[House Hearing, 110 Congress]
[From the U.S. Government Publishing Office]



 
                           ELECTION REFORM: 
                           AUDITING THE VOTE

=======================================================================

                                HEARING

                               before the

                       SUBCOMMITTEE ON ELECTIONS

                   COMMITTEE ON HOUSE ADMINISTRATION
                        HOUSE OF REPRESENTATIVES

                       ONE HUNDRED TENTH CONGRESS

                             FIRST SESSION

                               ----------                              

             MEETING HELD IN WASHINGTON, DC, MARCH 20, 2007

                               ----------                              

      Printed for the use of the Committee on House Administration


                       Available on the Internet:
   http://www.gpoaccess.gov/congress/house/administration/index.html




                     U.S. GOVERNMENT PRINTING OFFICE

35-804 PDF                 WASHINGTON DC:  2006
---------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Printing
Office  Internet: bookstore.gpo.gov Phone: toll free (866)512-1800
DC area (202)512-1800  Fax: (202) 512-2250 Mail Stop SSOP, 
Washington, DC 20402-0001



                   COMMITTEE ON HOUSE ADMINISTRATION

           JUANITA MILLENDER-McDONALD, California, Chairwoman
ROBERT A. BRADY, Pennsylvania        VERNON J. EHLERS, Michigan
ZOE LOFGREN, California                Ranking Minority Member
MICHAEL E. CAPUANO, Massachusetts    DANIEL E. LUNGREN, California
CHARLES A. GONZALEZ, Texas           KEVIN McCARTHY, California
SUSAN A. DAVIS, California
                                 ------                                

                       Subcommittee on Elections

                  ZOE LOFGREN, California, Chairwoman
JUANITA MILLENDER-McDONALD,          KEVIN McCARTHY, California
    California                       VERNON J. EHLERS, Michigan
CHARLES A. GONZALEZ, Texas
SUSAN A. DAVIS, California


                           AUDITING THE VOTE

                              ----------                              


                        TUESDAY, MARCH 20, 2007

                  House of Representatives,
                         Subcommittee on Elections,
                         Committee on House Administration,
                                                    Washington, DC.
    The subcommittee met, pursuant to call, at 2:50 p.m., in 
room 1309, Longworth House Office Building, Hon. Zoe Lofgren 
(chairwoman of the subcommittee) presiding.
    Present: Representatives Lofgren, Gonzales, Davis of 
California, McCarthy, and Ehlers.
    Also Present: Representative Holt.
    Staff Present: Tom Hicks, Election Counsel; Janelle Hu, 
Professional Staff Member; Matt Pinkus, Professional Staff/
Parliamentarian; Kristin McCowan, Chief Legislative Clerk; 
Gineen Beach, Minority Counsel; and Peter Sloan, Minority 
Professional Staff.
    Ms. Lofgren. Welcome to everyone. First, apologies for our 
tardiness, for some reason as soon as there is a Election 
Subcommittee notice, the House calls at least 5 votes and makes 
us late. And so obviously, we have to go over and do our duty 
of voting. But it does put us in a delay mode and we do 
apologize.
    Last week, we had a hearing on election reform and learned 
about some of the issues surrounding the tools of voting 
machines and software as well as access to those machines for 
all voters.
    This week, we shift focus to what happens when the polls 
close, auditing. The Help America Vote Act, or HAVA, addressed 
machines and established new procedures for voting lists and 
registration, but auditing was not addressed. Post election 
audits are a tool to increase voter confidence in the election 
process. But there is no national standard for auditing in 
Federal elections. There are States, such as California, where 
the paper count takes precedence over the electronic one, or 
other States like Nevada are the reverse.
    Discrepancies are not just about paper or electronic 
counts. It also goes to the extent of audits. Connecticut has 
performed post election audits using 20 percent of precincts. 
Other States are as low as 1 percent of precincts. In addition 
to the extent of audits, they need to be conducted so they are 
publicly observable. And audits must, as we know, also be 
entirely random to avoid the possibility of tampering.
    Greater transparency through increased election scrutiny is 
not a bad thing, and having a voter verified paper trail with 
an automatic routine audit may go a long way to increase voter 
confidence and to deter fraud.
    It is no secret that a number of recent elections have been 
determined by a very small difference of votes. And a handful 
of these are still in dispute. A failure to have paper records 
that can be audited could ultimately call into question the 
validity of a very close election. And it may be that an 
established national process for the audit of Federal elections 
would be worthwhile.
    I would like to call on the ranking member, Mr. McCarthy, 
for his opening remarks, and also invite the other Members to 
submit statements for the record. Mr. McCarthy.
    [The statement of Ms. Lofgren follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.300
    
    Mr. McCarthy. Well, I appreciate the Chairwoman for calling 
this hearing on election audits. I do believe that all Members 
of this panel as well as our colleagues in the House are 
dedicated to ensuring that Federal elections are conducted and 
tabulated in a free and fair manner. I do want to make sure, 
though, that the Federal interest in this issue should not 
cloud the work of the State and local jurisdictions and I am 
happy that we have two States here, Ohio and Arizona, that 
tabulate and do an audit as well, so we will hear from them on 
the ability that we can work together and make sure we have 
fair and honest elections. Thank you.
    [The statement of Mr. McCarthy follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.301
    
    Ms. Lofgren. Thank you. I would just like to note for the 
record that there was some confusion on witness notification 
and submission of testimony, and we are going to strive at the 
next hearing to have all our testimony submitted within the 48 
hours that is specified in the rules and make sure that 
everything is shared as promptly as possible between the sides.
    I would like now to call upon our first panel, and, we have 
Mr. Ion, is that how it is pronounced? Sancho, who is the 
supervisor of elections from Leon County Florida. He was 
elected to his first term as supervisor of the elections in 
November of 1988, re-elected in 92, 96, 2000, 2004 and serving 
his fifth term in--as of January second 2005. He has a JD from 
Florida State University Law School, a bachelor's degree from 
Stetson University.
    And we also have Mr. Matt Damschroder, who is the director 
of the Franklin County Board of Elections. And Mr. Damschroder 
has been the Director of the Franklin County Board of Elections 
since 2003 and serves as the president of the Ohio Association 
of Election Officials. And he was the executive director of the 
Franklin County Republican Party until his appointment to the 
Board of Elections.

STATEMENTS OF ION SANCHO, SUPERVISOR OF ELECTIONS, LEON COUNTY, 
    FLORIDA; AND MATT DAMSCHRODER, FRANKLIN COUNTY BOARD OF 
                           ELECTIONS

    Ms. Lofgren. We welcome both of you and we have a light 
system here. We will ask that you summarize your written 
statements which we will have and submit for the record and try 
and give your oral remarks in about 5 minutes that will allow 
for questions. So if we may ask, Mr. Sancho first and then the 
other witness.

                    STATEMENT OF ION SANCHO

    Mr. Sancho. Thank you, Chairwoman Lofgren, Ranking Member 
McCarthy and committee members. Thank you for the opportunity 
to testify today. My name is Ion Sancho. I am the elected 
supervisor for elections for Leon County, Florida. I am a 
member of the International Association of Clerks, recorders, 
election officials and treasurers, and the election officers, 
the two largest professional associations for election 
officials in the country. I have no party affiliation.
    In my testimony today, I will focus the problems Florida 
has encountered over the past 6 years and how our audits--or 
more accurately, the lack of audit--have contributed to the 
current crisis in confidence Floridians have in their electoral 
system today, a crisis of confidence that is so critical that 
the incoming governor, upon taking office, stated that in the 
State of Florida, there will be--there will be a paper trail 
for every vote.
    And that is a sea change from the previous administration.
    What are audits? One dictionary definition refers to an 
audit as an official examination and verification of a accounts 
and records. Another calls it a methodical examination and 
review. I would like to say that the reason that I had to 
resort to the dictionary is that audits have no definition 
under Florida law. Because Florida does not conduct audits. 
Before the election, we have to do a series of tests, which we 
call logic and accuracy, and after the election, we do a logic 
and accuracy that ensures that the machine can count a test 
step. But we never audit actual ballots or votes. And so as an 
election official, I had to resort to the dictionary to 
describe an audit.
    Which leads me to directly to the 2000 elections. In 
Florida, audits for any election, again, as I have mentioned, 
are not required.
    The closest thing that we have had in Florida election law 
to an audit was our pre-2000 recount provisions in chapter 102, 
which, depending on the closeness of the contest, could mean 
that every ballot had to be manually examined.
    These recounts are generally rare events. In my almost 20-
year career, I have overseen four of them, and only one of 
these, the presidential race of 2000, involved a Federal race, 
and that recount, the only audit that we could use was 
terminated by the United States Supreme Court. All of us in 
Florida were embarrassed by what happened, including election 
officials. And this embarrassment, which the Miami Herald 
recorded at 70 percent dissatisfaction with our electorate, 
forced our public policy makers to act.
    The governor ordered a bipartisan task force, which held 
hearings across the State and produced 35 excellent 
recommendations, including audits. But audits, recounts, voter 
intent, were all discarded out of that task force series of 
recommendations and substituted in those for those audits and 
recounts was the assumption that since voting machines could 
never make a mistake, it was illogical then to audit any result 
that was produced by voting machines. And this is the current 
State of Florida law today.
    Let me tell you that if we were going to have the same kind 
of statistical dead heat in Florida in 2008, we would have less 
than 1 percent of the ballots examined under current law. Why? 
Because under Florida law, the only ballots which can be 
examined in an audit are those ballots which could not be read 
by voting machine.
    That means every machine readable ballot is out of bounds 
in a recount to examine.
    In the last election, only 500 ballots out of over 92,000 
ballots cast would be examined under a Florida recount, roughly 
less than 1 percent of the ballots. Again because only these 
number of ballots, the overvotes and undervotes and optical 
scan ballots, either by those jurisdictions who use paper 
ballots in the precincts or those ballots received by the 
election official through the mail which are on paper, only 
those ballots which are nonmachine readable could be manually 
audited.
    Ms. Lofgren. Mr. Sancho, the red light means your time is 
up, so could we ask you to wind up and we will go to our next 
witness?
    Mr. Sancho. All right. I would urge this body to require 
that mandatory random and manual post election audits be made 
the law. Without them, no citizen, State or Federal in terms of 
the races that they focus on, can actually have any validity 
under Florida's current statutory scheme that is, in fact, 
their votes are being counted. Thank you.
    Ms. Lofgren. Thank you very much for that testimony. And as 
I mentioned earlier, your complete statement will be made a 
part of the record as well as the questions that follow.
    Mr. Damschroder, now it is your turn, and I probably should 
have said when the yellow light goes on that means you have a 
little time left to summarize.

