SPEAKERS CONTENTS INSERTS Tables
Page 1 TOP OF DOC
73–327PS
2001
IMPROVING VOTING TECHNOLOGIES:
THE ROLE OF STANDARDS
HEARING
BEFORE THE
COMMITTEE ON SCIENCE
HOUSE OF REPRESENTATIVES
ONE HUNDRED SEVENTH CONGRESS
FIRST SESSION
MAY 22, 2001
Serial No. 107–20
Printed for the use of the Committee on Science
Available via the World Wide Web: http://www.house.gov/science
COMMITTEE ON SCIENCE
Page 2 PREV PAGE TOP OF DOC
HON. SHERWOOD L. BOEHLERT, New York, Chairman
LAMAR S. SMITH, Texas
CONSTANCE A. MORELLA, Maryland
CHRISTOPHER SHAYS, Connecticut
CURT WELDON, Pennsylvania
DANA ROHRABACHER, California
JOE BARTON, Texas
KEN CALVERT, California
NICK SMITH, Michigan
ROSCOE G. BARTLETT, Maryland
VERNON J. EHLERS, Michigan
DAVE WELDON, Florida
GIL GUTKNECHT, Minnesota
CHRIS CANNON, Utah
GEORGE R. NETHERCUTT, JR., Washington
FRANK D. LUCAS, Oklahoma
GARY G. MILLER, California
JUDY BIGGERT, Illinois
JOHN ABNEY CULBERSON, Texas
W. TODD AKIN, Missouri
TIMOTHY V. JOHNSON, Illinois
MIKE PENCE, Indiana
FELIX J. GRUCCI, JR., New York
Page 3 PREV PAGE TOP OF DOC
MELISSA A. HART, Pennsylvania
RALPH M. HALL, Texas
BART GORDON, Tennessee
JERRY F. COSTELLO, Illinois
JAMES A. BARCIA, Michigan
EDDIE BERNICE JOHNSON, Texas
LYNN C. WOOLSEY, California
LYNN N. RIVERS, Michigan
ZOE LOFGREN, California
SHEILA JACKSON LEE, Texas
BOB ETHERIDGE, North Carolina
NICK LAMPSON, Texas
JOHN B. LARSON, Connecticut
MARK UDALL, Colorado
DAVID WU, Oregon
ANTHONY D. WEINER, New York
BRIAN BAIRD, Washington
JOSEPH M. HOEFFEL, Pennsylvania
JOE BACA, California
JIM MATHESON, Utah
STEVE ISRAEL, New York
DENNIS MOORE, Kansas
MICHAEL M. HONDA, California
Page 4 PREV PAGE TOP OF DOC
C O N T E N T S
May 22, 2001
Witness List
Hearing Charter
Opening Statement by Representative Sherwood L. Boehlert, Chairman, Committee on Science, U.S. House of Representatives
Opening Statement by Ralph M. Hall, Minority Ranking Member, Committee on Science, U.S. House of Representatives
Opening Statement by Representative James A. Barcia, Member, Committee on Science, U.S. House of Representatives
Panel
Dr. Stephen Ansolabehere, Professor of Political Science, Massachusetts Institute of Technology; Director, Caltech-MIT Voting Project
Dr. Rebecca Mercuri, Assistant Professor of Computer Science, Bryn Mawr College; President, Notable Software, Inc.
Mr. Roy G. Saltman, Consultant on Election Policy and Technology
Dr. Douglas W. Jones, Professor of Computer Science, University of Iowa; Member, Iowa Board of Examiners for Voting Machines and Electronic Voting Systems
Page 5 PREV PAGE TOP OF DOC
Discussion
Appendix 1: Opening Statements
Written Statement by Representative Sherwood L. Boehlert, Chairman, Committee on Science, U.S. House of Representatives
Written Statement by Representative Ralph M. Hall, Minority Ranking Member, Committee on Science, U.S. House of Representatives
Written Statement by Representative Constance Morella, Member, Committee on Science, U.S. House of Representatives
Written Statement by Representative Jerry F. Costello, Member, Committee on Science, U.S. House of Representatives
Written Statement by Representative Sheila Jackson Lee, Member, Committee on Science, U.S. House of Representatives
Written Statement by Representative Eddie Bernice Johnson, Member, Committee on Science, U.S. House of Representatives
Appendix 2: Written Testimony, Biographies, Financial Disclosures, and Answers to Post-Hearing Questions
Dr. Stephen Ansolabehere, Professor of Political Science, Massachusetts Institute of Technology; Director, Caltech-MIT Voting Project
Written Statement
Abstract
Page 6 PREV PAGE TOP OF DOC
Biography
Financial Disclosure
Dr. Rebecca Mercuri, Assistant Professor, Bryn Mawr College
Written Statement
Abstract
Rebecca Mercuri's Statement on Electronic Voting
Generic Security Assessment Questions
Questions for Voting System Vendors
Biography
Financial Disclosure
Mr. Roy G. Saltman, Consultant on Election Policy and Technology
Written Statement
Abstract
Biography
Financial Disclosure
Dr. Douglas W. Jones, Professor of Computer Science, University of Iowa; Member, Iowa Board of Examiners for Voting Machines and Electronic Voting Systems
Written Statement
Abstract
Biography
Financial Disclosure
Appendix 3: Additional Material for the Record
Submitted Testimony from the Federal Election Commission
Page 7 PREV PAGE TOP OF DOC
Letter from the Association for Computing Machinery
Types of Voting Technologies Used in the United States, 1998
Executive Summary of ''Residual Votes Attributable to Technology: An Assessment of the Reliability of Existing Voting Equipment,'' The Caltech/MIT Voting Project, March 30, 2001
''System Integrity Revisited,'' Communications of the ACM, January 2001
''Voting Automation (Early and Often?),'' Communications of the ACM
''Corrupted Polling,'' Communications of the ACM, November 1993
''Voting-Machine Risks,'' Communications of the ACM, November 1992
''Risks in Computerized Elections,'' Communications of the ACM, November 1990
''Finding Profit in Chad,'' U.S.News & World Report, March 6, 2001
''The Importance of Recounting Votes,'' November 13, 2000
'' 'Scalable' Ballot Fraud: Why One Tech Maven Fears Computer Voting,'' The Wall Street Journal, March 19, 2001
''Computer Scientists and Political Scientists Seek to Create a Fiasco-Free Election Day,'' The Chronicle of Higher Education, April 20, 2001
''Holes in Punch-Card System Noted Long Ago,'' USA Today, March 7, 2001
Page 8 PREV PAGE TOP OF DOC
IMPROVING VOTING TECHNOLOGIES: THE ROLE OF STANDARDS
TUESDAY, MAY 22, 2001
House of Representatives,
Committee on Science,
Washington, DC.
The committee met, pursuant to call, at 10:05 a.m., in Room 2318 of the Rayburn House Office Building, Hon. Sherwood L. Boehlert (chairman of the committee) presiding.
Committee on Science
U.S. House of Representatives
Washington, DC 20515
Hearing on
Improving Voting Technologies:
The Role of Standards
Page 9 PREV PAGE TOP OF DOC
Tuesday, May 22, 2001
Witness List
Dr. Stephen Ansolabehere
Professor of Political Science,
Massachusetts Institute of Technology;
Director, Caltech-MIT Voting Project
Mr. Roy Saltman
Consultant on
Election Policy and Technology
Dr. Rebecca Mercuri
Assistant Professor,
Bryn Mawr College
Dr. Doug Jones
Page 10 PREV PAGE TOP OF DOC
Professor of Computer Science,
University of Iowa;
Member, Iowa Board of Examiners for
Voting Machines and Electronic Voting Systems
Section 210 of the Congressional Accountability Act of 1995 applies the rights and protections covered under the Americans with Disabilities Act of 1990 to the United States Congress. Accordingly, the Committee on Science strives to accommodate/meet the needs of those requiring special assistance. If you need special accommodation, please contact the Committee on Science in advance of the scheduled event (three days requested) at (202) 225–6371 or FAX (202) 225–0891.
Should you need Committee materials in alternative formats, please contact the Committee as noted above.
HEARING CHARTER
COMMITTEE ON SCIENCE
U.S. HOUSE OF REPRESENTATIVES
Page 11 PREV PAGE TOP OF DOC
Improving Voting Technologies:
The Role of Standards
TUESDAY, MAY 22, 2001
10:00 A.M.–12:00 NOON
2318 RAYBURN HOUSE OFFICE BUILDING
I. Purpose
As a result of the 2000 presidential election, Congress has undertaken a review to implement significant election reforms. As Congress considers legislation to reform the voting process, a number of issues have emerged as part of the debate, such as whether changes are needed in the voting technologies used in the United States, and what should be the appropriate federal role.
As part of this congressional review, the House Science Committee is examining the role of standards in improving voting technologies. On May 22, 2001, the Committee will convene a hearing to ascertain the problems of our current election system and to explore potential solutions.
The issues to be addressed at the hearing include:
Page 12 PREV PAGE TOP OF DOC
(1) What problems have been identified in the various voting systems used throughout the United States?
(2) Which of these problems can be addressed by developing or improving standards for voting equipment? What kinds of standards need to be developed or improved, and why?
(3) What different types of research, testing, or data-collecting activities are necessary in order to develop effective voting standards?
(4) What are the major concerns, such as computer security, auditability, accountability, testing, certification, and accreditation, for new voting technologies?
II. Background
A. General Background:
Reports of problems in Florida and elsewhere in the nation during the 2000 election raised concerns about specific failures of voting technologies. One focus of current debate is whether more rigorous standards can provide useful guidance to elections officials. After election day, the media focused attention on specific problems with punch card voting. In the months since then, however, broader questions have arisen about error rates, costs, counting standards, and other issues with all types of voting technologies.
In the 1980s, the Federal Elections Commission (FEC) began developing voluntary standards for computer-based voting systems (see http://www.fec.gov/elections.html). Thirty-two states have now adopted all or parts of those standards, which were issued in 1990. However, the FEC standards have many critics, who consider them to be inadequate, suggesting that national standards must be expanded in scope to address factors such as ballot design, election management, and voter error. Solutions are likely to consider such diverse factors as cost, speed, accuracy, security, reduction in voter errors, and ease of use.
Page 13 PREV PAGE TOP OF DOC
B. Voting Technologies Used in the Last Election:
In the past election cycle, there were five different types of voting technologies in use around the country: Hand-counted ballots, mechanical lever machines, computer-tabulated punchcards, computer-tabulated optical scan ballots, and computer-based direct recording electronic (DRE) systems. Across the country, punchcard systems were the most common, used by about one-third of registered voters, while optical scan systems were used by about one-quarter.
The following is a description of each of the five types of voting technologies:
1. Paper Ballots. The oldest technology, paper ballots are still used in about 3% of precincts, mostly in rural areas. Paper ballots are counted manually. The percentage of voters using paper ballots has declined by half since 1992.
