelines Development Committee of the Election Assistance Commission in their hearings on September 20, 2004.
There is indeed good reason to phase out old mechanical lever voting machines and to abandon the use of Votomatic punched-card ballots. I told this to the United States Civil Rights Commission on January 11, 2001, and I stand by my recommendations of that date.
Unfortunately, with a few exceptions, we have put our focus on replacing voting machines when most of the problems from the election of 2000 had their roots in election administration. While punched-card technology is bad, the butterfly ballot was the product of a record number of minor-party presidential candidates, a poor choice by a county election administrator, and weak state oversight.
Most of the new voting systems that have been purchased in the past 4 years were built and tested to the Federal Election Commission's 1990 standards, although parts of some of these systems are now certified to the newer 2002 standards. I have written at length about the shortcomings of these standards, for example, in my testimony before the House Science Committee on May 22, 2001, and before the Federal Election Commission, on April 17, 2002.
I sent the House Science Committee detailed comments on the Help America Vote Act (HAVA) on November 26, 2001, and I stand by those comments. In the long run, HAVA will have positive consequences, if it is ever properly funded. The most important contributions of this act will be in the development of new voting system standards and in the documentation of the best practices in election administration.
I feel that the rush to fund the purchase of large numbers of new voting systems was a mistake. This mistake was compounded by the delay of over a year between the passage of the act and the appointment of members to the Election Assistance Commission (EAC), and the mistake has been further compounded by the failure of the Congress to allocate the promised funding for the standards and best-practices activities of the EAC.
Proponents of HAVA, at the time it was passed, wanted the new voting system standards promulgated by the EAC to be in place in time to have an effect on the voting systems purchased using HAVA funds. This will not happen. It takes several years for the industry to respond to new voting system standards, as is well illustrated by the fact that most of the voting systems purchased since the issue of the new 2002 voting system standards were certified to the older 1990 standards. If the EAC were, by some miracle, to issue new standards today, they would have little effect by the HAVA deadline of 2006.
For the past year, I have emphasized that no new purchase of voting systems is likely to benefit this November's election. A responsible local election official intent on buying new machines should begin the process about 18 months before the general election! For the past year, what we have needed to pursue is improved election practices. There is still time to make some small changes:
Consider the problem of processing absentee ballots. In many counties, 5 to 10 percent of all votes cast on absentee ballots go uncounted. Ballots marked with the wrong type of pencil or pen, circles around the candidate's name instead of filling in the bubble, and similar problems cost enough votes to swing many elections.
In some counties, however, these are not a big problem! In Miami-Dade County, all overvoted absentee ballots are routinely sorted out for inspection by the canvassing board on the first count, so that crossed out errors, smudged erasures and flyspecks will not be discounted as overvotes.
In Arlington Virginia, over 5 percent of votes cast on absentee ballots went uncounted until the county reset their ballot scanners so they could follow the procedure described above; now, the numbers are closer to 1 percent. This is according to Fred Berghoefer, Secretary of the Arlington Virginia election office, as reported at the September 20 hearing of the Technical Guidelines Development Committee of the Election Assistance Commission. This is a change counties can still make in time for the November election, where state law permits!
Consider the problem of informing voters of their right to a provisional ballot. Some civil rights organizations have insisted that, if a voter is not listed in the pollbook for a precinct, that voter should immediately be informed of their right to a provisional ballot.
Unfortunately, there are many states where a provisional ballot can only be counted if it is cast in the right precinct. In those states, giving out provisional ballot will disenfranchise voters who are at the wrong precinct. At their worst, provisional ballots therefore become a way to brush off troublesome voters by letting them think they have voted.
When election workers find voters who are not in the pollbook, they should begin by trying to help those voters find the right precinct. Only after it is determined that the voter is at the right precinct should a provisional ballot be offered. There is still time for many counties to amend their pollworker training materials to encourage this, and there is still time for civil rights organizations to inform their observers that they should insist on these priorities.
Despite laws that require election officials to carry out their activities in public, there are stories from around the country of election observers being turned away from polling places or county vote tabulating centers. We need to insist that observers at the polling places be permitted to watch not only the voting itself, but also the pre-election setup of the voting equipment and the crucial activities involved in closing the polls and reporting the results.
Similarly, at county tabulating centers, we must insist on the rights of election observers to watch all crucial activities. Not only must observers be able to watch, but they must be able to understand what they are seeing. Watching someone count X-marks on paper ballots is easy enough, but as the work of vote counting has been shifted into computers, the ability of an observer to distinguish between the legitimate an the illegitimate has declined.
We can and should ask that at least some of the observers present at the tabulating center be able to watch what is on the computer screen when an election official or technician enters commands into that computer. In addition, we must ask that the observer be in a position to understand what is being done, either because an explanation is offered for each such command, or because the observer has been given access to the relevant documentation far enough in advance that they had a reasonable opportunity to learn what the commands mean.
I believe that many election officials will resist extending these rights to observers, but I am not convinced that overcoming this resistance will require legislation. Current law regarding the rights of election observers may be sufficient, but it may take good lawyers and fast work with injunctions to assure that election observers can do their jobs.
Unfortunately, it is still the case that where you vote has too much impact on how likely your vote is to be counted. Despite this, my first recommendation to anyone who is concerned about this is, go vote. If you do not vote, you can be sure your vote will not be counted.
There are a few things we can still do to improve election conduct during the upcoming general election, but one of them is central. Be suspicious! Watch how the polls are run! Watch how the votes are counted! The presence of diligent election observers really does reduce the likelihood of both error and fraud.
National Committee for Voting Integrity
Jones Testimony before Civil Rights Commission, January 11, 2001
Jones Testimony before the House Science Committee, May 22, 2001
Jones Testimony before the Federal Election Commission, April 17, 2003
Jones Comments on the Help America Vote Act, November 26, 2001