                 STATEMENT OF MATT DAMSCHRODER

    Mr. Damschroder. Good afternoon, Madam Chairwoman and 
members of the committee. Franklin County is the most populous 
county of Ohio's 15th congressional district. And during the 
2006 congressional election cycle, the 15th congressional 
district election resulted in an official margin of victory of 
less than one half of one percent or less than 1,055 votes. 
That general election was administered in Franklin County using 
direct recording electronic voting machines, or DREs, with a 
voter verified paper audit trail.
    In 2004, the Ohio General Assembly had enacted legislation 
requiring the VVPAT to be included in all DRE's beginning in 
2006. I was the only election official in Ohio who testified in 
support of VVPAT technology. But I had two concerns with Ohio's 
legislation. And they are the two concerns that I have with 
H.R. 811.
    First, the VVPAT should not be, in my opinion, the ballot 
of record. With the VVPAT as the official ballot, it is 
possible that an otherwise properly cast and accurately 
recorded ballot on the DRE might not be recounted in a close 
election due to paper jams or poll worker error loading the 
paper backwards. The question what constitutes a vote, having 
long been determined in Ohio for punch cards and optically 
scanned paper ballots, has seemed wrong to me at the time and 
wrong to me now that we would institute a new voter intent 
question that could cause a voter's properly cast and 
accurately recorded vote to go uncounted.
    Second, the VVPAT requirement should go into effect only 
after the Election Assistance Commission had developed 
standards for the function and operation of the technology.
    I am concerned that by enacting portions of this 
legislation, Congress might be making precisely the same 
mistake that Ohio made by mandating changes to the technology 
by date certain before operational standards are established 
and the technology is fully developed and tested.
    Essentially in Ohio and in Franklin County specifically, 
our $14 million investment partially included HAVA funds and 
partially including local taxpayer funds would likely have to 
be dumped because our VVPAT technology does not conform to the 
requirements of this legislation.
    Beginning with our first use of the VVPAT in a special 
election of February of 2006, we began voluntarily conducting 
audits of the VVPAT to the electronic record.
    As a part of that, we determined that there were instances 
where the paper printers did jam or that poll workers did make 
errors in loading the paper tapes backwards. So as we prepared 
for the recounts, the mandatory recount in the 15th 
congressional district election, we knew that there were going 
to be problems and with no definition in Ohio law from the 
Secretary of State in answering that very important voter 
intent question on the VVPAT, the Board of Elections engaged 
both of the campaigns to basically create ground rules for the 
conduct of the recount of how we would treat a ballot in the 
event that we came to a jam or a blank tape.
    The hand tally of the VVPAT for this Federal recount took 
2,000 employee hours. And again, that was to hand tally 10 
percent of the total votes in Franklin County.
    I would like to bring one specific instance to the 
committee's attention that demonstrates the importance of clear 
and objective guidelines to define voter intent on the VVPAT. 
On one particular paper tape, the recount team encountered a 
paper jam and there was no indication on the paper tape that 
the voter's voting session had ended in a ballot properly being 
cast. At the conclusion of the recount of that tape, the total 
number of hand tallied votes excluding the so-called jammed 
vote on the paper tape equaled the total number of votes cast 
on the machine. To further verify that the so-called jammed 
vote should not be counted, the recount team examined the 
ballot image log and the electronic event log, both of which 
validated that there were only the specified number of ballots 
cast in the machine. And the event log actually showed that at 
the time displayed on the paper tape that the printer had 
jammed.
    Additionally, the poll workers followed their 
responsibilities and noted in their records that a paper tape 
had jammed, that they cancelled the ballot and allowed the 
voter to vote on a different machine.
    When using VVPAT technology which introduces new question 
of voter intent as I described, vague and subjective language, 
such as 811's phrase, preponderance of evidence or clear and 
convincing evidence is an open invitation to litigation. In 
Franklin County, that could have meant that litigation would 
have required weeks beyond the end of the recount and that 
possibly a representative from the 15th congressional district 
not being seated when Congress convened, also look towards 
2008, it could possibly mean that our recounts and other audits 
would not be concluded by the time the Federal electors are 
seated.
    Our experience in this recount demonstrates the accuracy of 
electronic voting systems and the benefit of State and local 
control over election audit recount definitions and procedures.
    [The statement of Mr. Damschroder follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.001
    