2. Lever Machines. First introduced in 1892, lever machines have no document ballot. Instead, a voter enters the voting booth and chooses candidates listed on a posted ballot by pulling a lever for each candidate choice. The votes are recorded by a counting mechanism in the back of the machine, eliminating the need to count ballots manually. Instead, poll workers read the numbers recorded by the counters. Since there is no document ballot, recounts and audits are limited to review of totals recorded by each machine. Write-in votes must be recorded on separate document ballots. About 22% of precincts currently use lever machines. That percentage has declined substantially since 1992 and is expected to continue to decrease because the machines are no longer manufactured, although parts are still available.
Page 14 PREV PAGE TOP OF DOC
3. Punchcards. The first technological approach utilizing computers to count votes was the punchcard system, introduced in 1964. In this system, considered among the most economical and efficient, especially for jurisdictions with large populations, the voter records choices by punching holes in appropriate locations on a paper computer card that is later fed into a computer reader to record the vote. The computer card serves as the document ballot on which the votes are recorded. As with other document ballots, punchcards can be manually recounted and audited. There are two basic types of punchcard systems:
VotoMatic type: A voter is given a ballot printed with numbered boxes, each box corresponding to a particular ballot choice printed in a booklet attached to the voting machine. The voter slips the card into the ''throat'' of the voting machine, where it rests on a set of rubber strips under the ballot book, and uses a simple stylus to punch out the chad for the box(es) corresponding to the candidate(s) chosen for each race or other item on the ballot. Turning a page in the booklet exposes another set of boxes on the card, corresponding to another set of ballot choices. This was the kind of system used in Palm Beach County, Florida. Write-in choices are not placed on the card itself but are written elsewhere, such as on the envelope in which the card is placed. About 33% of precincts use this type of system, the most widely used voting technology at present. The number of voters using the system has declined since 1992, and that decline is expected to continue.
DataVote type: A voter punches holes next to the names of candidates or other ballot choices that are printed on the cards themselves—there is no ballot book. The voter places the ballot card in a voting apparatus that has a stapler-like punching mechanism on a slide. Write-in votes can be placed directly on the card. About 4% of precincts use the Datavote system, and usage of this system has also declined.
Page 15 PREV PAGE TOP OF DOC
4. Optical Scan. This technology, which is also known as a ''marksense'' or ''bubble'' ballot system, has been used for decades in scoring standardized tests and first appeared for use in voting in the 1980's. In this system a voter, using a paper form and an appropriate writing instrument, darkens in a box or oval or completes an arrow corresponding to each candidate choice. A computerized device that senses and records the marks then scans the completed ballot. Write-in votes can be placed directly on the ballot. About 25% of precincts use marksense voting systems. The percentage of voters using this technology has almost doubled since 1992, and that increase is likely to continue.
5. Direct Recording Electronic (DRE) Voting. This technology, first introduced in the 1970's, is an electronic version of the lever voting machine, in which a voter's choice is recorded not on paper or a by a mechanical counter, but electronically by the computer. Depending on the equipment used, the ballot may be printed and posted on the voting machine, or it may be displayed on a computer screen. Voters make their choices by pushing a button, touching the screen or key pad, or using some other device. The voter submits those choices before leaving the booth, for example by pushing a ''vote'' button, and the votes are directly stored in a computer memory device such as a removable disk or nonvolatile memory circuit. If the voting equipment has a keyboard, write-in votes can be recorded electronically, otherwise, they must be recorded separately on a document. DRE systems have often been considered the most expensive (except perhaps for lever machines), but they are also arguably the most adaptable, with the greatest potential for speed. About 7% of precincts use DRE voting systems. Like the marksense systems, the percentage of voters using DRE has almost doubled since 1992 and is expected to continue to increase.
Page 16 PREV PAGE TOP OF DOC
The following table indicates the types of voting technologies and the incidence of their use in the U.S. in 1998.
73327a.eps
C. Internet Voting:
One form of electronic voting currently in development is Internet voting, in which voters make their choices online. Internet voting differs from DRE systems in several ways. First, it is often done using a personal computer rather than a custom-designed voting machine, although such machines can also be used. Second, results are not accumulated at the polling place but are sent to the tabulating computer when cast. Third, results (ballots or counts) are not sent over a direct modem connection or physically transported to the central tabulator, but are sent over the Internet. Those features make Internet voting a promising technology in some ways but pose special challenges for ensuring authentication, secrecy, and security in the voting process. The use of Internet voting is currently limited to demonstration projects. For example, for the November 2000 election, voters in several counties in California cast nonbinding votes online, from online voting machines placed in central locations. In the same election, 84 overseas military personnel cast their actual votes via the Internet through a small pilot project run by the Federal Voter Assistance Program (FVAP).
D. Current Standards for Voting:
In 1982 Congress directed the FEC to develop national standards for computer based voting systems that states might voluntarily adopt. In 1990, the National Association of State Election Directors (NASED) approved FEC's voluntary standards, which, at the time of the November 2000 elections, had been adopted in whole or part by 32 states, including Florida, although many states (including Florida) grandfathered in technologies introduced before the standards were developed. These standards were developed for both hardware and software and include functional and documentation requirements, performance characteristics, and testing procedures for punchcard, marksense, and DRE systems. The FEC plans to update these standards next year.
Page 17 PREV PAGE TOP OF DOC
Some have pointed to the occurrence of Florida's elections problems despite the state's adoption of FEC's standards as evidence that the standards are inadequate. They argue that such standards should be updated to include more robust standards for computer security, integrity, and accuracy of the election process. They also urge that the standards be expanded to include performance-based standards that address voting errors made in real voting situations that arise from such factors as voting machine design, ballot design, and election management (including maintenance). Such standards, for example, could require that a voting system prevent or reduce overvotes, or votes for more than one candidate.
Testing of voting equipment is performed at the national level by two independent labs overseen by the Elections Center, the professional association of election officials. These labs test voting machines, according to the vendor's specifications, to determine if they meet FEC hardware and software standards. Voting machines generally are not tested for their ability to meet performance-based standards under election-like conditions. In addition, while some states require that voting technologies meet requirements beyond those required by the FEC, few states have independent testing laboratories to certify voting equipment.
Additionally, the National Institute of Standards and Technology (NIST) has performed research on voting technologies for 30 years, including issuing reports in 1978 and 1988 detailing major problems with punchcard systems and other technologies. NIST's past voting expertise and ability to conduct standards research has led to proposals for NIST to create thresholds for accuracy, maintenance, and usability of voting systems. NIST has no ongoing role in setting the FEC standards.
Page 18 PREV PAGE TOP OF DOC
E. The Response of States, Localities, and Others:
The responses of state and local election officials since November have been mixed. While state and local groups have welcomed Federal financial assistance, they have warned Congress about mandating ''one-size-fits-all'' solutions. For example, the National Association of Secretaries of State adopted a February, 2001, resolution calling for updated, voluntary national standards and federal funding for voting system modernization, among other actions. The National Association of State Election Directors also adopted a resolution in February calling for increased federal funding to develop updated and expanded standards.
The National Association of Counties (NACO) and the National Association of County Recorders, Election Officials, and Clerks established a National Commission on Election Standards and Reform in November 2000. The Election Center, an association of election and voter registration officials, has established an Elections Reform Task Force to review concerns about election systems and recommend changes. Both groups are currently still deliberating.
The National Conference of State Legislatures has also established an Elections Reform Task Force to restore public confidence in state election systems, and is attempting to identify model practices and laws for states to consider. Reform legislation is pending in all 50 states, with more than 1,400 bills introduced in state legislatures this year on a wide range of election reform issues.
In December 2000, Florida Governor Jeb Bush established by executive order the bipartisan Select Task Force on Election Procedures, Standards, and Technology. The task force examined several issues associated with election administration and has issued its recommendations. As a result, Florida recently enacted major election reform legislation that eliminates punchcard ballots. Additionally, the legislation mandated a uniform election ballot design. At least four more states (Georgia, Maryland, Iowa, and Missouri) have proposed adopting a uniform statewide voting system, as well as other election reforms, and several have also proposed adopting systems that help prevent voter error.
Page 19 PREV PAGE TOP OF DOC
More than a dozen states have established task forces or other efforts to examine election reform needs, and some have produced recommendations. Also, a privately funded National Commission on Federal Election Reform, co-chaired by Presidents Carter and Ford, is examining a wide range of issues relating to voting technology and election administration. The bipartisan Constitution Project has established an Election Reform Initiative to develop consensus about improvements in election administration. And the California Institute of Technology and the Massachusetts Institute of Technology are engaged in a joint effort to determine how to improve the performance and reliability of voting systems.
III. Witnesses
There will be one panel of four witnesses:
(1) Dr. Stephen Ansolabehere, Professor of Political Science at the Massachusetts Institute of Technology and the Project Manager of the Caltech-MIT Voting Project. The Voting Project was created in December 2000 to prevent a recurrence of the problems that threatened the 2000 elections. Specific tasks of the project include evaluating the current state of reliability and uniformity of U.S. voting systems, and proposing uniform standards and quantitative guidelines for performance and reliability of voting systems. The Voting Project just completed a March 30, 2001 preliminary assessment of the reliability of existing voting equipment.
(2) Dr. Rebecca Mercuri, Assistant Professor of Computer Sciences at Bryn Mawr College, is a nationally recognized expert on voting technologies and standards. In October 2000, she successfully defended her Ph.D. thesis, ''Electronic Vote Tabulation Checks & Balances.''
Page 20 PREV PAGE TOP OF DOC
(3) Dr. Doug Jones, Associate Professor of Computer Science at the University of Iowa, has served on the Iowa Board of Examiners for Voting Machines and Electronic Voting Systems since 1994, and has chaired the board since the fall of 1999. This board, appointed by the Iowa Secretary of State, must examine and approve all voting machines before they can be offered for sale to county governments. The board meets whenever a manufacturer wishes to offer a new voting machine or a new modification of an existing machine for sale in the state of Iowa.
(4) Mr. Roy Saltman is a consultant and a retired employee of the National Institute of Standards and Technology (formerly, the National Bureau of Standards) who authored the 1988 National Bureau of Standards report, ''Accuracy, Integrity, and Security in Computerized Vote-Tallying,'' that first raised the difficulties of using punch cards and other machine-readable ballots. He also authored the 1978 National Bureau of Standards study, ''Science & Technology: Effective Use of Computing Technology in Vote-Tallying.''
Improving Voting Technologies: The Role of Standards
Chairman BOEHLERT. The Committee will now come to order. The Committee will now consider a request from the Ranking Minority Member regarding Subcommittee assignments. Mr. Hall.