    [GRAPHIC] [TIFF OMITTED] 35804A.002
    
    [GRAPHIC] [TIFF OMITTED] 35804A.003
    
    [GRAPHIC] [TIFF OMITTED] 35804A.004
    
    Ms. Lofgren. Thank you very much and congratulations on 
having a precisely 5-minute statement. I would like to begin 
the questions, if I may, and I will start with Mr. Sancho.
    You have been an election official for a long time and have 
had substantial experience. In your experience, do you have an 
opinion on what voting system or machine makes conducting an 
audit easier, more transparent and accurate?
    Mr. Sancho. Yes, I do. I believe that a document--right now 
the only thing we have is paper, marked by the voter's own 
mark, in their own handwriting, is the kind of evidence that 
would allow me to do a complete audit.
    I got into the area of elections in 1988 because in 1986 I 
ran for a local government position, and our supervisor of 
elections negligently programmed the voting machines so that 
they jammed on Election Day--through, it was not nefarious it 
was just negligence. But it meant that upwards of 5,000 Leon 
County voters could not vote on that Election Day. And when I 
ran for and got that position 2 years later, the number one 
thing on my agenda was to create a voting system and a process 
that would allow the voters to know what the votes were, even 
if I negligently programmed the voting machines.
    So I wanted a system that would have a backup to the 
technology that I could confirm the validity of the votes. And 
that led me to optical scan voting technology, which is 
emerging in the late 1980s and early 1990s. And it was a 
technology that would allow every voter to mark that paper 
ballot and then I could scan them and get the result to the 
media quickly and to the candidates. But if there were any 
disputes over my handling of the ballots or of the election 
workers handling of the technology, the evidence resided in the 
form of optical scanned ballots which no one need use 
technology to get a 100 percent accurate count. And that is 
what we selected on in 1992. And the results are clear. In the 
2000 botched election, Leon County had a residual or error rate 
of less than 1 fifth of 1 percent.
    My neighboring county had an error rate of 12.4 percent, 
which means that only 86 percent of the citizens that tried to 
vote in the presidential race actually even ended up having 
their vote properly recorded. That--all of us that used the 
optical scan technology in Florida had an error rate of less 
than 1 percent.
    And we were able to recount our ballots without a problem. 
The punch card and the central count systems failed the 
citizens and they were banned.
    Ms. Lofgren. The governor of your State has recently made--
is he--is the State of Florida moving toward that system?
    Mr. Sancho. We are making steps in a very, very good 
direction. House Bill 2166, which was introduced by Senator 
Villalobos, at the behest of the governor would replace every 
touch screen voting machine at the precinct on Election Day 
with an optical scanner. And that is a very excellent first 
move.
    Ms. Lofgren. Let me ask you, you have--correct me if this 
is wrong--but my understanding is that you did some testing to 
see whether voting machines could be hacked or not hacked. 
Could you describe or report to us on your findings in that 
regard?
    Mr. Sancho. Yes. After the 2004 election, I was approached 
by a British journalist who had been observing elections across 
the country, and I guess I struck him as sort of an honest type 
of guy. And he said, I think we can hack your voting system. 
And I said I don't think so. But he said, would you be open to 
having an independent test to see whether or not your system 
would be compromised? And I thought about it for a while. We 
had some negotiations. And I answered yes.
    And what we found over the course of six tests that took 11 
months to conduct, was in fact a huge, huge security flaw in 
the process that would allow anyone with insider access to 
actually prestuff the electronic ballot box and have those 
numbers alter the outcome of an election. And it would not be 
detected in the normal canvass process.
    This was shocking to me. And quite frankly, it confirms my 
own sense that even when you are using a paper-based system, 
unless you actually audit it, it can be vulnerable to a whole 
host of security attacks which the systems administrator would 
not even be aware is occurring.
    Ms. Lofgren. I am going to turn it over to my ranking 
member, Mr. McCarthy.
    Mr. McCarthy. Thank you, Madam Chair.
    Question for Franklin. Ohio has automatically, if you are 
in the half percent to 1 percent, you do the audit.
    Mr. Damschroder. That's correct.
    Mr. McCarthy. Now you have raised some good points here 
because you have gone through a little legislation that we may 
bring up later with the VVPAT. You said in your statement that 
the paper should not be the ballot of record.
    Mr. Damschroder. That is correct.
    Mr. McCarthy. Could you elaborate on that a little more and 
why?
    Mr. Damschroder. My belief is that the voter verified paper 
audit trail should not be the ballot of record, and the reason 
I believe that is when you compare an electronic touch screen 
system to an optical scan system, with the optical scan as Mr. 
Sancho mentioned, the voter as marking the ballots themselves, 
the electronic system in Franklin, Ohio with the voter verified 
paper audit trail the voter is marking the ballot 
electronically, and they are verifying that paper record.
    In situations with a recount of the--in Ohio of an optical 
scan system, you hand tally a small universe. The minimum 
requirement is 3 percent of the optical scan paper ballots. And 
then you run them through the tabulator and compare the hand 
tally to the electronic tabulator result. If they match, then 
you rerun all of the ballots electronically. Basically, the 
hand audit of the paper is to prove the veracity of the 
electronic.
    To me, I believe that it is not recommended to create a 
higher standard for an alternate voting system so that we 
should use the voter verified paper audit trail as it has been 
named in Ohio, and as it is named, I believe, in this 
legislation for that precise purpose to audit the electronic 
results. When the paper ballot becomes----
    Mr. McCarthy. So you are saying the paper should not be the 
final decision maker?
    Mr. Damschroder. Should not be, because as I shared in my 
testimony, in some instances, the paper jams----
    Mr. McCarthy. You said that paper jammed and it shows that 
it jams and the people did what was right even though you 
trained them, you gave them enough training and computer showed 
that it did jam?
    Mr. Damschroder. Correct.
    Mr. McCarthy. So did that ballot get counted properly and 
it wasn't double-counted?
    Mr. Damschroder. It was not double-counted because when the 
paper jammed, the machines are designed, such as the machine 
will not allow the voter to continue casting ballot when the 
paper is jammed or is not present. So when the, in this 
instance, when the printer jammed, the voter noticed that the 
machine stopped allowing him to vote or her to vote, notified a 
poll worker. The poll worker cancelled the ballot on that 
machine and took the voter to a different machine where she 
could cast her ballot and the jammed machine was not used again 
until a technician could come and unjam the paper.
    Mr. McCarthy. Now you said it took 2,000 hours to do this 
audit or recount.
    Mr. Damschroder. Correct.
    Mr. McCarthy. And you believe audits are good.
    Mr. Damschroder. Yes, I do. And that is why we did three 
audits before the 2006 general voluntarily.
    Mr. McCarthy. How many races in Ohio come within 3 percent?
    Mr. Damschroder. I don't know for sure. In Franklin County, 
it is very few. This congressional district was very unique in 
the political balance of the jurisdiction, also the 
characteristics of the contest.
    Mr. McCarthy. Would you audit every race?
    Mr. Damschroder. I would not. In the situation of the 
previous audits that we did, we randomly selected a handful in 
two of the contests that were a single issue. In the primary, 
we randomly selected, I think, about a dozen contests of the 
probably 50 that were on the countywide ballot and tallied 
those from 3 percent of the machines, so I don't believe it is 
necessary to tally the entire universe. I am not a 
statistician, but all the statisticians that I have spoken with 
lead me to believe that with that form of mathematics, you can 
test a small universe to learn with virtual certainty the 
characteristics of the entire universe.
    Mr. McCarthy. Fortunately, we have a person that knows math 
very well on the committee that could probably answer that for 
us. But let me ask one question to the individual from Florida. 
Would you audit every race?
    Mr. Sancho. No, I would randomly select races, but I think 
that you with particularly the focus here at Federal races, I 
think you have to look at the Federal races that has to be 
included one of the Federal races.
    Mr. McCarthy. Would you pick it based upon the difference 
in the race? Or would you pick every Federal race? Would you 
say if a race came between 3 percentage points, 5 percentage 
points, 1 percentage point, 10 percentage points?
    Mr. Sancho. Well, in Florida, we basically have no more 
than three Federal races on any particular general election 
ballot. So if you wanted to do all the feds, you could do that. 
The 2,000-hour time frame that was described by Matt is 
essentially a series of 4, I believe it is 5, 4-person teams 
Matt?
    Mr. Damschroder. Mmh-hmm.
    Mr. Sancho. And how long did that take you to do?
    Mr. Damschroder. Seven days.
    Mr. Sancho. Seven days, 5 four-person teams, the number 
2,000 sounds large, but an audit that takes 7 days that 
examines 10 percent of the vote in the largest jurisdiction in 
Ohio, actually I commend him on doing it within 7 days.
    Mr. McCarthy. So could you, timewise, if you were required 
to do every Federal race, could you do it?
    Mr. Sancho. Absolutely in three races, random selection--
and not--I would like to say I think you have to do a--my own 
analysis of the statistics involved, if the race is very close, 
you would not want to use the same number to statistically 
sample a race in which the margin is 25 percent apart. I think 
a tiered system, the closer the race is, the larger your sample 
is going to have to be if you are going to get the level of 
confidence--and in my mind, you are talking about 99 percent 
plus confidence rate--in order to establish that what you do 
has meaning.
    Because that is really what we are trying to accomplish 
here. We are not trying to just do an audit so that we do an 
audit and we have done an audit. We are trying to do an audit 
so that the individuals involved in the race, their supporters 
and everyone concerned can know with confidence that result is 
the result.
    Ms. Lofgren. I would like to call Judge Gonzalez, our 
colleague from Texas.
    Mr. Gonzalez. Thank you, very much, Madam Chairwoman. And 
the question will go to--is it Damschroder?
    Mr. Damschroder. Damschroder. Correct.
    Mr. Gonzalez. I am new to the committee and appreciate this 
opportunity. It is a education for all of us, the general 
public out there many times will be confusing audit with 
recount and the manner in which you would conduct that. I guess 
my question really goes--and I will be asking Mr. Sancho a 
question in a minute--but regardless, it appears that you are 
still depending on a paper trail to serve a very important 
function even in the auditing process, because by your 
testimony a minute ago, it would be the paper ballot or the 
printout from the electronic recording of the vote that would 
basically measure some percentage of the machines, then you 
would feel that you would go forward after meeting that 
particular threshold. Is that correct?
    Mr. Damschroder. Yes. And then there is an important, I 
think, distinction, I mean, a recount as we conducted it in the 
case of the 15th congressional district election last year was 
statutorily required function to come up with to basically 
essentially re-examine every ballot that was cast, basically 
recanvass in the election.
    For us in the previous voluntary audits, those were not 
mandated by law and didn't have the force of a canvass or a 
vote-certifying function. It was an examination of the paper 
records to determine the veracity and validity of the 
electronic results, which in the case of the 15th congressional 
district we found that not one vote changed as a result of the 
examination of the paper trail to the electronic. The only 
votes that changed were during the hand tally of and the 
recount of the optically scanned paper ballots.
    Mr. Gonzalez. Yes, sir. Absent a paper trail, how would you 
conduct an audit? In other words, it lends an important tool. 
The paper trail is--I guess what I am really getting at, even 
if you look in Texas in my county, Barrett County, we don't 
have a paper trail. So if you want to conduct an audit, you are 
going to have to do it absent that threshold test that you have 
already described so----
    Mr. Damschroder. And in Franklin County, we had used 
paperless DREs since 1992. In the case of the 15th 
congressional district, both of the candidates had been elected 
and reelected multiple times for different offices using that 
paperless system.
    That is, frankly, why I believe in the voter verified paper 
audit trail, and supported it in Ohio in 2004 was that I 
believed that it was important to have a permanent independent 
record from the electronics who audit the electronic--the 
electronic record. In the past, recounts have been conducted 
simply by rereading the electronic results from the electronic 
cartridges from the voting machines. So by having this 
permanent paper record that was voter verified independent of 
the electronic record, we have something to audit the 
electronic record with.
    Mr. Gonzalez. The other question I think you had posed at 
the outset of your testimony was something what constitutes a 
vote, and I am just thinking in terms of I go in there, I vote, 
I hit the key, whatever, it is registering internally, I am 
going to be able to read it, but it is also printed out but I 
am going to need both components in order to constitute the 
vote. Is that correct? The actual recording by the machine 
itself internally, electronically but also then the manual 
printout, the paper trail.
    Mr. Damschroder. Under Ohio law, and under, based on my 
understanding of this legislation, the paper is the official 
ballot of record for purposes of the recount. It is my view 
that the paper record should be used to audit the electronic 
and that the voter, you know, has cast their ballot 
electronically, they have saw it in the electronic, but they 
saw it on the paper, but as I said, it does introduce a new 
voter intent question that is wildly open to litigation by 
simply saying the paper is the ballot of record, and giving 
election officials no guidance on whether to count what appears 
on the paper as a jam or what does not appear in the event of a 
blank tape. Those issues--if the paper is going to be used as a 
ballot of record, States must give local election officials 
guidance on what constitutes a vote. It is the same issue we 
ran into in Florida and other jurisdictions with punch cards 
same issue a lot of jurisdictions wrestle with in optical scan.
    Mr. Gonzalez. My time is just about up.
    Mr. Sancho, quickly, you said your preference would be the 
optical scanner where someone actually marks the ballot and 
then it is read electronically and such but you still have 
that, is that correct?
    Mr. Sancho. Yes, it is, and I really think that right now 
it is the most auditable kind of election. I sort of come back 
at this from the way of audits. I am--the reason that voter 
verification even began was the introduction of the touch 
screen voting machines which were not transparent and auditable 
under any circumstances to begin with.
    So the idea was to produce a piece of paper that the voters 
could look at to give them the level of confidence that the 
machine was, in fact, recording that.
    Now the issue is, I like the direction that the Holt bill 
is discussing in saying that that piece of paper has to be 
durable. One of the reasons that election officials all across 
the country are just not very happy with the current technology 
is that the thermal printers tear and jam. Whether you look at 
the report emerging out of Cuyahoga County in May, where 10 
percent, or even leaving that jurisdiction aside and going to 
Charlotte--excuse me, Mecklenburg County in North Carolina, 
where on Election Day, I have--I received their copy of all the 
Election Day problems. Over 58 percent of all of the reported 
election problems were on that printer.
    The Holt bill says that if you are going to use a paper 
that comes out of a DRE, that has to be durable, it can't be 
the these toilet paper thermal rolls. In other words, what they 
have done is they have established what needs to be audited and 
left it to the industry to do that. And I will tell that you in 
the Senate hearing that I was at this week, in Florida, I was 
approached by my vendor, who told me they have new equipment 
under development designed to meet that Holt standard, and I 
was very, very pleased because these thermal papers are awful.
    Ms. Lofgren. Thank you. And the chairman--ranking member of 
the full committee, Mr. Ehlers.
    Mr. Ehlers. Thank you Madam Chair and, I thank you for 
holding this hearing.
    I just want to mention, first of all, that we held two 
hearings on this topic last year, and I hope that will continue 
to be part of the record as we review the whole issue.
    Ms. Lofgren. Certainly.
    Mr. Ehlers. First of all, Mr. Sancho, I am very pleased to 
meet you because I am a physicist, and I have been working on 
ions most of my life. It is really a pleasure to meet one face 
to face finally.
    Mr. Sancho. And I am a charged particle.
    Mr. Ehlers. You look like it.
    I have been in this voting business since the early 1970s, 
and initially not as a candidate, but as working with other 
elections. I have watched with interest the attitudes 
developing about the bill that Mr. Holt has offered.
    I am not opposed to the bill in principle. I am opposed to 
some of the specifics in the bill. I think it is important that 
we have some method of auditing, as both of you said.
    Let me first say precisely what my standard is, to start 
with, I believe that every voter should have complete assurance 
that the vote they cast will be counted accurately and 
completely. That is number one.
    There is a second part here, which I haven't heard 
mentioned at all, and that is the voter also has to have the 
assurance that their vote is not diluted by others casting 
fraudulent votes or some manipulation of the votes which 
introduces fraud. Now, I haven't heard you mention fraud at 
all. To me, that is a very important factor. When you look back 
at the history of it, originally everyone had paper ballots. 
But the reason that voting machines were developed and put in 
metropolitan areas is to avoid fraud because there was so much 
shuffling of the paper. Once you have paper ballots, people can 
hide them, destroy some, ballot boxes can mysteriously appear 
that were overlooked earlier, questionable marks, et cetera. So 
I think it is very important to look at both aspects of that as 
we consider this bill and the problems that arise.
    The other problem is, you both mention optical scan 
ballots, which are used in my home community as well. Again, a 
paper trail, I am adamantly opposed that the Congress would 
require that you absolutely have to take one of the options and 
say that is the official record. I think it should be left to 
local and State election officials to decide on site which 
record is more accurate more reliable when they make their 
decisions.
    A good example is during the last presidential election, 
optical scanned ballots were used in Los Angeles County. There 
were 30--over 3,600 ballots cast for President in which voters 
blackened off the ovals for 8 of the 9 candidates and did not 
blacken that one for President Bush. Now, how do you interpret 
that? Does this mean that they were blackening out all the 
others because they wanted to get rid of them but they really 
wanted to vote for President Bush? Or were they saying, anyone 
but Bush. In other words, I will vote for eight people, but I 
don't want to vote for President Bush.
    So you can have errors with the paper ballots, too. I don't 
know how the machine would read them, but the advantage of an 
electronic machine is that it spots the error immediately and 
tells the voter, you cannot do that. So we have to recognize, 
it is one of the big advantages of electronic machines.
    So I think it would be a bad error to say, well, it is 
paper and paper only that counts.
    I think it is extremely important that we set good 
standards. One or both of you mentioned that. If we are going 
to go to that system I think we have to take the time for the 
National Institute of Standards and Technology to once again 
set the standards and make certain that the firms that check 
these machines are doing it accurately. I think that is a weak 
link in what we did the last time. We did not empower the 
standards community to watch the companies that are testing the 
equipment. So that is something that should be included as 
well.
    So with that, I don't have any questions. I really 
appreciate your experience. Much of what you say fits together. 
Some of you have different views on some things but the basic 
idea is to use an electronic machine, a properly designed one 
for voting and recording the vote. Use some alternative 
verification. I would like to talk about 2 reproducible paths. 
It could be 2 computers frankly. I know computers. I could 
easily design a system where you have one computer that people 
vote on.
    You have an alternate computer that looks at the same key 
strokes, analyzes them, records them in a separate fashion, and 
then you can electronically compare the two. Or you can have a 
computer and a paper ballot, whether it is optical scan or a 
tape or whatever. But absolutely I am unalterably opposed to us 
in the Congress saying we know what is best for you people, and 
you have to use the paper trail as a final one, or you have to 
use the computer. I think the local officials, who have a much 
better wisdom than the Congress in deciding which one it should 
be. So I have no questions. Thank you very much.
    Ms. Lofgren. Thank you very much and, we have my colleague 
from California, Mrs. Davis from San Diego.
    Mrs. Davis. I wonder if we can talk about a few of the 
basics, because I don't quite know this. You mentioned the 
number of hours that were required, just cost generally, how do 
we look at this? This shouldn't be the biggest factor involved 
at all, but how expensive is it actually to audit and does that 
end up playing a role in terms of what counties do sometimes?
    Mr. Damschroder. Well, in Franklin County in the recount 
that we did to get just 10 percent of the votes from one 
Federal contest out of, if you look at 2008, we will have 5 
Federal contests on our ballot, it took us 2,000 person-hours 
over the course of several days to do just that one Federal 
contest.
    If we were to do all five--and again, depending on the 
percentage that we are auditing to, it would obviously take far 
more than that. The costs for us to prepare and to conduct 
that--that recount just in terms of overtime alone was $50,000 
just for that time frame.
    So I view it as a significant cost. If the Federal 
Government is to going to mandate that local jurisdictions on 
top of their other election responsibilities and canvassing and 
verifying provisional ballots and all of the rest to come up 
with a certified election result, that that poses a significant 
burden financially on local jurisdictions.
    Mrs. Davis. Did you want to comment on that?
    Mr. Sancho. Again, since we don't routinely do recounts, 
and the last one we did was 2000, legally, the issue of how 
many bodies that we have to hire, what we did was we brought 
in--it is after the election. We brought in all of our precinct 
chairmen. So we brought in--we had 150 additional people who we 
have trained over the course of several years. And we were able 
to do our portion of the recount within a day in the 
presidential election.
    So again, if you are well organized, if--and again Leon 
County is not the size of Cuyahoga, we are 150,000 voters, not 
a million voters----
    Mrs. Davis. I understand.
    Mr. Sancho. But the issue in Florida is what has it cost us 
in not verifying our elections, and I would want to say you 
don't want to pay that cost if you are an election official or 
a voter.
    Mrs. Davis. Could you talk a little about absentee ballots 
and how they are integrated into the ballot?
    Mr. Sancho. They are all optically scanned. And that is one 
of the things--and Congressman Ehlers made an excellent point. 
You are going to have problems with any system. Voter education 
is top of my list of where you must make an expenditure. Voter 
education is how you interact with the citizens and correctly 
inform of them of what the possible pitfalls are and inform 
them over and over and over again how to do this right. If you 
are not going--if you are not willing to put money into voter 
education, I don't care what technology you introduce, you are 
going to introduce problems with new technology.
    So voter education has to be considered excellent money 
spent up front. The transition from any technology to another 
one makes that more critical. And that is not where you make 
shortcuts.
    Mrs. Davis. Where you have two technologies, or three in 
some cases, because you have absentee ballots coming in though, 
is there some things as we consider the importance of the audit 
that should be discussed?
    Mr. Damschroder. Yes, Congresswoman, and I think, in fact, 
that is why it is so important that many of the provisions that 
this legislation seeks to Federalize are handled at the State 
level. In Ohio, we have 88 different counties. And there are, 
if I remember correctly, three different voting systems and 
nine different combinations of those voting systems being used 
in each of the counties, depending on, for instance, our county 
uses all touch screens in the precinct, but by mail, absentee 
ballots are on optical scanned paper ballots.
    So when we are doing those recounts, we have to have 
separate procedures and guidelines on how we conduct each of 
those recounts or audits. I think--back to your previous 
question about the cost. I think one of the additional costs 
that this legislation imposes is the requirement of using 
auditors with auditing backgrounds, as Mr. Sancho says, he uses 
poll workers. We used a lot of our staff and poll workers as 
well. None of those individuals have professional auditing 
backgrounds or experience that would qualify under the 
requirements of this Federal legislation. That would again 
raise the cost of obtaining those people.
    Mrs. Davis. This is somewhat controversial, perhaps, but 
the issue of people waiting in line and we all know that 
certainly plays a factor in the ability of people to stand in 
line with children in the rain, and terrible situations happen.
    Is it worthwhile to look at this issue in terms of a time 
factor? Is this just a matter of people you know being 
negligent about not having enough machines for people to use? 
Where does that issue play into this discussion?
    Mr. Damschroder. I think it is an important issue. But it 
is one that can't be looked at in a vacuum by just looking at 
just one of many different factors. The length of the ballot, 
the length of the language on the individual issues, all of 
those play into the part of the issue the time it takes. 
Different States' requirements, some require ID, some don't. 
All of these different factors, including the number of the 
extent of the equipment, the number of machines, all of those 
play a factor in the time it takes for an individual from the 
time it takes to park the car to the time they return, I don't 
think you can--it is an issue that has to be examined, but I 
don't think it is an issue yet that is ripe for legislation, 
but I don't think all the facets are properly studied.
    Mr. Sancho. I have a little different view on this. Coming 
into this as an aggrieved citizen who saw the system melt down 
my orientation is from the voter's perspective. We built our 
precinct voter system and technology around a voter never 
having to wait longer than 20 minutes to vote, period. That is 
the commitment that we make to the voters, we built the number 
of precincts, and the optical scan voting system actually is a 
faster voting system than using, certainly, an audio-based 
touch screen.
    And we have successfully, since 1992, been able to carry 
out that have responsibility of nobody waiting more than 20 
minutes. And when I see lines of citizens waiting four and 5 
hours to vote, I am outraged. I personally, would like to see 
it a crime for citizens to be turned away from the poll without 
being allowed to vote. To me, my role as the gatekeeper of 
democracy imposes on my position the requirement that I do what 
I must to ensure that that gate is not closed on the voter.
    Ms. Lofgren. The time has expired, and we have been joined 
by the author of the bill, not a member of the committee, and I 
would ask unanimous consent could to allow Mr. Holt to 
participate in the questioning for as long as he is able to be 
here. You are next if you want to ask a question, Rush.
    Mr. Holt. I have just one brief comment. As I came in, I 
heard some discussion about auditing. And then I also heard a 
brief comment about waiting times. With regard to the 
particular legislation that is under consideration here, the 
idea of an independent board to oversee the procedures of the 
audit, it does not require that they all be professional 
auditors. In fact, election poll workers might qualify. It is 
just that it should be overseen by a person in the State who 
has that independence from the election procedure. And that is 
the point, to establish some measure of independence.
    With regard to waiting time, I would say that is the 
subject of other legislation that I've introduced or that 
others have introduced that I think is worthy of consideration, 
but probably at another time. Thank you.
    [The statement of Mr. Holt follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.302
    
    [GRAPHIC] [TIFF OMITTED] 35804A.303
    
    Ms. Lofgren. Thank you, and welcome and thanks to this 
panel. We keep the record open for additional questions for 5 
days, so there may be additional questions to either one of 
you, and we appreciate so much your taking your time to share 
your experiences with us as we look at this important piece of 
legislation. So thank you very much and we will ask that next 
panel to come forward.
    Welcome to all of you. I was mentioning earlier as soon as 
we notice a hearing, votes are called so we always start these 
hearings late. So we will ask therefore that you summarize your 
written statements in 5 minutes. When the yellow light goes on, 
it means you have about a minute left to summarize. And your 
full statements will be made part of the official record.
    I would like to introduce, in order, Candice Hoke, who is a 
professor, a law professor, a nationally recognized and widely 
cited expert on constitutional federalism and major Federal 
regulatory programs in the context of election law. She is a 
graduate of Yale law school, former judicial clerk on U.S. 
Court of Appeals in the first district and a former staff 
member of the North Carolina governor's office. She leads the 
center for election integrity as its director and had been 
appointed as a member of the three-person Cuyahoga election 
review panel.
    Mr. Doug Lewis is executive director of the National 
Association of Election Officials. He has been the executive 
director of the National Association of Election Officials 
since 1994. Also a 501 C-3 organization known as the Election 
Center. He developed the professional education program, 
extensive training program for elections and registration 
officials which awards the certified elections registration 
administrators certificate upon completion of the program along 
with other honors and activities.
    Lawrence Norden is the counsel for the Brennan Center for 
Justice, a think tank and Public Law Institute at New York 
University School of law. He works in the areas of voting 
technology, voting rights and government accountability and is 
the lead author of several books relating to machines and 
voting systems.
    He is a graduate of the University of Chicago and the NYU 
School of Law.
    Tammy Patrick is a Federal compliance officer of Maricopa 
County, Arizona elections department, and is responsible for 
the recent expansion of the Maricopa County Elections Voter 
Assistance and Board Worker Enhancement Training Program, which 
won the National Association of Counties Achievement Award in 
2005 and 2006 and is tasked with ensuring compliance with the 
Voting Rights Act, the Americans with Disabilities Act, and 
HAVA. Ms. Patrick collaborates with community organizations and 
voter empowerment groups.
    And finally, Pamela Smith, who is the president of Verified 
Voter and Verified Voting Dot Org, and Verified Voting 
Foundation. She provides information and public testimony on 
verified voting issues on State and local levels and has 
testified before Maryland's legislature, California's Secretary 
of State's voting system and procedures panel, the San Diego 
county board of supervisors among others. She has co-authored 
written testimony on several voting systems and has made 
legislative recommendations and reports on accessibility and 
audibility issues for voting systems and other research, and 
she has been a small business and marketing consultant for many 
years originally from Chicago Illinois. She is now a resident 
of Carlsbad, California.

     STATEMENTS OF CANDICE HOKE, DIRECTOR, CLEVELAND STATE 
   UNIVERSITY CENTER FOR ELECTION INTEGRITY; R. DOUG LEWIS, 
EXECUTIVE DIRECTOR, NATIONAL ASSOCIATION OF ELECTION OFFICIALS; 
  LAWRENCE NORDEN, COUNSEL, BRENNAN CENTER FOR JUSTICE; TAMMY 
PATRICK, FEDERAL COMPLIANCE OFFICER, MARICOPA COUNTY ELECTIONS 
    DEPARTMENT; AND PAMELA SMITH, PRESIDENT, VERIFIED VOTER

    Ms. Lofgren. So welcome to all of you. And if we may begin 
with Ms. Hoke and move forward for all of your statements.

                   STATEMENT OF CANDICE HOKE

    Ms. Hoke. Good afternoon, Chairwoman Lofgren, Ranking 
Member McCarthy and Committee members. Thank you for inviting 
the Center for Election Integrity of Cleveland State University 
to participate in this hearing on Federal Election Auditing. 
Since 2006, our Center has held the appointment as Public 
Monitor of Cuyahoga County Election Reform.
    In October, we initiated and obtained Cuyahoga Board of 
Elections approval to conduct an unprecedented Collaborative 
Public Audit of the November 2006 General Election. This 
independent audit was conducted jointly by the County 
Democratic party, the County Republican party and three 
election reform organizations. Professional auditors were 
participants in its planning and execution.
    Our Center may be somewhat unusual in that we have been 
guided by a principled commitment to Federalism as the basis of 
our election systems. My own scholarly and consulting 
background prior to undertaking work in election reform focused 
on the 10th Amendment and Supremacy Clause. In the mid-1990s, 
for instance, the U.S. Senate's Judiciary Committee, 
Subcommittee on the Constitution, enlisted me to work on 
legislation to stem Federal preemption of State law.
    I have not advocated Federal legislation as the solution to 
all electoral administrative problems and don't believe it 
alone can cure the inadequacies present today. But even as a 
strong proponent of State powers and Federalism, I would urge 
that Federal election legislation plays a critical, though as 
yet, underutilized role in achieving quality elections 
throughout the Nations. The election system can become a prime 
example of cooperative Federalism.
    Given the expertise present to discuss election audits 
today, I would like to address only 2 issues in oral testimony: 
first, the importance of requiring independence of the election 
auditing structure from the election officials; and secondly, 
the need to ensure that essential administrative duties for 
achieving valid trustworthy Federal election audits will be 
fulfilled.
    First on independence: to achieve voter confidence that all 
valid ballots were counted and counted accurately, the 
auditors' independence from the election officials and the 
system responsible for conducting the election is imperative.
    As my written testimony details, independence is critical 
for a number of reasons that have been documented in a variety 
of studies including the SEC's own work on auditing.
    Any federally mandated auditing process that does not 
include true independence will waste taxpayers' time and money 
and will likely retard rather than augment voter confidence.
    Turning to the second recommendation that we made in our 
written testimony, we would ask that you ensure that the 
essential administrative tasks for achieving valid trustworthy 
Federal election audits will be fulfilled. In Cuyahoga County, 
our collaborative audit was constructed so that the audit 
policy decisions and choices of races to be audited were 
invested collectively in the participating political parties 
and advocacy organizations. Our center was to identify the 
audit materials needed and to work with the Cuyahoga Board of 
Elections staff to ensure the logistics and the chain of 
custody of all needed materials.
    The Board of Election Board Members unanimously approved 
the audit process. We expected staff cooperation and few 
problems. But I was surprised by the number of impediments that 
were interposed. First, the existing administrative procedures 
for processing Election Day materials upon arrival in the BOE 
meant that we had a difficult time securing the materials in a 
chain of custody. Second, we discovered software design and 
electronic election data availability issues that we had not 
anticipated. Third, certain hardware design issues impeded 
effective audits. And fourth, we found to our chagrin and 
sadness that there was staff nonperformance or obstructionist 
conduct that delayed our audit.
    My written testimony provides details about the first three 
of these types of impediments. I would like to focus on the 
fourth, staff nonperformance or opposition.
    In my experience in Cuyahoga County and elsewhere, I found 
that many election officials approached their jobs with a high 
commitment to achieving the best possible elections 
administrative record with a verifiably accurate and legal set 
of results. But unfortunately, there is a ``group 2,'' or 
pockets of election officials who have enjoyed historically 
unchecked, broad discretionary authority over election 
performance and over reported election results. Members of this 
second group tend to disfavor public accountability and 
independent verification of election results, and thus view 
audits as inimical to their interests.
    I see I need to conclude. I would ask that you keep in mind 
when you develop the Federal approach that we need to have a 
process articulated preferably through the States' Secretary of 
State or their chief election officers that will anticipate 
various kinds of procedures that need to be in place for 
effective and expeditious election auditing to occur. A State 
based system of data collection and other procedures must be 
articulated well in advance of the audits which will also 
include specified consequences for local officials if they do 
not perform. Thank you.
    Ms. Lofgren. Thank you.
    [The statement of Ms. Hoke follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.005
    