Mr. HALL. Thank you, Mr. Chairman. First, I hereby notify the Committee that Congresswoman Lynn Woolsey has relinquished her seat on the Subcommittee on Research. And second, by direction of the Democratic Caucus of the Committee on Science, I have been asked to nominate Mr. Honda to fill the vacant Democratic seat on the Research Subcommittee. His personal background in K–12 education and science education makes him a very strong addition to that panel. And I have read some of his resume, and the thing that struck me first is that he raised almost $2 million to run on. And I am going to ask him to be my treasurer for the next go-round. And if he turns me down I am going to ask his opponent who is behind him with $1.3 million, and in my district that ain't bad. So—but I think he is a fine Member of Congress, a great American, and it's an honor to have him here. And I know that he will be supported, hopefully, on both sides of the aisle. Mr. Chairman, I ask unanimous consent that the nomination of Mr. Honda be ratified.
Page 21 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Hearing no objections, so ordered. I have had the pleasure of sitting down and having a conversation with Mr. Honda and enjoy his enthusiasm and welcome him with open arms to the Committee. This is a very productive committee, and I am sure you will make valuable contributions. We look forward to your service.
Mr. HONDA. Thank you very much, Mr. Chairman.
Chairman BOEHLERT. I would like to welcome everyone today to the Science Committee's first look at what I consider to be one of the most important aspects of election reform: ensuring that voting technologies are secure, reliable and accurate. Last November, as the world placed Florida under a microscope to scrutinize its election, America saw for the first time the extent to which one of our most sacred tenets of democracy was vulnerable to error. As the problems with the now-infamous punch card ballot unfolded, America watched in amazement, and the rest of the world watched in amusement. They watched as America, the most technologically advanced nation in the world, who preached democracy to others, tripped on its own shoelaces. Now, several months later, Congress is beginning to respond. And while it is tempting to criticize this institution for not acting sooner, I would suggest that the meantime has not been wasted.
We have learned many things about our nation's various voting technologies. And we have begun to answer many questions about how to fix our voting systems. For example, the problems that plague Florida also affected many other states and counties across the country. But would banning punch cards solve the problems for these areas? It turns out that it depends on what technology you replace the old punch cards with, for some are just as bad. Should states simply embrace the future and move to adopt voting by Internet? Probably not. For there still seems to be too many uncertainties over security, privacy and equal access to ensure that this potential solution would be fair or problem-free. How about electronic voting machines with touch screens or push buttons that automatically tabulate votes. Should states buy those? That is not really clear either, for some studies have shown those machines to be as error-prone on average as punch cards.
Page 22 PREV PAGE TOP OF DOC
So while our election problems are undeniably pressing, the last few months have shown us that we must proceed carefully in order to get the answers right, and not set ourselves up for potentially bigger problems down the road.
So what can the Federal Government do to help? First, let me state clearly what the Government should not do. It should not mandate a one-size-fits-all solution. It should not insist, for example, that every voting precinct in all states buy the same technology for its elections. Nor should it trample on the rights of localities and states to conduct elections in ways that make the most sense for them.
But the Federal Government can and should offer states the kinds of information that it is best suited to provide. For example, I believe the Federal Government should conduct research, not just into ways of improving the inner workings of voting machines, but also into how people interact with them to make sure that voting equipment is user friendly for ordinary people. It can also develop robust technical standards for voting equipment manufacturers to meet. And it can accredit labs around the country to help certify that voting equipment meets those standards so that states can confidentially choose voting technologies that are secure, reliable and accurate.
As the Committee that oversees the National Institute of Standards and Technology, I believe we have an important role to play in these areas. Next month we plan to work together in a bipartisan way to report out legislation that will help ensure that the Federal Government plays its proper role in strengthening our election systems.
Page 23 PREV PAGE TOP OF DOC
Obviously, there are other important ways the Federal Government can help, too. For example, by providing money to help states and localities upgrade their voting systems. As a former county executive, I can tell you how tight the budgets of local governments are stretched and how low a priority elections are as a result. Because they are competing with roads and so many other projects that traditionally are the main responsibility of local government.
But now it is time for elected officials who often like to lecture voters about how important voting is to our democracy to put our money where our mouths are. This is a complex problem with complex solutions. Solving it will take some time, some effort, and some investment. But in the end it is worth it. After all, in this case, it is no exaggeration to say that nothing less than our democracy is at stake.
I look forward to the testimony of our witnesses and to a productive discussion. The Chair now recognizes Mr. Hall.
Mr. HALL. Mr. Chairman, that was a good statesman-like rendition of your opinion of it, and I subscribe to it. I can add some things to it because, like you, I have had experience at the local level as county judge for twelve years of a small county, and the Texas Senate 10 years. I have observed the voting process. As a matter of fact, the absentee votes in our home county, they put them in a box and they had 3 keys and 3 locks on there. One for me, as county judge, 1 for the District Clerk, and 1 for the Sheriff. Now they could have been safe if they would have just had a fourth key and given that to Mother Theresa. But that is—without throwing it away to be absolutely sure that somebody wasn't going to look at some of those boxes.
Page 24 PREV PAGE TOP OF DOC
We—I add my welcome to you who are—have given your time and travel and all that to this hearing this morning. And I want to thank Chairman Boehlert for holding this hearing on election technology and the role of standards.
I think before November a lot of people reading this hearing title might have wondered about the connection between elections and standards. But the last election highlighted that we need to improve our voting equipment across the country. And I think our voting equipment has to be accurate and has to reliable. And good standards play a great role in that. I believe that the Science Committee can make a contribution to the ongoing debate of improving voter technology. This is especially true in the area of standards for voting technology and the research required to support the development of those new technologies.
The Chairman has assembled a very capable group of witnesses. And I want to thank him for working with us on this hearing. I also want to thank the witnesses for taking time from their busy schedules to appear before the Committee today. I understand that the legislature in Florida has passed some legislation. I am sure you have all studied that. Someone from a standpoint of humor said that they had a motion to reconsider and that they didn't know how it came out because they were down there in Florida and they were still counting. I am not sure that that's true, but you might expect that of Republicans. Because us Democrats, like up in Chicago the mayor said, every time two democrats voted they counted all 4 of them, by golly. And so there is a lot of improvement on both sides of the docket.
I think this Chairman is going to lead us into that improvement. It is going to take a little time, but we thank you for being a part of this this morning. I yield back my time unless you want some more of my stories. Yes, I do. Mr. Barcia is here and I am not sure I have any time to yield to him. But if I do, I want to yield the balance of my time to Mr. Barcia. He has been working on this issue and has a bill on this issue. And I support his bill. I think I am on the bill with him. And I yield you 15, 20 minutes, Mr. Chairman.
Page 25 PREV PAGE TOP OF DOC
Chairman BOEHLERT. The Chair is pleased to recognize Mr. Barcia.
Mr. BARCIA. Mr. Chairman, I will try to be brief. But I want to, of course, thank Ranking Member Hall and also for those kind comments. And I also want to thank you, Chairman Boehlert, for holding this very important and timely hearing. After the last election 1 issue that was highlighted was the need to improve the accuracy, integrity and security of the voting products and systems used in Federal elections. I felt that this was something that needed to be addressed immediately as states and local municipalities were likely to rush to replace existing voting equipment with newer equipment. I was struck also by the lack of objective information on how voting equipment performs. And that we lack a set of up-to-date, comprehensive standards for voting equipment. I believe that we need a set of comprehensive, voluntary, technology neutral standards that address the accuracy, integrity, and security of voting products and systems. In addition to the development of these standards, we should have an independent certification procedure as well as the means to accredit independent labs to certify voting equipment.
Finally, we need to—we need research on voting technology. This should include things like human factors, security and authentication issues surrounding Internet voting, and open standards, to name but a few. For 100 years we have relied on the National Institute of Standards and Technology, or NIST, to ensure our systems of weights and measures are accurate and reliable. And in that effort NIST has done an outstanding job. Common sense indicates that they could play a significant role in assisting in the development of standards for voting technologies. As far as back as 1975, NIST has been reporting on voting technology. And if we had heeded the recommendations in the 1975 report, and a subsequent 1988 report, we likely would not be holding this hearing today. I do not believe that NIST can be the sole arbiter of election standards, but by working with the industry, the FEC, and state and local officials, I do believe that a comprehensive set of voting standards can be developed.
Page 26 PREV PAGE TOP OF DOC
That is why I, along with a number of my Science Committee colleagues introduced H.R. 1165, the Election Voting Systems Standards Act. This bill authorizes NIST to work with the FEC to establish a commission to develop voluntary technology neutral voting standards, and an accreditation certification procedure for use by state and local election officials. The legislation is not meant to replace the current system of standards set by the FEC, but to strengthen the standard setting process.
In addition, the bill establishes a national research lab on voting technologies. The jointly chaired NIST/FEC commission would decide on the structure of the lab. This lab would not only fund research grants, but would serve as a central repository for research that would be open to all. We consider this legislation to be the first step in addressing this important issue.
Again, I would like to join Ranking Member Hall in commending Chairman Boehlert and also Ranking Member Ehlers for holding this timely hearing. I look forward also to the witnesses' views on the role of standards for improving our election equipment. And I want to thank them for appearing before the Committee today.
Chairman BOEHLERT. Thank you very much for that statement. Do you yield back the balance of your time? All right. Thank you. The Chair was somewhat indulgent on the opening statements. And all other members will have leave too with opening statements in the record.
At this juncture we will go to our first and only panel of the day, a very distinguished panel, I might add. Dr. Steven Ansolabehere, Professor of Political Science, Massachusetts Institute of Technology, and Director of the Caltech-MIT Voting Project. Dr. Rebecca Mercuri, Assistant Professor at Bryn Mawr College. Dr. Roy Saltman, Consultant on Election Policy and Technology. And Dr. Doug Jones, Professor of Computer Science, University of Iowa, Member of the Iowa Board of Examiners for Voting Machines and Electronic Voting Systems.
Page 27 PREV PAGE TOP OF DOC
And I might add as, is customary in this Committee, all of our panelists are very distinguished and have outstanding reputations. And, therefore, we are most appreciative of you being willing to serve as resources for this Committee as we are trying to come to grips with something that is very important to all of us.
With that, let me say we would ask that you try to summarize your statement in 5 minutes or so. I will not be arbitrary because you have got a lot to say and we have got a lot to hear, hopefully. And then we will open it up to questions. And we look forward to a productive couple of hours. Doctor, you are up first.
STATEMENT OF DR. STEVEN ANSOLABEHERE, PROFESSOR OF POLITICAL SCIENCE, MIT; DIRECTOR, CALTECH-MIT VOTING PROJECT
Dr. ANSOLABEHERE. Thank you for inviting me to speak today. I would like to begin by telling you a little bit about our project, and then tell you some specific findings of our project as they relate to technical standards and innovation in voting technology.
The week after the 2000 election President David Baltimore of Caltech called President Chuck Vest of MIT with an idea. And the idea was that our two institutions should collaborate to develop new voting technology. The problems observed in counting the vote in Florida and elsewhere ultimately originated with technology and often computing technology. Presidents Vest and Baltimore assembled a team of computer scientists, mechanical engineers and social scientists. The Carnegie Corporation and our two institutions have funded this endeavor.