    [GRAPHIC] [TIFF OMITTED] 35804A.006
    
    [GRAPHIC] [TIFF OMITTED] 35804A.007
    
    [GRAPHIC] [TIFF OMITTED] 35804A.008
    
    [GRAPHIC] [TIFF OMITTED] 35804A.009
    
    [GRAPHIC] [TIFF OMITTED] 35804A.010
    
    [GRAPHIC] [TIFF OMITTED] 35804A.011
    
    [GRAPHIC] [TIFF OMITTED] 35804A.012
    
    [GRAPHIC] [TIFF OMITTED] 35804A.013
    
    [GRAPHIC] [TIFF OMITTED] 35804A.014
    
    Ms. Lofgren. Mr. Lewis.

                   STATEMENT OF R. DOUG LEWIS

    Mr. Lewis. I obviously cannot cram 8 pages of testimony 
into 5 minutes, so let me just try to summarize as best I can 
for you all.
    We do a form of audits now. Every jurisdiction in America 
does a form of audits. It is called canvass the vote, and this 
is one of those things that--a whole lot of people who are not 
involved in elections do not understand it, but we go through a 
process of where we actually do try to reconcile all of the 
numbers and make sure that the numbers are what the numbers 
should be; that we go back and look when the numbers do not 
make sense, and in fact, usually whenever there is a problem 
with an election, it is because an election official found that 
the numbers did not make sense somewhere, and then he went back 
to investigate to find out why.
    In terms of whether or not we ought to have an audit that 
verifies that the equipment is counting correctly, quite 
frankly, I do not think you are going to find that most 
election officials disagree with that. It is how it gets done. 
It is whether or not what you are doing in H.R. 811 actually 
gets to that and does it in such a prescriptive manner that it 
really makes it very difficult to do this.
    So I would say to you, as Congressional people, rather than 
trying to tell us exactly how to do each and every one of these 
things, determine the areas that you want to have policy 
decisions made in, and instruct us to figure out a way to do 
that. Tell the States and the locales what the basic standards 
are that you want to achieve, and then let the laboratory of 
7,800 jurisdictions in America go to work. You will find there 
are several right answers more often than not, and I have been 
on record as saying that having an audit that proves that 
voting systems work is not a bad deal. That is a good deal. It 
is just a matter of how we get there.
    The problem, I think, is that if we get this thing so 
complicated, instead of achieving your goals of higher 
believability and acceptance of the process, we end up 
destroying the very thing that you are trying to improve, and 
so let us make sure that we do not get to the point that we 
complicate this so much to the point that nobody has faith in 
it at all. We are at the point where we have got to be careful 
that we keep beating on the system, alleging that somehow the 
system is wrong when the system works pretty cotton-picking 
well. It may not work well in the estimation of some who are 
very, very partisan in their viewpoint on this (and that is on 
both sides of the aisle depending on what the issue is) and so 
it is one of those that we have to look at. If you go to the 
point of trying to build escalating audits of all of the 
Federal races and going through counting all of them, you are 
going to find that humans count far less accurately than 
equipment does. Every time we go into one of these recounts, 
every time we even do the audits in the States that actually do 
these audits, you find that the real problems come from humans, 
not from the equipment. I mean, we do have some equipment 
problems on occasion, but it is rare when compared to the 
problems that humans make, and so it seems to me that where we 
need to be in this is, do not assume that humans are going to 
be the better answer for this whole process because, as 
election officials, we know after recounts and after audits of 
doing this, when we find an error, it is usually the humans who 
have made the error and not the equipment.
    In terms of when you look at the costs of all of this, 
there always is a tendency, I think, of any legislative body, 
that when it wants something, it always underestimates the cost 
of what it is going to cost to accomplish the objectives, and 
it is always easy to shift that cost to somebody else that you 
are not paying for, to always push that down to the lowest 
level and say, well, the jurisdictions can pick up that cost.
    The problem with all of this is that we have started to 
make the jurisdictions responsible for all costs and, you must 
know that they are almost at the breaking point in terms of the 
cost of election changes. My concern is this: If Congress 
begins to assume that it knows more about the management of 
practices of elections and that it knows more about what will 
work than local governments do, that it wants to be uniform 
across the United States, then I think democracy is going to be 
in for a very rough period.
    Secondly, in wrapping this up, let me say to you election 
officials in America are dedicated people. They do this with 
little money, not a whole lot of support most of the time. They 
get vilified as somehow people who are manipulating elections 
when none of that seems to be able to be proven in 99 percent 
of the cases. It seems to me we need to recognize these are 
good people trying to do a decent and honest job.
    Thank you.
    [The statement of Mr. Lewis follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.015
    
    [GRAPHIC] [TIFF OMITTED] 35804A.016
    
    [GRAPHIC] [TIFF OMITTED] 35804A.017
    
    [GRAPHIC] [TIFF OMITTED] 35804A.018
    
    [GRAPHIC] [TIFF OMITTED] 35804A.019
    
    [GRAPHIC] [TIFF OMITTED] 35804A.020
    
    [GRAPHIC] [TIFF OMITTED] 35804A.021
    
    [GRAPHIC] [TIFF OMITTED] 35804A.022
    
    Ms. Lofgren. Thank you very much for your testimony.
    Mr. Norden.

                  STATEMENT OF LAWRENCE NORDEN

    Mr. Norden. Thank you.
    Thank you to the subcommittee for holding a hearing on 
post-election audits, I think a subject that probably would 
strike most people as rather boring but that, in fact, is 
critical if we are going to secure our elections, improve them 
and restore full public confidence in our electoral system.
    My name is Lawrence Norden, and for 18 months, I was chair 
for the Brennan Center Task Force on Voting System Security. 
This task force included the Nation's leading computer 
scientists and security professionals, including scientists 
from the National Institute of Standards and Technology, and 
the former chief security officer of Microsoft and former cyber 
security czar for President George W. Bush.
    What the task force found, I think at this point, is a 
consensus view among security experts and computer scientists 
who have looked at electronic voting machines, and that is that 
there are serious security and reliability vulnerabilities in 
these machines. The good news is that there is also a 
substantial consensus on how to address these vulnerabilities, 
and one of the most important things, it is agreed, that we can 
do is to use post-election audits to compare voter-verified 
independent records to the electronic tally. Fortunately, this 
is already happening in many States. As the chair mentioned, 
Connecticut in the last election audited 20 percent of its 
precincts that were using optical scan machines. Illinois 
regularly audits 5 percent of its precincts, and California 
audits, as many on the subcommittee know, 1 percent of its 
precincts. It has for a long time. In many cases, this is 
counting up to 200 races in a single county. Unfortunately, 
this is not being done in most States, and it is the reason why 
we need Federal legislation mandating good post-election 
audits. What would that look like?
    The Brennan Center has been working with the University of 
California at Berkeley's Samuelson Clinic, and with many 
leading election officials and academics to answer that 
question, and we will be issuing a report on that subject 
shortly. I detail many recommendations in my written testimony. 
I am going to briefly mention a few of them.
    The first is that it is crucial that the selection process 
for audits be random and transparent, and Cuyahoga County, in 
2004, is an excellent example of why this is necessary. In 
2004, after the general election, election officials were doing 
a 3-percent post-election audit and pre-sorted the ballots to 
make sure they matched with the electronic ballots, so the 
audits showed that the paper and electronic records matched. 
But this actually was not a check on the system; it did not 
prove anything. If the process for selecting the machines to be 
audited were public and it were clear to the public that the 
selection was random, this is a problem that could have been 
avoided.
    Secondly, we need to increase scrutiny, audit scrutiny, in 
close elections. In most Federal elections, a 3-percent audit 
will be sufficient to have confidence that a full recount would 
not change the results, but in very close races where just the 
corruption of a few votes or a mis-tally could change the 
outcome of an election, a 3-percent audit does not give you, 
really, any confidence that you will have caught an error that 
might have occurred. To his credit, Congressman Holt introduced 
this concept in his legislation.
    Third, a good chain of custody is key, and again, the 
Cuyahoga example is a great example for why this is necessary. 
If poll workers are going to presort ballots or take voting 
machines home with them, we cannot really have much confidence 
in the audit.
    So, in conclusion, if voter-verified paper records are 
going to be much more than an inconvenience to poll workers and 
election officials, we really need to use them to check the 
electronic tally. The debate over voting system security is not 
really a debate over paper or what technology we use. It is a 
debate about increasing public confidence in our voting system, 
and ultimately, good post-election audits are what will 
increase that public confidence.
    [The statement of Mr. Norden follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.023
    
    [GRAPHIC] [TIFF OMITTED] 35804A.024
    
    [GRAPHIC] [TIFF OMITTED] 35804A.025
    
    [GRAPHIC] [TIFF OMITTED] 35804A.026
    
    [GRAPHIC] [TIFF OMITTED] 35804A.027
    
    [GRAPHIC] [TIFF OMITTED] 35804A.028
    
    [GRAPHIC] [TIFF OMITTED] 35804A.029
    
    [GRAPHIC] [TIFF OMITTED] 35804A.030
    
    [GRAPHIC] [TIFF OMITTED] 35804A.031
    
    [GRAPHIC] [TIFF OMITTED] 35804A.032
    
    [GRAPHIC] [TIFF OMITTED] 35804A.033
    
    [GRAPHIC] [TIFF OMITTED] 35804A.034
    
    [GRAPHIC] [TIFF OMITTED] 35804A.035
    
    [GRAPHIC] [TIFF OMITTED] 35804A.036
    
    [GRAPHIC] [TIFF OMITTED] 35804A.037
    
    [GRAPHIC] [TIFF OMITTED] 35804A.038
    
    [GRAPHIC] [TIFF OMITTED] 35804A.039
    
    [GRAPHIC] [TIFF OMITTED] 35804A.040
    
    [GRAPHIC] [TIFF OMITTED] 35804A.041
    
    [GRAPHIC] [TIFF OMITTED] 35804A.042
    
    [GRAPHIC] [TIFF OMITTED] 35804A.043
    
    [GRAPHIC] [TIFF OMITTED] 35804A.044
    
    [GRAPHIC] [TIFF OMITTED] 35804A.045
    
    [GRAPHIC] [TIFF OMITTED] 35804A.046
    
    [GRAPHIC] [TIFF OMITTED] 35804A.047
    
    [GRAPHIC] [TIFF OMITTED] 35804A.048
    
    [GRAPHIC] [TIFF OMITTED] 35804A.049
    
    [GRAPHIC] [TIFF OMITTED] 35804A.050
    
    [GRAPHIC] [TIFF OMITTED] 35804A.051
    
    [GRAPHIC] [TIFF OMITTED] 35804A.052
    
    [GRAPHIC] [TIFF OMITTED] 35804A.053
    
    [GRAPHIC] [TIFF OMITTED] 35804A.054
    
    [GRAPHIC] [TIFF OMITTED] 35804A.055
    
    [GRAPHIC] [TIFF OMITTED] 35804A.056
    
    [GRAPHIC] [TIFF OMITTED] 35804A.057
    
    [GRAPHIC] [TIFF OMITTED] 35804A.058
    
    [GRAPHIC] [TIFF OMITTED] 35804A.059
    
    [GRAPHIC] [TIFF OMITTED] 35804A.060
    
    [GRAPHIC] [TIFF OMITTED] 35804A.061
    
    [GRAPHIC] [TIFF OMITTED] 35804A.062
    
    [GRAPHIC] [TIFF OMITTED] 35804A.063
    
    [GRAPHIC] [TIFF OMITTED] 35804A.064
    
    [GRAPHIC] [TIFF OMITTED] 35804A.065
    
    [GRAPHIC] [TIFF OMITTED] 35804A.066
    
    [GRAPHIC] [TIFF OMITTED] 35804A.067
    
    [GRAPHIC] [TIFF OMITTED] 35804A.068
    
    [GRAPHIC] [TIFF OMITTED] 35804A.069
    
    [GRAPHIC] [TIFF OMITTED] 35804A.070
    
    [GRAPHIC] [TIFF OMITTED] 35804A.071
    
    [GRAPHIC] [TIFF OMITTED] 35804A.072
    
    [GRAPHIC] [TIFF OMITTED] 35804A.073
    
    [GRAPHIC] [TIFF OMITTED] 35804A.074
    
    [GRAPHIC] [TIFF OMITTED] 35804A.075
    
    [GRAPHIC] [TIFF OMITTED] 35804A.076
    
    [GRAPHIC] [TIFF OMITTED] 35804A.077
    
    [GRAPHIC] [TIFF OMITTED] 35804A.078
    
    [GRAPHIC] [TIFF OMITTED] 35804A.079
    
    [GRAPHIC] [TIFF OMITTED] 35804A.080
    
    [GRAPHIC] [TIFF OMITTED] 35804A.081
    
    [GRAPHIC] [TIFF OMITTED] 35804A.082
    
    [GRAPHIC] [TIFF OMITTED] 35804A.083
    
    [GRAPHIC] [TIFF OMITTED] 35804A.084
    
    [GRAPHIC] [TIFF OMITTED] 35804A.085
    
    [GRAPHIC] [TIFF OMITTED] 35804A.086
    
    [GRAPHIC] [TIFF OMITTED] 35804A.087
    
    [GRAPHIC] [TIFF OMITTED] 35804A.088
    
    [GRAPHIC] [TIFF OMITTED] 35804A.089
    
    [GRAPHIC] [TIFF OMITTED] 35804A.090
    
    [GRAPHIC] [TIFF OMITTED] 35804A.091
    
    [GRAPHIC] [TIFF OMITTED] 35804A.092
    
    [GRAPHIC] [TIFF OMITTED] 35804A.093
    
    [GRAPHIC] [TIFF OMITTED] 35804A.094
    
    [GRAPHIC] [TIFF OMITTED] 35804A.095
    
    [GRAPHIC] [TIFF OMITTED] 35804A.096
    
    [GRAPHIC] [TIFF OMITTED] 35804A.097
    
    [GRAPHIC] [TIFF OMITTED] 35804A.098
    
    [GRAPHIC] [TIFF OMITTED] 35804A.099
    
    [GRAPHIC] [TIFF OMITTED] 35804A.100
    
    [GRAPHIC] [TIFF OMITTED] 35804A.101
    
    [GRAPHIC] [TIFF OMITTED] 35804A.102
    
    [GRAPHIC] [TIFF OMITTED] 35804A.103
    
    [GRAPHIC] [TIFF OMITTED] 35804A.104
    
    [GRAPHIC] [TIFF OMITTED] 35804A.105
    
    [GRAPHIC] [TIFF OMITTED] 35804A.106
    
    [GRAPHIC] [TIFF OMITTED] 35804A.107
    
    [GRAPHIC] [TIFF OMITTED] 35804A.108
    
    [GRAPHIC] [TIFF OMITTED] 35804A.109
    
    [GRAPHIC] [TIFF OMITTED] 35804A.110
    
    [GRAPHIC] [TIFF OMITTED] 35804A.111
    
    [GRAPHIC] [TIFF OMITTED] 35804A.112
    
    [GRAPHIC] [TIFF OMITTED] 35804A.113
    
    [GRAPHIC] [TIFF OMITTED] 35804A.114
    
    [GRAPHIC] [TIFF OMITTED] 35804A.115
    
    [GRAPHIC] [TIFF OMITTED] 35804A.116
    
    [GRAPHIC] [TIFF OMITTED] 35804A.117
    
    [GRAPHIC] [TIFF OMITTED] 35804A.118
    
    [GRAPHIC] [TIFF OMITTED] 35804A.119
    
    [GRAPHIC] [TIFF OMITTED] 35804A.120
    
    [GRAPHIC] [TIFF OMITTED] 35804A.121
    
    [GRAPHIC] [TIFF OMITTED] 35804A.122
    
    [GRAPHIC] [TIFF OMITTED] 35804A.123
    
    [GRAPHIC] [TIFF OMITTED] 35804A.124
    
    [GRAPHIC] [TIFF OMITTED] 35804A.125
    
    [GRAPHIC] [TIFF OMITTED] 35804A.126
    
    [GRAPHIC] [TIFF OMITTED] 35804A.127
    
    [GRAPHIC] [TIFF OMITTED] 35804A.128
    
    [GRAPHIC] [TIFF OMITTED] 35804A.129
    
    [GRAPHIC] [TIFF OMITTED] 35804A.130
    
    [GRAPHIC] [TIFF OMITTED] 35804A.131
    
    [GRAPHIC] [TIFF OMITTED] 35804A.132
    
    [GRAPHIC] [TIFF OMITTED] 35804A.133
    
    [GRAPHIC] [TIFF OMITTED] 35804A.134
    
    [GRAPHIC] [TIFF OMITTED] 35804A.135
    
    [GRAPHIC] [TIFF OMITTED] 35804A.136
    
    [GRAPHIC] [TIFF OMITTED] 35804A.137
    
    [GRAPHIC] [TIFF OMITTED] 35804A.138
    
    [GRAPHIC] [TIFF OMITTED] 35804A.139
    
    [GRAPHIC] [TIFF OMITTED] 35804A.140
    
    [GRAPHIC] [TIFF OMITTED] 35804A.141
    
    [GRAPHIC] [TIFF OMITTED] 35804A.142
    
    [GRAPHIC] [TIFF OMITTED] 35804A.143
    
    [GRAPHIC] [TIFF OMITTED] 35804A.144
    
    [GRAPHIC] [TIFF OMITTED] 35804A.145
    
    [GRAPHIC] [TIFF OMITTED] 35804A.146
    
    [GRAPHIC] [TIFF OMITTED] 35804A.147
    
    [GRAPHIC] [TIFF OMITTED] 35804A.148
    
    [GRAPHIC] [TIFF OMITTED] 35804A.149
    
    [GRAPHIC] [TIFF OMITTED] 35804A.150
    
    [GRAPHIC] [TIFF OMITTED] 35804A.151
    
    [GRAPHIC] [TIFF OMITTED] 35804A.152
    
    [GRAPHIC] [TIFF OMITTED] 35804A.153
    
    [GRAPHIC] [TIFF OMITTED] 35804A.154
    
    [GRAPHIC] [TIFF OMITTED] 35804A.155
    
    [GRAPHIC] [TIFF OMITTED] 35804A.156
    
    [GRAPHIC] [TIFF OMITTED] 35804A.157
    
    [GRAPHIC] [TIFF OMITTED] 35804A.158
    
    [GRAPHIC] [TIFF OMITTED] 35804A.159
    
    [GRAPHIC] [TIFF OMITTED] 35804A.160
    
    [GRAPHIC] [TIFF OMITTED] 35804A.161
    
    [GRAPHIC] [TIFF OMITTED] 35804A.162
    
    [GRAPHIC] [TIFF OMITTED] 35804A.163
    
    [GRAPHIC] [TIFF OMITTED] 35804A.164
    
    [GRAPHIC] [TIFF OMITTED] 35804A.165
    
    [GRAPHIC] [TIFF OMITTED] 35804A.166
    
    [GRAPHIC] [TIFF OMITTED] 35804A.167
    
    [GRAPHIC] [TIFF OMITTED] 35804A.168
    
    [GRAPHIC] [TIFF OMITTED] 35804A.169
    
    [GRAPHIC] [TIFF OMITTED] 35804A.170
    
    [GRAPHIC] [TIFF OMITTED] 35804A.171
    
    [GRAPHIC] [TIFF OMITTED] 35804A.172
    
    [GRAPHIC] [TIFF OMITTED] 35804A.173
    
    [GRAPHIC] [TIFF OMITTED] 35804A.174
    
    [GRAPHIC] [TIFF OMITTED] 35804A.175
    
    [GRAPHIC] [TIFF OMITTED] 35804A.176
    
    [GRAPHIC] [TIFF OMITTED] 35804A.177
    
    [GRAPHIC] [TIFF OMITTED] 35804A.178
    
    [GRAPHIC] [TIFF OMITTED] 35804A.179
    
    [GRAPHIC] [TIFF OMITTED] 35804A.180
    
    [GRAPHIC] [TIFF OMITTED] 35804A.181
    
    [GRAPHIC] [TIFF OMITTED] 35804A.182
    
    [GRAPHIC] [TIFF OMITTED] 35804A.183
    
    [GRAPHIC] [TIFF OMITTED] 35804A.184
    
    [GRAPHIC] [TIFF OMITTED] 35804A.185
    
    [GRAPHIC] [TIFF OMITTED] 35804A.186
    
    [GRAPHIC] [TIFF OMITTED] 35804A.187
    
    [GRAPHIC] [TIFF OMITTED] 35804A.188
    
    [GRAPHIC] [TIFF OMITTED] 35804A.189
    
    [GRAPHIC] [TIFF OMITTED] 35804A.190
    
    [GRAPHIC] [TIFF OMITTED] 35804A.191
    
    [GRAPHIC] [TIFF OMITTED] 35804A.192
    
    [GRAPHIC] [TIFF OMITTED] 35804A.193
    
    [GRAPHIC] [TIFF OMITTED] 35804A.194
    
    [GRAPHIC] [TIFF OMITTED] 35804A.195
    
    [GRAPHIC] [TIFF OMITTED] 35804A.196
    
    [GRAPHIC] [TIFF OMITTED] 35804A.197
    
    [GRAPHIC] [TIFF OMITTED] 35804A.198
    
    [GRAPHIC] [TIFF OMITTED] 35804A.199
    
    Ms. Lofgren. Thank you very much, Mr. Norden.
    Ms. Patrick.