Page 28 PREV PAGE TOP OF DOC
I am Steve Ansolabehere, a Professor of Political Science at MIT. I am one of the social scientists, and co-director of this project. My counterpart at Caltech is Tom Palfrey, an economist. Our team spans many different fields. The engineers bring expertise in electronic security, user interface design, machine design and performance standards. I speak for the group.
We are in the initial phase of our project, which I consider the learning phase. Over the last four months we have met with many voting machine manufacturers and elections administrators to ascertain where the problems are and to explore some ways that we may contribute to the solutions.
We have also conducted studies of voting machine performance and design, the public finances of election administration, and voter registration practices. A complete report of our work over the last 4 months will be out July 16.
What are the problems with voting technology? Well, there are many detailed problems you could list. But we are going to focus on 4. First, is the high rates of spoiled, unmarked and uncounted ballots, such as we saw in Florida. Our first step in this project was to assess the performance of existing equipment in order to establish some benchmarks. I have distributed a report, which is the second version of a report that we issued in the end of March on the extent to which equipment affects the rate of uncounted, unmarked and spoiled ballots. The report shows or finds that punch cards, indeed, performed relatively badly compared to the other technologies, according to the standard of uncounted, unmarked and spoiled ballots. Surprisingly, electronic equipment did as well. Paper ballots that are hand-counted, lever machines and optically scanned ballots performed best by this metric.
Page 29 PREV PAGE TOP OF DOC
In our assessment, these differences are due to ballot and user interface designs more than they are due to the durability of the counting process. We feel that greater private or government research and development is required to improve ballot design and user interface design in this regard.
The second problem, errors in the voter registration data bases. Voter registration is a very large data base management problem, and as such, a technical problem. The roles must have high levels of integrity and they must be accessible at the polling places. An audit of Michigan's voter roles, after they went to a qualified voter file at the state level, revealed that 1 million voters had duplicate registrations out of about 9 million. According to the 2000 census, 7 percent of voters who did not vote reported registration problems as the reasons for not voting. So we see that there are some big issues with access to the poles and the quality of the data or the integrity of the data that need to be addressed.
The third issue, accessibility. Put simply, blind people really can't vote without assistance.
The fourth issue, security of ballots and tabulation. The traditional process is having many eyes on the process to check the quality of the tabulation. As we move to electronic equipment that is becoming impossible. So the natural check on the integrity of the count is being lost. Today we are on the verge of Internet voting. Simply put, we need standards for secure electronic voting and we see those as lacking.
What are standards and how can they help? The first set of standards is minimum criteria. We have those and the existing standards that are implemented by NASED, minimum criteria for accuracy of tabulation and durability of equipment. We think that there need to be minimum criteria for usability, accessibility and auditability.
Page 30 PREV PAGE TOP OF DOC
Second, best practices. There are model programs out there for registration, for equipment maintenance, equipment development, and this should be developed further. A Federal agency should compile and distribute information about equipment performance, cost, and administration. That is essential if counties are to become smart consumers of equipment.
The third standards are specifications. Our group is against the idea of uniform equipment, which in the extreme is a specification that everybody must use the same form of equipment. However, there are areas where uniform specifications might help, such as in electronic security. Here, we see a special role for NIST and the Science Committee. Internet voting is on the verge of being here and will grow over the next 10 years, the pressures are very strong. The biggest objections have to do with security. And one thing we feel strongly about is that this Committee and NIST begin a process now to develop standards for secure Internet voting.
Chairman BOEHLERT. Thank you very much. Dr. Mercuri.
STATEMENT OF DR. REBECCA MERCURI, ASSISTANT PROFESSOR, BRYN MAWR COLLEGE
Dr. MERCURI. Thank you and good morning. I am Dr. Rebecca Mercuri of Lawrenceville, New Jersey, an Assistant Professor of Computer Science at Bryn Mawr College, and also president of Notable Software, Incorporated, a New Jersey computer consulting firm. Thank you for the opportunity to address your committee on this important matter.
Page 31 PREV PAGE TOP OF DOC
Over the last decade I have investigated voting systems with particular emphasis on the electronic equipment used to collect and tabulate ballots. Through this research I have identified numerous flaws inherent to the application of computer technology to the democratic process of elections. As it turns out, these flaws are both technologically as well as sociologically based, so a quick or even long-term fix is not readily apparent. For example, present and proposed computer-based solutions are not able to resolve, and in some cases, may even increase the likelihood of vote selling, coercion, monitoring, disenfranchisement, and fraud in the election process. Some of the problems with electronic balloting and tabulation systems are identified as follows:
First, fully electronic systems do not allow the voter to independently verify that the ballot cast corresponds to that one that was actually recorded, transmitted or tabulated. Any programmer can write code that displays one thing on the screen, records something else, and prints out something else as an entirely different result. I have freshmen, by the way, who can do this. There is no known way to ensure that this is not happening inside of a voting system.
Number two. Electronic balloting and tabulation makes the tasks performed by poll workers, challengers, and election officials purely procedural and removes any opportunity to perform bipartisan checks. The process is thus entrusted to the small group of individuals who program, construct, and maintain the machines.
Number three. Although, in many states, convicted felons and foreign citizens are prohibited from voting in U.S. elections, we don't have any laws regarding voting system manufacturers, programmers and administrators in that regard.
Page 32 PREV PAGE TOP OF DOC
Fourth, each election season, newly deployed voting equipment fails to perform properly in actual use. Communities will continue to run the severe risk that they will administer an election whose results may be contested.
Five, electronic balloting systems without individual printouts for examination by the voters do not provide a wholly independent audit trail. As all voting systems are prone to error, the ability to perform a manual hand-count of the ballots is essential.
Six, some new electronic systems actually make the balloting process more lengthy, tedious, and confusing. The use of such devices has been viewed by some as a modern-day literacy test.
Seven, encryption cannot be relied on to provide end-to-end privacy assurance. Nor can it assure the accuracy of ballot data recorded and tallies rendered.
Eight, Internet voting provides avenues to the entire planet for malicious denial of service attacks. Off-site Internet voting also creates unresolvable problems with authentication, leading to possible loss of voter privacy, and increased opportunities for vote selling. Internet voting may make it easier for the techno-savvy elite to cast ballots, while potentially disenfranchising or creating a digital divide for poor, elderly, rural and disabled voters.
And ninth, it is not possible to construct a standardized voting system that could be used in all municipalities without treading seriously on states' rights issues, and without mandating changes in many conflicting election code laws to provide conformity.
Page 33 PREV PAGE TOP OF DOC
I could go on, but these are just some of my concerns, many, many more exist. Now the computer industry has already established standards for secure certification mandated by Congress under the Computer Security Act of 1987. NIST typically administers this certification for devices purchased by the Department of Defense. But Congress, though, exempted itself from compliance with the Act, and has never certified the accuracy and integrity of computer-based voting systems used in Federal elections under this code. This loophole should be changed. The existing standards such as the ISO Common Criteria or its predecessor, TCSEC/ITSEC, are not perfect, but they are the best assurance mechanism that the computer industry has at present.
I have long recommended that the NIST standards be applied to voting systems. As a part of my Doctoral Dissertation at the University of Pennsylvania, I performed a detailed analysis of the Common Criteria against the features that I thought should be necessary in voting systems. The painstaking description in the thesis provides an excellent starting point for the development of a voting standard. I have provided it to the Committee.
I have also formulated a list of questions that voting system vendors should be required to answer about their products. I would suggest that first a trial standard be developed, along with an assessment procedure. And then voting systems applying different state requirements should be constructed and assessed to see what level of conformance is possible using current technologies.
In conclusion, I would like to remind the Committee that technology can not and does not, at present, provide a solution to the balloting and tabulation problem. Our society has become increasingly enamored with computers, yet we have all experienced, first-hand, their sometimes catastrophic failures in the products that we use every day. The same will be true and is true for computer-based voting systems. But here there are no warranties, and there is no insurance if we have some problems with the results.
Page 34 PREV PAGE TOP OF DOC
It is, therefore, crucial that we continue to maintain and impose human checks and balances throughout our election process. This is the only way to ensure that our democracy does not become one that is by the machines, of the machines and for the machines. Thank you.
Chairman BOEHLERT. Thank you very much, Dr. Mercuri. Boy, you gave us something to ponder on. I am glad you weren't like David Letterman and went through a top ten list. I am glad you stopped at nine. It was getting discouraging.
Dr. MERCURI. I stopped at nine. Thanks.
Chairman BOEHLERT. But thank you very much for an outstanding testimony.
Dr. MERCURI. Thank you.
Chairman BOEHLERT. Mr. Saltman.
STATEMENT OF ROY SALTMAN, CONSULTANT ON ELECTION POLICY AND TECHNOLOGY
Mr. SALTMAN. Mr. Chairman and and members of the Committee, thank you for permitting me this opportunity to present my views. I want to first indicate to you my background. As Mr. Barcia has mentioned the two reports of the National Institute of Standards of Technology, and it is important for you to know that I was the author of both of those reports.
Page 35 PREV PAGE TOP OF DOC
It is interesting to note, however, that the money for those reports did not—was not appropriated to NIST. The first report was sponsored by the General Accounting Office with the approval of Mr. Staats, the Controller General at the time, Sam Hughes, the Assistant Controller General, and the Office of Federal Elections, which was then the predecessor of the Federal Election Commission and was in the General Accounting Office.
The second report was sponsored by—in 1988 was sponsored by the John and Mary R. Markle Foundation. And the reason was that there had been some New York Times articles that said that fraud was a possibility and due to computer program manipulation. And the Markle Foundation wanted to know what was the issue. So that money came from there. At no time did NIST ever appropriate any money for the—for projects on election administration.
I want to tell you what I think the Federal Government could do in this area. Surely, the Federal Government should not conduct elections. But in other civil—areas of activity like education and health and highway transportation, the Federal Government has significant activities but does not undertake the delivery of services.
Specifically, the Federal Government should do this. It should undertake data collection, analysis and reporting. It should carry out research. It should establish robust national standards. It should accredit independent testing laboratories. It should support state-wide voter registration systems. And it should have a grant program for states and local governments.
Specifically, documentation is needed for equipment—reports of equipment performance and incidents of difficulties with either machines or procedures. Data concerned with the ability or inability of voters to use voting machines successfully with collection of under-vote and over-vote data, for example, it would be useful to obtain on an organized basis.
Page 36 PREV PAGE TOP OF DOC
In research, studies of the ease of use, that is, voter usability of different vote-casting methods should be undertaken. Product independent research into new types of voting systems and use of the Internet should be undertaken. Analysis of techniques for improving the capability of sight-limited voters to vote without assistance should be done. And development of new methods of voter identification that could be applied to precinct located voting or to remote voting should be undertaken.