                   STATEMENT OF TAMMY PATRICK

    Ms. Patrick. Madam Chair, members of the committee, thank 
you.
    My name is Tammy Patrick. I am the Federal Compliance 
Officer for the Elections Department in Maricopa County, 
Arizona. In last November's general election, we conducted a 
statutory hand-audit of 2 percent of the precinct-cast ballots 
and 1 percent of our early ballots in a total of four races 
that combined both Federal and State offices. I refer you to my 
written testimony for more information on that audit, and I am 
going to confine my oral remarks to three very brief points.
    First of all, a little background. Maricopa County has 1.5 
million registered voters, 1,142 voting precincts, and about 
half of our voters vote by mail-in early ballots. At the same 
time that we were conducting this new hand-audit, we also had 
well over 100 workers working from 6:00 in the morning to 12:00 
p.m. Midnight, processing our provisional ballots. We had 
difficulty securing the 144 individuals who were needed to 
conduct the hand recount of that 1 percent of the early ballots 
and 2 percent of the ballots cast at the precinct. We also had 
additional difficulty keeping them long enough to actually 
finish the audit. I cannot fathom how we would conduct a hand-
audit of 10 percent of the ballots, especially under the 
requirements of H.R. 811.
    Secondly, it is critical to recognize the limits of a human 
count. For example, our procedures manual, which is written by 
the Secretary of State, directed audit boards to count ballots 
in lots of 25. We found that, in application, that was too 
difficult. It seems easy enough, but in fact, we had several 
boards that we had to instruct to stack in groups of 10 in 
order to have the counts come out correctly. Our State law did 
not expect perfection from the human count and recognized an 
acceptable variation between the hand-audit and what was done 
by the machine tabulation. We could not identify and address 
every discrepancy as required by H.R. 811 within any realistic 
time frame. An example of this would be in instances where, in 
one precinct, the audit board and the electronic machine came 
up with the exact same number of ballots but had discrepancy in 
the total number of votes cast, so the boards were not quite 
understanding in some instances an overvote, for instance, they 
voted for two when it should have only been for one, so they 
allocated each one of those to two separate candidates.
    Finally, it goes without saying that we cannot conduct a 
recount that includes provisional ballots until we have 
finished counting them. The requirements in H.R. 811 to retain 
the paper ballots, including those cast on a DRE, in a manner 
so as not to enable them to be tied to a voter, would either 
preclude any provisional ballot from being cast on a DRE or it 
would immediately impact the ability of the audit's totals to 
match reported totals. So, by virtue of the parameters set 
forth in H.R. 811, a successful hand-audit is impossible. 
Additionally, we could not sort early ballots in with our 
election day ballots, as required by H.R. 811 without 
compromising the inherent security of random order retention. 
Although we report our results of our early ballots at a 
precinct level, we tabulate and store them in a random, mixed 
batch as they come in from the post office, and that is how we 
count to preserve both the security and the privacy of the 
voter. So, in our audit, we take those batches of 200 and audit 
them in their entirety as a batch, and we would not be able to 
do that, thus resulting in those ballots being handled multiple 
times.
    We are very fortunate that everyone in this room has a 
common purpose. Whether you are an election official 
responsible for conducting and tabulating an election, a 
political observer overseeing this process, an elected official 
whose name appears on the ballot or all of us simply as voters 
ourselves, the outcome of the election is that we are all 
working for the same end, that it is an accurate reflection of 
the will and intent of the people. But the ability to 
administer an audit and to implement change in response to the 
unique environments within which we all work on our local 
jurisdictional level is an integral characteristic of success 
of such an audit and must remain at the local level.
    I thank you for the opportunity to testify today, and I 
look forward to answering any questions.
    [The statement of Ms. Patrick follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.200
    
    [GRAPHIC] [TIFF OMITTED] 35804A.201
    
    Ms. Lofgren. Thank you very much.
    Our final witness is Ms. Pamela Smith.

                   STATEMENT OF PAMELA SMITH

    Ms. Smith. Thank you, Chairman Lofgren and Ranking Member 
McCarthy and members of the committee. Thanks for the 
opportunity to be here and to testify today.
    VerifiedVoting.org and the Verified Voting Foundation were 
actually started by a computer scientist. We believe that, if 
our voting systems are reliable and those election officials 
and the voting public have a way to assure themselves that they 
are accurate, then that will go a long way to increasing public 
confidence and, therefore, participation in the process.
    We think transparency in elections is so important that we 
launched an election transparency project in 2006 to promote 
observation of all phases of the election by citizens, 
including the audits. We think a key purpose of an election 
audit is to convince the losers that they have lost. The 
winners pretty much always think they have won, so that part is 
easy. I will confine my remarks today to talking a little bit 
about the successes that States have seen that are doing audits 
already and also about costs.
    Not every State has voter-verified paper records that they 
can audit at this time, but of those that do, still only a 
fraction are doing statewide audits. In the ones that are doing 
statewide audits, there have been successes that illustrate 
that audits can be done effectively and cost-effectively, and 
Congress can help us make our elections more transparent and 
reliable by passing a requirement.
    A great example of what an audit can tell us about a voting 
system, not just about the accuracy of the outcome but about 
the system itself, occurred in a place that does not yet 
require audits. Pottawamie County, Iowa, in June of 2006, on 
the election night of the Republican primary, an experienced 
county election official there by the name of Marilyn Jo Drake 
noticed something odd in the vote tallies when the returns were 
coming in. A popular incumbent seemed to suddenly be losing to 
an unknown newcomer who had scarcely even campaigned. On her 
own authority, she performed an audit of a single sample 
precinct to see if the voting machines were counting correctly. 
They were not. There was a programming error, which meant that 
votes for candidate A started going to candidate B. A full 
hand-count showed the correct result. The machine count was 
wrong. Suppose both candidates had been unknown, neither one an 
incumbent, nothing to give her the idea that something was 
amiss. A routine audit would still have uncovered the problem. 
Ms. Drake could have just chosen expediency over accuracy. 
There was no rule requiring her to do this audit. It made more 
work, but it resulted in two important things--the right 
candidate's being seated in office and the exposure of a glitch 
in the system that could then be repaired or prevented the next 
time.
    Auditing really means choosing accuracy over expediency. 
Yes, it takes some time. Yes, there is some cost involved, but 
if it is going to be a safety net for voter intent, we have to 
see it as just one step in what it takes to get our elections 
right. It is no different than taking care to program the right 
names on the ballot or recruiting enough poll workers for 
election day, and you cannot just do it once in a while. You 
have to do audits every time. You cannot say, ``See, the 
tallies match every time. Now we do not have to do this 
anymore.'' that would be a dangerous perspective that fails not 
only to comprehend voting system security but also Murphy's 
Law. There will be problems at every election. What audits and 
voter-verified paper records do is make those problems 
solvable.
    To talk about costs just a little bit in the States, there 
is workload. There is a big difference in what one State is 
doing compared to another, but interestingly, some of the 
States are actually increasing the amounts that they audit, 
voluntarily choosing to expand their law or go higher.
    What does this add to the budget? It is far from 
prohibitive. In some cases, the cost is very low, indeed. One 
of your key costs is what you pay the people doing the 
counting, plus your supervisory and planning costs, of course. 
It takes a time--a big difference in time is as variable as the 
type of ballots you audit. Paper ballots are much easier to 
count than thermal paper rolls.
    What would this look like as a national investment? From 
what we have been able to find out from the States, it just 
would not cost that much, but to compare, take the Washington 
State gubernatorial recount from 2004. The actual cost total 
was about $900,000. They looked at one race on 2.8 million 
ballots. It worked out to about 31 cents a ballot. If the 
provisions in H.R. 811 had been in effect in 2004, nationwide, 
we would have audited just under 7,000 precincts, and 
nationwide, we would have checked votes from two or three races 
per ballot, call it 11.5 million votes checked. Even at a more 
conservative rate of, say, 35 cents a ballot, it is a little 
over $4 million. It is not that much.
    What we recommend is for audits to be effective, ensure the 
selection process is verifiably random; make sure the audit 
happens after all of the ballots have been counted and initial 
results are in; include all of the ballot types, including 
absentees. You know, I wanted to say that, as to absentee 
ballots, it is important to be able to track those for also 
being able to secure the vote and to secure the chain of 
custody, so whatever we could do to track absentees is 
important. Make sure that officials invite the public in to 
observe. Do not just allow them in, but actually invite their 
participation, and they will appreciate that and participate 
more as time goes forward.
    Thank you so much.
    [The statement of Ms. Smith follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.202
    