I agree that the standards established by the Federal Election Commission need some updating. That process is underway. But it also requires significant amounts of resources to undertake and continue the undertaking of that project.
The accreditation of independent testing laboratories is one area where NIST could be of value. Their national laboratory—or their National Voluntary Laboratory Accreditation Program is perfectly suited to assist in this effort.
State-wide voter registration systems need considerable support. As Professor Ansolabehere mentioned, there is a significant problem. I believe that it can only be solved through state-wide voter registration systems with Federal assistance, with county-level voter registration systems coming off that, and with terminals from two of the precincts and from the precincts during election administration on election day.
I want to point out that I do not have the same concern about fully electronic voting that Dr. Mercuri does. However, I have always stated that the auditability problem of non-ballot systems is a serious matter that must be undertaken with great care to assure that the public has confidence in the results produced by an all electronic system.
Page 37 PREV PAGE TOP OF DOC
I have proposed in my 1988 report two items for the design of these direct recording electronic systems one of which was introduced into the Federal Election Commission standards and the second was not. The second involves recording the non-voting possibilities of voters and will allow a second chance effort by voters to replace their under-votes.
Thank you for your discuss—for your listening. I have in my written testimony my discussion of the advantages, disadvantages and recommendations for change of the different voting systems. Thank you for your attention.
Chairman BOEHLERT. Thank you very much, Mr. Saltman. Dr. Jones.
STATEMENT OF DR. DOUG JONES, PROFESSOR OF COMPUTER SCIENCE, UNIVERSITY OF IOWA; MEMBER, IOWA BOARD OF EXAMINERS FOR VOTING MACHINES
Dr. JONES. Thank you. I am very glad to be here and hope that my testimony is useful. Going last I feel that is rather difficult to say more than amen. But there are few things which I have to contribute, which are different.
In 1992 I was appointed by the then Republican Secretary of State, Paul Pate, to the Iowa Board of Examiners for Voting Machines and Electronic Voting Systems, and I have served on that board now for 9, going on 10 years, and in that time I have had the opportunity to examine a large number of voting machines. In this position, I have dealt with the regulatory process as a unit, the combination of Federal regulations and state regulations. And I have had numerous opportunities to deal directly with vendors, representatives, to directly speak to vendors, software people and hardware people. And to deal with a fair number of county elections officials and state elections officials within the State of Iowa. I now chair that commission. And it is in that role that I am here.
Page 38 PREV PAGE TOP OF DOC
I have over the past almost a decade of experience looking at voting machines, seen a large number of systems come before our board that have been deficient, that we have turned down for one reason or another. And almost every one of these systems has been—has passed the current Federal Election Commission Voluntary Certification Process, has been examined by Wile Labs, which is the only independent testing authority currently approved. Almost every one of these systems that we have had difficulty with has been approved for use in a large number of states. I don't believe that we set unreasonable standards. In fact, I don't believe that we can set the standards we ought to right now because we cannot turn down a machine for use if it does—if it passes the legal requirements. And there have been many occasions where I wished we could turn down a machine.
Because of this, I think we have to look very carefully at several issues. I want to very strongly back the—Rebecca Mercuri's comments about electronic voting, fully computerized voting systems where there is no independent record of the voter's action which the voter could—which the voter could check before depositing a ballot in a ballot box, is extremely problematic. I don't believe it is impossible. I think I am more optimistic than Rebecca is about that. But I don't want to encourage people to rush for computerized voting systems. And here I refer to direct recording electronic voting systems.
Internet voting is even worse. I would, quite frankly, like to forbid Internet voting until we have some really solid standards in place. I feel very strongly, as Mr. Saltman does, that we need much harder audit requirements. I believe that the over-votes and under-votes in every race should be included as part of the official canvass brought forward to the level where it is included, so that you can verify that the sum of votes for each candidate, plus under-votes for that race, plus over-votes for that race equals the number of ballots. That is such a trivial accounting trick. And it would account for a huge number of the lost ballot boxes and strange things like that that we find in typical recounts.
Page 39 PREV PAGE TOP OF DOC
It is very important to realize that Mark Twain was right when he said that democracy is the worst of all systems, except for all the others. It is fair to say as a corollary that every approach we have to conducting a democratic vote is bad, except that the alternative is worse. There is no perfect voting technology available today. Because of this, I am very wary about the current rush to abandoned the ''old tech'' voting systems, and immediately adopt the highest tech that we can find. I would find it very inappropriate for the Federal Government, for example, to fund a massive emergency purchase of new voting technologies in counties across the country. Because I feel that in doing so we might be simply replacing the bad system we know we have with the system whose failings we don't understand yet.
I recommend that we slowly phase out, deliberately, voting systems whose problems are clearly worse than the others. And among these I would say punch card voting is clearly bad. And I would suggest that we move cautiously forward on new technologies. I don't like the idea of declaring an emergency and throwing a lot of money at purchase of new hardware, only to discover we have made a mistake. Thank you.(see footnote 1)
In my experiments with this machine, I was able to reproduce some of the problems reported in the press that led voters to complain that even when they hammered on the punch, they were unable to cleanly punch their ballot. The cause of this problem is chad accumulation between braces inside the machine and the back of the ballot. This problem only occurs where the voting position is over these braces.
I strongly recommend that, if Votomatic voting machines remain in use, the voting positions above these internal braces should never be assigned to candidates!
Page 40 PREV PAGE TOP OF DOC
See http://www.cs.uiowa.edu/jones/cards/chad.html for photos and additional documentation of this issue. I will update this web page as I learn more.
DISCUSSION
Chairman BOEHLERT. Thank you very much. You are quoting Mark Twain. I will quote Pogo. We have met the enemy and he is us. You know, Mr. Saltman, you said that NIST didn't appropriate any money for your studies. Well, NIST doesn't appropriate, we do. And shame on us for not doing what we should have done before. But we are not going to concentrate on the past, we are going to look to the future.
But how many days, 1,275 to the next election? Can we do it in time? I am talking about the presidential election. Is that a realistic goal to establish, that we take corrective action to prevent a repeat of year 2000 in the next 1,275 days? Or is this going to require longer? Does anyone have any comment on that? Dr. Jones.
Dr. JONES. It will take considerably longer to put new standards in place and see a new generation of voting systems that satisfy the requirements of those standards.(see footnote 2)
I believe that the root of our disagreement rests in the height of the goals we inferred from Chairman Boehlert's question. I agree with Dr. Mercuri that we can indeed take some corrective actions quite quickly—we can get rid of the worst of our current voting systems, for example. I am not so optimistic about the speed of the federal or state rulemaking processes, though. Some of the changes we must make involve legislation in the 50 states, and if we have a new voluntary standard, it will take time for the states to adopt this standard.
Page 41 PREV PAGE TOP OF DOC
I am also not optimistic about the ability of the voting system vendors to meet a new standard. There is a shortage of technical manpower in the world of secure systems, and this is the manpower pool we must tap if we are to insist on voting systems that meet the Common Criteria, as Dr. Mercuri has advocated so effectively. My experience is that few of the vendors have much software expertise in this domain, so my guess is that, even if we promulgate a perfect standard and all 50 states adopt it, we would see few new products that meet our demand.
Chairman BOEHLERT. You—do you others agree with that or do you think we can do it? Dr. Mercuri.
Dr. MERCURI. I am actually rather optimistic. I have been pushing since the early 1990's to apply the NIST standards that are already in place for secure computer systems. These standards are the best that we have, and we can apply these to voting systems. It is just that nobody has mandated it and nobody has voluntarily done it. I have spoken directly with a number of the vendors and asked them to voluntarily comply with it.
In my analysis there are seven levels of the Common Criteria. And in my analysis I said we could get up to satisfying level four with some systems. There are systems out there that are used for the Department of Defense that do satisfy level four. So we could theoretically come up with some system that I would actually feel comfortable with. I am not as harsh about it as I might have sounded.
Chairman BOEHLERT. Mr. Saltman. Yes, sir.
Page 42 PREV PAGE TOP OF DOC
Mr. SALTMAN. First, permit me to apologize for using the word appropriation. What I meant was internally appropriated funds.
Chairman BOEHLERT. You don't have to apologize. I just wanted to put the blame where the blame——
Mr. SALTMAN. Oh, okay.
Chairman BOEHLERT [continuing]. Should rest, and that is with Congress.
Mr. SALTMAN. What I am distinguishing between is internally assigned money for projects as opposed to other agency funds of which NIST has considerable amounts of. That is what I am distinguishing.
Chairman BOEHLERT. Okay.
Mr. SALTMAN. But in addition, I think the problem is that we have a many older systems in the United States. We have 19 percent of the voters vote on lever machines, and——
Chairman BOEHLERT. Which incidentally, if I may interrupt here——
Mr. SALTMAN. Yeah.
Page 43 PREV PAGE TOP OF DOC
Chairman BOEHLERT [continuing]. Interject reading some of your material, some of the advanced testimony. The lever machines seem to be the least error-prone. And they are not even manufactured anymore.
Mr. SALTMAN. The problem with lever machines is as I was told by a very astute election administrator, who I am sorry, Connie Morella, isn't here right at this moment. But she knows very well a Marie Garber who was the election director of Maryland's Montgomery County and later State Director of Maryland Elections. And Marie told me that the problem with lever machines is that they hide their errors.
There is a problem there that—a test was done on lever machines and it was found that the number of votes equaling 99 for any candidate was more likely than any other number. And that shouldn't be on a random basis. And the reason is that the inertia of the counters turning from the units to the tens to the hundreds positions is more—is greater. The inertia is greater and the issue is that it would be more——
Chairman BOEHLERT. We don't want to get too technical.
Mr. SALTMAN. I am sorry.
Chairman BOEHLERT. So we shouldn't all just jump on lever machines right now.
Mr. SALTMAN. Yes.
Page 44 PREV PAGE TOP OF DOC
Chairman BOEHLERT. They are the most accurate of the various——
Mr. SALTMAN. They have their own errors. All right. In addition, we would like to get rid of prescored punch cards, most of us would here at this panel. But I am told, for example, by Connie Morella—I am sorry, Connie McCormack of Los Angeles County, that—which is the largest user of prescored punch cards in the nation, that she likes them. And it would take an enormous amount of money to replace them with electronic machines. So it—the likelihood of there being taken out in a short time is very small.
Chairman BOEHLERT. You know what boggles, I think, people that I talk to back home all the time, I am sure my colleagues go through the same experience, I can go virtually any place in this country today, from a major urban center to a small four-cornered town, and get cash instantly out of a machine for a modest service fee. Now we don't want people to have to pay to vote. But the issue is the technology. I mean, the technology is there, it is basically very secure. I can only get cash from—all right. Dr. Mercuri.
Dr. MERCURI. I totally beg to disagree with you on that. The banking industry loses billions of dollars at those ATM's. The difference is, and I sort of alluded to that in my concluding statement, that the banking industry is insured. They also have auditing. And when you go to the bank to deposit money or take it out, it is registered to you. But we don't do that with our votes. And in fact, if we recorded them and had a little camera there, then that would not be exactly what we want to have in the way of an anonymous ballot.
Page 45 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Yes.
Dr. MERCURI. So the way that banks actually audit their banking systems, we can't do. So it throws it into this sort of dark, gray area where the electrons are taking this away. And we have to be assured that the electrons are going in the right pockets and in the right holes. And so it is not at all the same. I do address that quite at length in my thesis.
Chairman BOEHLERT. Oh, good.
Dr. MERCURI. So please read that section.
Chairman BOEHLERT. Okay. I will.
Dr. MERCURI. It is not at all the same.
Chairman BOEHLERT. My time is up but I will let you respond to my question before I go to my colleagues. Mr. Saltman and then Dr. Ansolabehere.
Mr. SALTMAN. The point I wanted to make is that the advantage of a banking system is that every depositor has an account and is always avail—capable of obtaining the value of the amount of money in the account. Unfortunately, in elections this is not the case. If you try to analyze elections with banking in somebody else's account, the candidate's account, not your own. And, therefore, there is no way of determining by yourself how many votes were deposited in that account. And that is a difficulty with the election system which makes it even more difficult than the banking system in the terms of an auditability situation.
Page 46 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Dr. Ansolabehere, did you want to——
Dr. ANSOLABEHERE. I wanted to address your first question.
Chairman BOEHLERT. Okay.
Dr. ANSOLABEHERE. There is a constant change in technology. Counties are constantly upgrading. The industry is about $150 to $200 million in sales a year. So you don't get that at small business, but you don't get that much revenue without there being something going on on an annual basis. The map over here of voting equipment by counties is going to change a lot by 2002, not just 2004. So there will be no more punch cards in Florida, we already know that one. There will be no more punch cards in Indiana. There will be—a lot of California uses punch cards and those are being phased out. So the states are working right now to phase out punch cards. Florida was enough of an embarrassment that people have lost confidence. That leaves really Ohio and Illinois as the two large population centers—large population states that are still using it. And I think those will teeter.
By 2004 then I think punch cards will largely be gone. And in their place will be one of two types of machines, either optical scan or electronics. And the big question is, how much confidence do we have in each of those. That is a short-run solution. Then there are the long-run solutions that have to do with developing new technology. And it is pretty clear that even these two newer technologies are not what we want them to be.
Page 47 PREV PAGE TOP OF DOC
Chairman BOEHLERT. I have got as many questions as votes were disallowed in Florida. But my time has expired and I go to Mr. Hall.
Mr. HALL. Mr. Chairman, thank you. I have always heard, I don't particularly believe it, but everybody—that people that live on the ocean and live on the sea don't like fish. Now you are all computer experts. And yet, if I understand your testimony, none of you recommend computer-based equipment. And there has been a push in my area to upgrade voting equipment with computer-based equipment. All of you seem to suggest that local officials shouldn't rush to purchase new equipment. Well, I don't know how they are going to get it if they don't purchase it. Or how they are going to upgrade their equipment if they don't turn to computers. So what do you see as the most common downfall in their turning to computers?
Dr. ANSOLABEHERE. I think it is the user interface design of the computer. The ballots are somewhat confusing to voters and a lot of voters seemed to be intimidated by them. There are a lot of people who walk out of the ballot——
Mr. HALL. Well——
Dr. ANSOLABEHERE. So improving the user interface is a very important thing. There is no standard ballot toolbox. So if you are a county administrator, you have to format. In LA County you have to format 5,000 ballots for five different languages. All right. Very big task. The computer company has to do it. There was recently a bake-off in LA County. And of the five computer companies that were trying to sell these machines only one of them could meet that goal. So just formatting the ballots in a user-friendly way is very difficult.
Page 48 PREV PAGE TOP OF DOC
And with hand—with printed paper ballots it is a much easier task a lot of the time. Not always.
Dr. MERCURI. My concern is the technology itself. I think because we are all so familiar with the technology and we live with it every day, we know the corruption that can actually occur internally. The fact is that we have no way of eradicating viruses in our own industry. And that people in businesses are plagued by this on a daily basis and it costs millions of dollars to eradicate some of these huge viruses. We don't want to see that type of thing happening with our election systems.
And so it is the technology, the underlying technology itself. And to come up with a way where we can mitigate these problems that are just in the computer industry and to try to eliminate them in our election systems. If we could do that in our industry itself, we could do it for election systems. But we are at a loss to be able to do that.
Mr. HALL. Gentleman, yes.
Mr. SALTMAN. I wanted to reiterate my insistence on the importance of the auditability issue in non-ballot systems such as DRE's. I fully agree with Rebecca that Internet voting is way beyond our capabilities at this particular moment. However, I think that the use of DRE's in inevitable and will increase in this country very soon. Because of that case, it is extremely important that we establish a very tightly established set of procedures to assure that first all of the computer programs that are used in DRE equipment are fully tested for correctness. That security procedures be used to assure that that—those computer programs are not manipulated or tampered with. Third, that the——
Page 49 PREV PAGE TOP OF DOC
Mr. HALL. Is that the major danger that you see?
Mr. SALTMAN. Excuse me?
Mr. HALL. Is that the major danger that you see, manipulation or fear of tampering?
Mr. SALTMAN. I see it as a danger. I couldn't say that it is a major danger. I think that simple computer bugs are probably more likely than deliberate manipulation. But computer bugs do happen all the time. And people tend to——
Mr. HALL. You got some buggy people counting votes, too.
Mr. SALTMAN. Well, yes. That is true. In any way, we want to be sure that the computer programs that are used in the computer are the same ones that are tested exactly. And I think security techniques can be sure of that.
And in addition, that we apply to the machines a procedure, and I think Doug Jones mentioned it, that the cross-check between the number of people using the machine and the number of people who voted for different candidates or didn't vote or over-voted are equal to each other. So that is a good cross-check. And that also makes it possible to apply the second-chance opportunity for voters to try again if they fail to vote in a race that was important to them.
Page 50 PREV PAGE TOP OF DOC
Mr. HALL. Yes, sir, Dr. Jones.
Dr. JONES. I think that the question of what is the greatest problem with electronic voting systems has two sides to it. The human factor side is the greatest immediate risk of adopting this technology. Because in the very short-run this technology is—has a poorly developed set of user interfaces. Our track—our experience is not very great on how to do it right. In the long-run we risk creating a situation where we end up putting our democracy in the hands of a small number of programmers with poor oversight. And any time you put your democracy in the hands of a small number of people, those people will emerge as the Praetorian Guard eventually and select a new emperor without consulting the people.
Mr. HALL. Well, my time is up. But let me just point out, Mr. Chairman, in an article by Florence Olsen, that I think Dr. Ansolabehere is familiar with, the lever machines showed only a 1.6 percent of ballots spoiled while electronic voting, that you are downing down here today and criticizing, had the same percentage as punch card voting. Each of them 3 percent of ballots spoiled. And lever machine 1.6, the lowest of all. And the paper ballots, that everybody is complaining about, was lower than either the punch card or the electronic voting. So it doesn't get right down to whether or not they are honest and whether or not people are counting the votes or given the number of votes that one candidate has to get. And I found most people were most upset over the fact that the military people weren't getting to vote. And you can cure that by sending them to them early and getting them back early. And setting a time certain to cut them off.
But I don't know, it seems like we need to improve on people more than we need to improve on the method of voting.
Page 51 PREV PAGE TOP OF DOC
Chairman BOEHLERT. The Chair now recognizes Dr. Ehlers.
Mr. EHLERS. Thank you, Mr. Chairman. And I will have to be fairly brief because I am due on the Floor to speak on the education issue. But I have a lot of experience at voting, as many do here. I have served at the local level as well as the state and Federal levels. And a problem that most Americans don't recognize is that there are a lot of people out there deliberately trying to subvert this system, or at least change the result. We tend to have a great faith in fellow human beings, especially fellow Americans. And so it is a huge surprise when it happens. But it does happen. I was just talking to one of my colleagues from a state recently and he commented that one of his county chairman got in trouble because not only had he voted in two different precincts but his dog had voted in three different precincts. And he said, the public was not upset that his dog voted, but that his dog voted three times. That was just—that was going beyond the pail. So this is—happens far more regularly than people think.
Let me just—I want to get on the record what I think is the problem. First of all, we need more accurate registration. We haven't talked about that here at all. I think probably the greatest fraud occurs in the registration process in terms of felons registering and non-citizens registering and so forth. Then we need an accurate people-friendly, fool-proof voting system. And I would emphasize the people-friendly and fool-proof more than I would the accuracy. We can solve the technical problems fairly easily. The human factors of voting are important.
As an example in the electronic voting that you have discussed, one of the problems is you only get one shot at it. And even Regis Philben always says, ''Is that your final answer?'' And we should do the same in voting on the display. Every time you vote, ''Is that your final vote?'' We also, of course, should print out a sheet for the voter, but also one for the tellers so there is a dup—paper duplicate of the electronic tally.
Page 52 PREV PAGE TOP OF DOC
We need accurate, honest counting. And that is a factor that is often overlooked as well. It—we have poll workers who are the salt of the earth. They are wonderful people, I know many of them. But they only do this once or twice a year. And it is easy to forget the instructions, procedures and so forth. So we have to develop very accurate and honest methods of counting.
Finally, we certainly need public confidence in the result. And I am afraid that we do not have that at this point. I think we probably had too much public confidence in the result in the past. The public wasn't aware of the many errors and illegal activities that occur. Now we have gone the other way and the public doesn't have much confidence.
I—just one final comment in response to the criticism I have heard here of punch card systems. I am a little surprised at that. Because properly maintained, the punch card system, and properly designed books to go on it, is surprisingly good. And in fact, in my community, my major community, Grand Rapids, Michigan, they not only have a punch card system but they also have tabulators right at the polling booth that the voter puts in the—their ballot in the tabulator. It is counted immediately. If they have over-counted it is rejected and they get a second shot. So you correct one common error right at the polling place. I think properly operated punch card is valid and probably more valid than the electronic systems given the present state of electronic systems.
I will be happy to hear any comments or criticisms of these statements from anyone here. Yeah.
Page 53 PREV PAGE TOP OF DOC
Dr. ANSOLABEHERE. I think you are exactly on target that giving voters lots of feedback is the first thing you want to try to build into a system. Punch card counters and scanners are very important innovations. The optical scan systems that work best are ones where you can scan your ballot right in the precinct. It is not, you know—anywhere that they have worked perfectly. But that degree of feedback is very important for solving these human issues.
Mr. EHLERS. And it is also, if I may interrupt, it is very important to have a paper trail, too.
Dr. ANSOLABEHERE. And public confidence, I think that is why we are all here, to get the public confidence back. And so that people are using a machine that they believe in is very important.
Mr. EHLERS. Mr. Saltman.
Mr. SALTMAN. I think your comments are very well taken, sir. However, and I would say that my written testimony talks that precinct counting is far better than central counting for the very reason that you have indicated. However, it is my feeling about prescored punch cards that the whole requirement that makes them good, which is that election administrators and their poll workers have to sit around and personally remove the chad. It seems to me to be a contradiction of computerized voting. I know that they do this in Washington State before they run the machine—the card through the machines. And of course, it is certainly true that the difference between a hole and a no-hole when they are totally obvious is certainly very easy for a computer reader to do. However, when there is chad and it is not removed, and a lot of voters just simply fail to do that, that creates a problem. And in my opinion, it creates much of a problem—more of a problem than optical scan cards, which are not destroyed in the process of voting them. And always are there for the voter intent to be determined and you are never—and you can be sure, as you can't in prescored punch cards, that some chad has either fallen out or not fallen out or pressed back in. So I think in general, my feeling is that optical scan cards are superior in their use.
Page 54 PREV PAGE TOP OF DOC
Mr. EHLERS. Dr. Jones?
Dr. JONES. I share this feeling about punch cards. I agree that properly maintained and operated they are probably not as bad as some people hold today after the previous election. But they should be phased out and deliberately.
However, I would like to go back to an earlier comment you said about issuing receipts to voters so that the voter can be assured of how they voted. That receipt can become a commodity which can be sold to the party willing to pay the highest for that vote. Vote buying and selling is something none of us would like to believe exists, but it has happened in the United States since we began voting. If you provide the voter with any sure-fire way to prove that they followed the instructions of their union boss or their employer or anyone else who was trying to buy their vote, then you produce something very dangerous. In fact, it is the absolute prohibition of this receipt that is required in order to maintain the integrity of our system, which leads to so many of the problems with electronic voting.
Mr. EHLERS. I agree with that. That receipt has to be handed in before they leave the room. Dr. Mercuri, very quickly. My time has expired.
Dr. MERCURI. I would like to also state my support for the scanned machine—the scan-type of machines. Because they do provide the voter with some tangible way of looking at the ballot that they prepared. It is much better if the names of the people are actually listed on that ballot. One of the problems we have with the punch card is you just have this encrypted bunch of holes and you can't identify it yourself. So if you have that visual feedback, then that is really good.
Page 55 PREV PAGE TOP OF DOC
But our entire government is based on a system of checks and balances—what we don't have and what we lose when we feed those ballots in and they are read by a single machine. We don't know that they are being tallied correctly. I like the fact that the press got involved in Florida and decided to do their own recount. I would like to see that various agencies, the different parties, the League of Women Voters, maybe the press, they each submit their machines for tallying. They have to be certified in advance, and then there is multiple recounts. And if there is some huge percentage difference then we maybe go to a manual recount of those ballots. I think that would also give us that checks and balances, and the feeling that these ballots aren't just whisked away and there is the total and you have to be stuck with it. That you actually do get the competing parties, they submit them in advance and that type of thing. I would like to see that happen.
Mr. EHLERS. My time has expired. Thank you, Mr. Chairman.
Chairman BOEHLERT. Dr. Jones, except for the good of the cause, we will very briefly.
Dr. JONES. I have been advocating for a while the idea that selected—that there should be a mandatory hand recount of at least one race in at least one precinct of every jurisdiction after each election using recountable ballots of any kind. Simply number one, to keep people skilled in the art of hand recounts. If you do it after every race with one race in one precinct, there will be people who have done it before. And also simply to keep track, to do a quality control check on the voting machines that we are using to automatically count those ballots.
Page 56 PREV PAGE TOP OF DOC
Chairman BOEHLERT. It boils down to dollars, doesn't it? Mr. Barcia.
Mr. BARCIA. Thank you, Mr. Chairman. I will try to be brief. Dr. Jones, in the 10 years since the FEC issued its standards, only 20 states have opted to require a conformance. Why have most states not required conformance with the current FEC standards? And when the FEC issues its revised standards, do you believe there will be wider acceptance by state authorities?
Dr. JONES. I—there is a mistake in that figure. When you are trying to prepare testimony with three day's notice you don't always proofread carefully. It is actually a small majority of the states. We are over 30 now. But it has taken a decade to get there. I think the states have been hesitating to sign onto this process, number one, because signing onto this process requires that you tell your state election authority that their inspections haven't been adequate. And most states would like to believe they are doing just fine. It also means that the state is handing over something that is traditionally the state authority to the Federal Government. Although, signing onto a standard from 1990, if that standard is revised, the state would have to enact the revision into the law, too.
But the fact is, voluntary standards, I don't think that a lot of states are anxious to increase the complexity of the process. Incidentally, one very serious problem with the current standards is that people keep talking about having multiple laboratories around the country certified to test the machines. We have had this standard for a decade now and there is only one laboratory that is certified to do testing under that standard.
Page 57 PREV PAGE TOP OF DOC
I seriously doubt we are going to find much competition in this field, even if we made a mandatory Federal standard. It is simply a very onerous job to test the machines. And machines aren't being introduced quickly enough to make it worthwhile for more than a few—more than one, maybe two laboratories to be involved.
Mr. BARCIA. Well, if I would ask one more question of you. Do state officials, in your opinion, or the FEC have the technical expertise to address the security and accuracy issues you have raised? If not, who do they rely upon should they not have it?
Dr. JONES. Well, it—I can only speak for Iowa. And I have only had indirect contact with people from other states. I believe that by-and-large, the election directors and the state legislature feel that by signing onto the FEC standards they—they don't have to do any more. They view this not as a minimum standard. It is not the minimum acceptable standard. I think this: Most of the states that have signed on view this as the insurance they need, as setting a very high standard. And that the need to legislate anything beyond that standard is viewed with great caution. Because after all, the Federal—the FEC knows what it is doing.
Now my own feeling is that the FEC may not have been bringing in the right people. I think by-and-large, given the fact that all of the new voting systems offered for sale are heavily computerized, it is very evident that the standard has only involved a few people involved with computers. And that the standard has not been updated, as I believe it must be, on a fairly regular basis to reflect our growing experience with the systems that are out there. And do reflect changes in the technology that is being offered.
Page 58 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Let me interject. What would be a fairly regular basis, Dr. Jones? You said, updated on a fairly regular basis.
Dr. JONES. Oh, that is a tough question. It is very expensive——
Chairman BOEHLERT. Give me a tough answer.
Dr. JONES. It is a very expensive thing to do a complete rework of a standard on a regular basis. You can't afford to do that more than once a decade. On the other hand, when problems are—when problem come up, someone has got to notice that there is a pattern in the problems that leads to a need for revision. Right now what I am afraid is, is that people—that when we have incidents involving problems with voting machines, very frequently they are not being reported to anybody. Or if they are reported, they are reported to the state administrator of elections who dumps them in the file marked incidents and they disappear there. Someone should be going over those incident reports probably nationwide in consolidating the big picture of which systems are proving troublesome, what kinds of problems are showing up.
Similarly, someone has got to go out there and look at the numbers for recounts. Because if you look at recounts, that is the best accuracy test we have for any of the ballot counting systems around. When you find errors in a recount—when you recount the votes on a different machine, you get a measure of the accuracy of the machines involved. Those figures should be consolidated. Someone needs to be watching in order to see if problems are being disclosed by this.
Page 59 PREV PAGE TOP OF DOC
Chairman BOEHLERT. You know, it has been suggested in the cloakrooms, if I may use that phrase, that while Florida was under the microscope, far greater problems existed elsewhere, only they weren't as visible.
Dr. JONES. When you have got a vote that comes within 5,000 votes statewide, problems are disclosed. If the margin is bigger, much bigger problems may never be noticed.
Chairman BOEHLERT. Thank you. Mr. Barcia.
Mr. HALL. Mr. Chairman?
Chairman BOEHLERT. Sir?
Mr. HALL. Forty years ago we had an entire precinct vote alphabetically in Terrance County. And nothing was done. We think if they are slow counting the votes down in Miami, and when you get right down to it, Gore didn't lose the election in Miami, he lost it in his home state.
Chairman BOEHLERT. And I am sure you have some interesting stories about Linden's great victory.
Mr. HALL. It wasn't Linden. It was in a governors race.
Page 60 PREV PAGE TOP OF DOC
Chairman BOEHLERT. The Chair recognizes Ms. Morella.
Ms. MORELLA. Thank you, Mr. Chairman, and thank you for holding this hearing. I very much appreciated the testimony from our experts on this panel. You all presented it very succinctly with a lot of information and rather provocative. And again, I am trying to get a handle sort of on this whole standard concept. And much has been mentioned in the last questions that were asked. But, Dr. Jones, you have made many statements in your testimony about the difficulty, the problem with current standards. Do you—and mention was made of the need to revise the FEC standards. One of the questions I would like to ask is, you mentioned that there is a contract with the American Management Systems to come out with recommendations. You seem to imply that this would be something worthy of looking at very closely. Maybe you would like to comment on that.
Dr. JONES. Are you questioning the role of American Management Systems?
Ms. MORELLA. I am wondering about—about what they will be coming out with.
Dr. JONES. Well, it is extremely important that we examine the new FEC standard. We expect the standard to be released in a month for public comment. Part one in a month, part two I think in October. I don't remember the exact dates. And I am extremely interested to examine that standard and see whether—see how that standard would have treated the problems that I have encountered over the past decade. It might be a good standard. It might be what we are looking for. I don't think I am that optimistic in hoping that it patches all the problems. But I suspect it will make a significant improvement.(see footnote 3)
Page 61 PREV PAGE TOP OF DOC
After completing this exercise, I have concluded that the new proposed FEC standard is an improvement over its predecessor, but it is far from perfect. The most glaring problem, and one it admits outright, is that it does not address human-factors issues, and, as was made very clear in Dr. Ansolabehere's testimony, this is an area where we are in desperate need of standards.
The new draft standard does not incorporate the Common Criteria, as Dr. Mercuri has strongly advocated, and it is generally weak in the area of security and auditability.
As I have noted in note 2 above, there is a shortage of expertise in the entire field of secure systems, and this has an effect not only on the staffing problems faced by voting system vendors, but also on the staffing available to standards organizations and the FEC. If we hope to correct this shortage, we must revitalize certain research fields that have been stagnant for most of the last 20 years.
Ms. MORELLA. If the Federal Government were to try to mandate a uniform set of standards, what would you like them to be? And I might ask the rest of you to pick up on that question also.
Dr. JONES. Well, I have certainly mentioned some of the things I would like to see.
Ms. MORELLA. Yeah.
Dr. JONES. I would like to see counting of all those over-votes and under-votes brought forward, all the way through the canvassing process. I would like to see standards governing electronic communication between voting machines. I would like to see standards covering the fault tolerance that is used. Right now the standards suggest that—require duplicate copies of the totals to be maintained, either paper ballot and an electronic count, or two separate electronic counts.
Page 62 PREV PAGE TOP OF DOC
But there is no rule about how to deal with the duplication in the case you find a problem. If you have a direct recording electronic voting machine, there are no reconciliation rules in the event that you find your two sets of ballot images differ. That is a very important area to cover.
I would like to see standards governing the testing of voting machines. I would like to see standards that govern the recounting. A very bad practice right now is that people insist when they do a recount, they will use exactly the same machine to count the ballots on the second time as on the first time. Because they know if they counted on a different machine they will get different totals because the photo sensors on the different optical mark-sense machines are calibrated differently. The fact is, that calibration difference is something that we should be exposing not sweeping under the rug.
Ms. MORELLA. Are we capable of coming up with standards, such as all of you have articulated in different ways, federally? I mean, is this plausible, practical, or is it elusive? Dr. Mercuri?
Dr. MERCURI. We actually do have the standard which is called the Common Criteria. It is actually an international standard and there are various levels. You deploy the level that you need for the types of things that you are trying to certify. It is really a full end-to-end type of standard. Because even if we did test it, let us say in the lab, a voting system, how do you know that that voting system actually is the one that the people purchased or you got 200 copies of that.
Page 63 PREV PAGE TOP OF DOC
Let me just—bear with me for a second. I just want to run down just the types of things that the Common Criteria deals with. System requirements, functionality, correctness, accuracy, accountability, disclosability, reliability, integrity, availability, fault tolerance so that—you know, you would be able to recover in case it has a problem, data requirements, confidentiality. Just a few more. Retention and recountability, user requirements, administrator requirements, interface usability, documentation testing, paths, facility management, recovery, system distribution and compliance with laws and regulations. And these are all in the Common Criteria. And NIST does administer this. I would like to see that direction and that level. It is a 700-page standard. So we do have that. It is at our disposal to use. And I would like to see it be used in a way that directs it toward the——
Ms. MORELLA. So what roles do you see for us, for the Federal Government, for the legislative branch?
Dr. MERCURI. I would like to see some agency charged with the responsibility of actually creating a standard that would be a modified, perhaps, Common Criteria, but that is specific to voting systems.
Ms. MORELLA. Do you see NIST as the laboratory to do that?
Dr. MERCURI. Yes, definitely. Toward that end, I had started doing that myself. But in the recommendation was that NIST would administer it, and that NIST would actually be responsible for it. Because NIST is responsible for administering the Common Criteria right now.
Page 64 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Does the gentlelady yield on that point?
Ms. MORELLA. Yes, indeed.
Chairman BOEHLERT. Is there any disagreement? We—the Committee are looking to NIST. And that is our mindset. Is there anyone that disagrees? Mr. Saltman.
Mr. SALTMAN. My hope is that there will be a unified agency that can undertake the Federal responsibility in election administration. What I mean is that policy direction, determination of what research projects and subjects there should be, what data should be collected, how should the Federal Government assist the states and local governments. All of those issues deserve to be carried out from a unified single agency, which I believe should be bipartisan.
Yes, it is certainly true that technical skills in computers information processing, measurement technology and industrial psychology for use—for ease of use activities need to be part of that process. And in 1993 I wrote an article that has been—it is on the web because it was published by Computer Professionals for Social Responsibility mentioning that—and I said in that article that Congress lacks the scientific assistance that it needs to assist it in election administration.
Now I think that this certainly is an organization that has some of this capability available. Whether it is the only organization that could do this is not clear to me. But the need to have a unified policy direction and research direction and standards direction certainly requires, in my opinion, control from the top in a bipartisan manner.
Page 65 PREV PAGE TOP OF DOC
The problem right now is that NIST is a member of the Executive Branch. And the Executive Branch is headed by a person of a single party who appoints his cabinet officers to be supportive of that party. I don't see how the Executive Branch, in my opinion, would be the proper place for policy direction for election administration because it is a bipartisan issue.
Ms. MORELLA. You would see the FEC also falling into that category, too.
Mr. SALTMAN. It has members of the Commission who are, six right now who supposedly are three of each party. One of the problems has been that the commissioners in the past have represented the issue of campaign finance, which they know very well. Danny McDonald, of course, was a former election administrator. And he is very conversant with the problems. And the other commissioners have made it clear that they understand the issues now, and they are willing to push—put forward the problems of election administration and have submitted a supplemental appropriation.
They all but one attended the recent advisory panel meeting of the Office of Election Administration in Baltimore. And it would appear that they understand the need for improvements in their efforts toward election administration.
Chairman BOEHLERT. Thank you very much, Mr. Saltman.
Ms. MORELLA. Thank you, thank you, Mr. Chairman.
Page 66 PREV PAGE TOP OF DOC
Chairman BOEHLERT. Thank you, Ms. Morella. Mr. Weiner.
Mr. WEINER. Thank you, Mr. Chairman. You know, frequently we are seduced by the capabilities of technology. And we sometimes expect it to do things that perhaps it is not best suited to do. It seems to me in the evolution of the technology we have gotten an increasingly technical, increasingly electronic, and increasingly fallible.
And I would ask the panel, is the move toward increasingly sophisticated technology out of a desire to count faster, tally things more quickly? And is that what has been driving the move of technology? Because I don't think anything is as user-friendly and as impervious to fraud as the older technologies. I mean, the person putting an ''X'' next to a box, a person going into a—those beat-up booths that the Chairman and I use in New York. You know, there are—it is cumbersome and it is clumsy, and the little flipper things fall off frequently, which make it impossible to vote for anyone in that case. But in terms of its fallibility, according to your testimony, those earlier systems are the least fallible. Dr. Jones, you are shaking most vehemently.
Dr. JONES. Well, the MIT-Caltech report pointed out correctly that lever machines have the smallest under-vote and no over-votes on the top of the ticket issues. But it turns out that there is another piece of research, I forget the author. I am sorry. That shows that the—that the fall-off in the vote, that is the participation on tail of the ballot issues, falls off far more rapidly on lever machines than on paper ballots. And this research was done long enough ago that paper ballots are—were the only competing technology studied.
Page 67 PREV PAGE TOP OF DOC
Mr. WEINER. Well, yeah—if I could just interrupt you. But you are addressing a different point than perhaps this Committee is even considering today. I mean, if you want to—if you are trying to encourage participation or you are trying to get involvement, I am sure you would come up with a different design of the polling—of a polling device than you would use if you wanted to reduce fraud. The Internet is an example of that. You would have far greater participation, far greater incidents of fraud. But if you can focus on my point, which is the fallibility of the system, which is the system screwing something up and people's votes not being counted, it seems that we are going—we have two lines. We have a line of fallibility and a line of technicality. And is there anyone on the panel who would disagree with the reason we are going in that direction is to make it more quickly that we can calculate everything. And if not, what would you say is leading us in that direction, except just our lure for things with bells and whistles. Mr. Saltman.
Mr. SALTMAN. I think you raise a very fundamental, philosophical issue that deserves far more discussion than we can give it at this particular moment. However, yes, the move to computers, and even the move to lever machines from paper ballots, was certainly done to improve the speed and supposedly the accuracy of the methods of tallying votes. Now you know the well-known point that of the three, speed, quality and what is the other one? Speed—and cost. You can get any two of the three, but you can't get all three.
Unfortunately, we are in a situation where low cost has been a hallmark of what we do because the Federal Government has not put in a sufficient amount of money to this area. Nor have the states, for that matter. Yes, every system has its advantages and disadvantages. And I—if you would go through my 1988 report, you would see that even lever machines, as I tried to say but it is too technical to point out in short—in a short time, they have their own problems. And transcription errors is not—is one of the major ones. And one of the——
Page 68 PREV PAGE TOP OF DOC
Mr. WEINER. Right. But, you know, being something—and I'm sure Congressman Boehlert is as well, being something of an expert on lever ballots. Even transcription errors, there is a way you go back and check. You know, the poll watchers who reported on election night sometimes flip numbers. But unlike an electronic machine where at a certain point the bits can be compromised. I mean, the technology is not getting us closer to the infallible standard. And I would say, and perhaps this is heresy in the Committee on Science to say such a thing, that perhaps the more ancient technologies, the more basic pen-to-paper, the lever that gets a click and can't be clicked twice, perhaps this is the kind of things, as ironic as it might be, should be a future standard.
Because, frankly, it is—if our concern is spoiled and unmarked ballots, if our concern is over-votes, New York, which has plenty of problems, we don't have that problem. Now sometimes we have machines that disappear on the way to the warehouse, machines that show up with my name not on them. That is a whole other thing. But I think if we are concerned about the fallibility of the system, it is interesting to note the more bells, the more whistles, the more computer chips, the more zero-to-1 digital codes, the more mistakes that are getting made. Thank you, Mr. Chairman.
Chairman BOEHLERT. Thank you very much. Mr. Johnson.
Mr. JOHNSON. I think you probably all each collectively or individually addressed this issue. What—looking, you know, three or four decades down the line, which I suppose is beyond any of our likely tenures or your testimony, to a nirvana, what would you like to see—either each of you individually or collectively—in a system that is not only a system that is workable and fair, but something that in light of technology and economics over the course of these next few decades is probably feasible as well? I know that is almost a Pollyanna question, but I would be interested to know your response.
Page 69 PREV PAGE TOP OF DOC
Dr. ANSOLABEHERE. Did you ever talk to a computer organization and they—one of the questions was just along these lines. And one of the answers was, well, next we are going to go to all electronics and Internet voting. And after that fails, we are going to go back to paper and pens or maybe chads. So it is outside of my realm. That is way too far in the future. Especially as a social scientist, I am used to looking backwards rather than forwards. So I will pass that up.
Dr. MERCURI. I really want to see the human aspect maintained in the election process. The whole system that we have of poll workers and individuals, the human factors of stepping up to the poll where somebody from your neighborhood remembers that you moved out six months ago and really shouldn't be there voting anymore. That type of, you know, bio-identification, it is right there and we have it at the precinct. It is not perfect, but I would like to see that maintained. The fact that we need to have some way of actually manually doing the recounts. And in the future not to abandon that, that humans always be involved with that at a very deep level.
Mr. SALTMAN. If I might add something. I think, yes, the human factor is extremely important. And despite the fact that all electronic systems are becoming wide-spread, and I have recommended considerable amounts of protections for them, optical scan balance allow for personal view of what you have accomplished. And you can see, look at the card before you drop it in the ballot box. It does allow—precinct voting I think is a significant social—of particular social importance. People who have—many people have—thoughtful people have talked about the fact that Internet, remote Internet voting would destroy this public situation where we come and meet our neighbors at the precinct. I think that is a very important issue. I have—the importance of the human factor, the management factor about election administrators, the need to do research undertake and probably by psychologists. Not by information technologists and computer scientists about what is best for the voter in terms of the type of system to be used. All of those factors mitigates the use of totally electronic systems. Although they are coming in, and if the protec