    [GRAPHIC] [TIFF OMITTED] 35804A.203
    
    [GRAPHIC] [TIFF OMITTED] 35804A.204
    
    [GRAPHIC] [TIFF OMITTED] 35804A.205
    
    [GRAPHIC] [TIFF OMITTED] 35804A.206
    
    [GRAPHIC] [TIFF OMITTED] 35804A.207
    
    [GRAPHIC] [TIFF OMITTED] 35804A.208
    
    [GRAPHIC] [TIFF OMITTED] 35804A.209
    
    [GRAPHIC] [TIFF OMITTED] 35804A.210
    
    [GRAPHIC] [TIFF OMITTED] 35804A.211
    
    Ms. Lofgren. Thank you very much. This is very interesting 
and useful, and as I was listening, I was thinking about the 
years I spent on the Santa Clara County Board of Supervisors 
when we had elections that we were in charge of, and it always 
ended up that we were the last to be counted, I think, in the 
State of California because we were so cheap about new 
equipment, but none of us at the time really ever considered 
that it would be inaccurate. It is only as elections have 
become closer and closer that we realize that the technology--
you know, that there is no 100 percent, and it is something 
that most of us really were not aware of and that we did not 
focus in on, and really, although when it comes down to a 
recount situation and one side is going to lose, you know, you 
see partisan issues, and it is just a normal part of American 
life. But the fact is, if there is a defect, it can benefit or 
disadvantage either party. So, you know, I am willing to run an 
election and either win or lose. What I want to make sure of is 
that the people who are making the decisions are the voters, 
not anyone else, and that really gets to the question of these 
voting machines because we cannot see it in the same way, and 
Mr. Ehlers is right. I mean, there is a history. You can 
defraud the electorate with paper ballots, too, and we know 
historically there have been times when that has happened, and 
that is wrong, but the concern about the machines is that you 
would not even know it, and so here is the question.
    Ms. Hoke, you mentioned in the audit--and I am not 
familiar, really, with the audit--that there were software 
issues, that there were hardware problems. Can you briefly 
describe what those issues were and how we could set a standard 
to avoid something like that?
    Ms. Hoke. Yes, Madam Chairman, I would be happy to.
    The software issues--I apologize for my phone. I should 
have turned it off.
    Ms. Lofgren. It is a beautiful sound.
    Ms. Hoke. It is about to quit.
    The software situation is that we specifically requested, 
using generic terms, the electronic files that we needed to be 
able to compare in with the DRE machine totals, meaning the 
long reports from the DRE units and the optical scan totals. 
Those files were not made available to us although I had stayed 
up 36 hours to catch the files at the conclusion of the 
election.
    So, as it turned out, the IT and Ballot Department managers 
said that the files that we needed did not really exist. After 
a long series of discussions--and we still do not know whether 
a singular file exists, since we do not have access to the 
Diebold information. We ask: what is possible to obtain? What 
files are created? Were there really two usable files that 
local officials simply did not tell us about? Was it that the 
software does not produce files that make it easy to audit? We 
do not know the answer, but we were told that we would need a 
whole range of files and that we would have to take file A and 
back out the data from file B from file A, and it was a 
complicated series of steps to end up with a number that then 
we could compare against certain precinct totals.
    One of the reasons that we very much need a supervisory 
authority who has access to the--shall we call it--I do not 
agree with ``proprietary information'' but, to whatever degree 
it is, for the authority to be able to study what files should 
be made available that are appropriate for the audit, that 
mandate at what time during the tabulations those files will be 
saved, and then produced at what time. We must have been at the 
Board of Elections five different times just on a mission to 
obtain the electronic files. That is a problem.
    As far as the hardware, the optical scan devices of at 
least the two major election vendors do not count ballots, 
optical scan ballots. They only count ballot pages, which means 
that it is not possible simply by feeding thousands of ballot 
pages through a scanner to know how many ballots have actually 
been counted. It is a separate step that must be undertaken. It 
is perhaps a hardware design issue plus software issue for us 
to be able to report accurately how many ballots were counted.
    Ms. Lofgren. Mr. Norden, you have been with the Brennan 
Center for a long time and have looked at these technology 
issues. Do you have recommendations for us to address the kind 
of issues that have just been described?
    Mr. Norden. Well, one of the things I would say that would 
be very helpful, particularly in terms of problems with 
software that we have seen, is that there are often problems in 
aggregating votes from a number of machines--and this has 
occurred several times in the aggregation tally software. So 
one of the things is that there are additional kinds of audits 
that we can do and that many States do that will ensure that, 
number one, we are deterring any kind of attacks against the 
voting system but, more importantly, that we are detecting the 
kind of errors that might cause us to get wrong tallies.
    So, for instance, one of the things that they do in 
California and that they do in New York and several other 
States--Texas--is there is a reconciliation process to ensure 
that we look at the vote totals that come out of the precincts 
from the machines and compare them to the tally server totals, 
and make sure that what happened at the tally server, the 
totals of what the tally server are giving us, are correct. 
That is a check against the software there.
    Another is making sure that the number of voters who signed 
in to vote is relatively close to the number of voters who the 
machines are telling us who voted. In the end, it is 
unavoidable, given the amount of software that these machines 
run on and the many sources that they come from, that we are 
not going to sometimes have problems with the software. The 
key, I think, is to make sure that we are making the 
appropriate checks and reconciliations that are necessary.
    Ms. Lofgren. My time has expired, so I will turn to Mr. 
McCarthy.
    Mr. McCarthy. Well, thank you, Madam Chair.
    This has been a great hearing. I mean, I appreciate all of 
the input everyone has given.
    Doug Lewis, you are the executive director of the National 
Association of Election Officials. Is that what all of the 
secretaries of state belong to?
    Mr. Lewis. It is all of the election officials from 
townships, cities, counties, and States.
    Mr. McCarthy. Okay. I was reading your 8-page one here. It 
is very thorough. You are giving us some different ideas. My 
question is: Did you look at this based upon--you wrote a lot 
of this about H.R. 811. Did you look at any timeframe of what 
that would do to the Electoral College?
    Mr. Lewis. I think it is a legitimate question here.
    The truth of the matter is we saw in election 2000 that the 
electoral college had to meet before we could actually finish a 
recount. If we are going to add in and layer on an audit 
procedure that is a multiple audit procedure of multiple 
levels, I think we are really looking at some point in another 
close presidential election that we may not be able to actually 
get to it and get it done by the time that the electoral 
college is going to meet. So it seems to us that it is a 
question that you all have to think about. You have to look at 
this as to what your desires are and also recognize that you 
then make it impossible to get to the point that the public 
knows who won.
    Mr. McCarthy. Thank you.
    Pam Smith, I was reading yours, too, here. You said, on 
page 2, procedures should be defined governing what to do, in 
essence, saying that paper should be the final on the audit; is 
that correct?
    Ms. Smith. Yes.
    Mr. McCarthy. Now, were you here earlier for the last 
panel?
    Ms. Smith. Yes.
    Mr. McCarthy. We had an individual from Ohio who raised a 
very good point where he had the VVPATs, the DREs, and the 
paper did not come out to be the same as the computer; whereas, 
the computer came out with the right number of the number of 
people who voted, but the paper, knowing that we have people 
who are going and feeding the paper who do not do this every 
day for a living, got jammed. And the computer was right, and 
the paper was not, and his opinion was different than yours. I 
do not know if you could give me a little elaboration on yours.
    Ms. Smith. Sure.
    I think we have to think of this not just in terms of DREs 
and VVPAT-type paper trails but in general. You have a computer 
record that voters never get a chance to see, and you have a 
paper record that they can confirm is accurate. Now, in the 
case that he cited, his very smart poll workers were able to 
redirect the voters so they were voting on the machine so they 
could check the paper.
    Mr. McCarthy. So you would say they were right in what they 
did?
    Ms. Smith. Yes, that was absolutely right to redirect the 
voters so that they could have an opportunity to check on 
paper.
    Mr. McCarthy. And they were able to do that because the 
machine showed it?
    Ms. Smith. In that particular case. In some cases, the 
machines may not tell you, and the voter may not know. They may 
not have been advised that you really should check this paper 
record because this is the hard-copy record of your vote. Make 
sure it is accurate. That is something--that notification, I 
think, is part of that, H.R. 811. You need to--if you are 
expecting voters to use a new system they are not familiar 
with, you have to guide them somewhat.
    Mr. McCarthy. Should you mandate one over the other as 
being the final tally?
    Ms. Smith. Say again?
    Mr. McCarthy. Should you mandate paper or the computer to 
be the final tally, one over the other? I mean----
    Ms. Smith. I think you can only mandate to be the final 
arbiter of the voter intent the document that a voter has had a 
chance to look at.
    Mr. McCarthy. Okay.
    Ms. Smith. If they do not get a chance to look at it, which 
is true about computer bits and bytes in the ether, then they 
cannot confirm that that is right.
    Mr. McCarthy. Okay.
    Lawrence, you and Pamela talked about--was it Pamela? No, 
it was Candice. You talked about post-election--is that 
correct?--that voters do not feel that the elections are being 
held honestly or they have questions as to the honesty of the 
elections; is this right?
    Mr. Norden. I think, unfortunately, that there is certainly 
a large segment of the population that--and I think it is 
incredibly corrosive to our democratic system--do not have 
confidence in the results of elections.
    Mr. McCarthy. Just because my yellow light is on, if you 
have any polling on that, if you could show me--because the 
executive Director of the National Elections cites CNN says 88 
percent in the last one, but you looked at all of the phrases--
I guess Pam did. In all of the phrases of auditing, have we 
ever talked about when we are auditing this election on 
auditing who actually votes? Has anyone brought that up? 
Because it seems like, okay, we are auditing the people who 
voted, but have we ever checked from the very beginning that 
the people who did vote were the individuals who have the right 
to vote? If not, would you propose that in legislation for 
identification? Anybody? Yes.
    Ms. Hoke. There is a kind of reconciliation or audit 
process as far as checking to see whether the--at least in 
Ohio--about the number of people who signed in versus the 
number of votes. There have been some additional studies from 
time to time to ascertain whether people did have their IDs 
checked, and in Cuyahoga County, we did do a further study--I 
do not have the results on that, but I will try to obtain that 
data for you if you would like. This concerned whether people's 
IDs were properly checked and, therefore, valid voters. I will 
try to provide that.
    Mr. McCarthy. Because the only thing--I know my time is up. 
If it is true the voters do not believe whether it is true or 
not, if we are going to audit in the end, which in some way we 
do need to have checks and balances, we also need to check and 
balance who actually went there.
    Mr. Norden. Well, one point I would like to make, 
Congressman, is that--and a point I tried to make earlier is 
that, at this point, there is near universal agreement among 
computer scientists and security experts about the security 
vulnerabilities of these machines.
    Mr. Norden. I am certainly happy to have the Brennan Center 
come back and talk about allegations of individual voter fraud 
and issues of voter ID, but there are two things I can say 
about that issue for certain. One of the things we looked at in 
our security report is that the number of votes that could be 
affected by problems with the voting machines dwarf in 
comparison to the number of votes that could be affected by any 
kind of retail voter fraud that you are talking about. And the 
second is we do not have that same kind of consensus at this 
point about whether or not there are instances of voter fraud, 
individual voters who are not entitled to vote who are coming 
to vote.
    Ms. Lofgren. The time has expired.
    I would note that we have all been handed that report, 
which we will make part of the record.
    Ms. Lofgren. Now, Mr. Holt, it is your opportunity to ask 
some questions.
    Mr. Holt. Thank you, Chairwoman Lofgren and Ranking Member 
McCarthy. Thank you for allowing me to join you. Thank you for 
holding this hearing for the legislation H.R. 811.
    I know there has been a lot of talk about the security and 
accessibility of equipment and the openness of software and any 
number of other things, but the heart of the legislation is the 
auditability and the use of that auditability, the audit, 
itself. There are too many unresolved irregularities hanging 
out there that are really undermining our belief in our ability 
to govern ourselves, and as I often say to middle school 
students and others, self-government works only if you believe 
it does. I think there is a lot that has shaken that faith that 
we need to address, and so at the heart of the legislation are 
routine, random, independent audits, and this concept, 
certainly, has been endorsed by many organizations, including 
some represented here today, and I think the specifics have 
been examined. The specifics have been examined and, in many 
cases, endorsed as well.
    Ms. Hoke or Mr. Lewis, let me ask you: Do you think that 
the proprietary software printout at the end of the day after 
the polls are closed would satisfy your definition or a 
reasonable person's definition of an ``independent audit''--and 
``independent auditability''--I beg your pardon--and then 
whether that would, by itself, count as an audit?
    Ms. Hoke.
    Ms. Hoke. We took a position in Cuyahoga County that an 
audit would not be sufficient unless independent. Part of the 
reason--I mean, our audit report should be available probably 
within the next 10 days, and you will have some very important 
data that will show that it is important to get underneath the 
tabulation software's, shall we say, massage of the numbers to 
find out what the real raw data are. You may find that there 
are differences between the raw data and what is actually 
reported by the software program. It was our view that it is 
important to know that.
    Now, it may be that that is a software design problem that 
could be corrected, but certainly, our software engineers who 
are working with our audit felt quite strongly that this is one 
way to discover whether there is rogue code in the tabulation 
server or in some place other than the voting devices 
themselves that could cause the report of the election results 
not to match the real raw data, and so we very much wanted that 
raw data. And we did do a study of only three races, and we 
found that there were differences between the two election 
results tables.
    Mr. Holt. Mr. Lewis.
    Mr. Lewis. Congressman, I will say to you that there has 
not ever been at this point one verifiable circumstance in 
which anybody has been able to show that electronic voting has 
been manipulated.
    Mr. Holt. That is not what I asked.
    Mr. Lewis. I understand.
    Mr. Holt. That is not at all what I asked, because there 
are very many documented unresolved irregularities. I can list 
15 counties off the top of my head. So, clearly, what I asked 
is, does a software printout at the end of the day constitute 
auditability, in your mind?
    Mr. Lewis. If you look at the international standards that 
are created for computer memory in terms of storage--and all of 
these systems have at least a double redundancy and most a 
triple redundancy in terms of their memory and storage of 
ballot images that are retained within the system meeting the 
Federal voting systems guidelines. If you look at that, those 
standards are far higher than the standards for paper ballots. 
We can faithfully reproduce the votes as the voters voted them 
from the Ballot Image Retention. We can show you a ballot image 
of how every voter voted.
    Mr. Holt. Of how every voter voted or----
    Mr. Lewis. Ballot image retention is required by Federal 
voting systems guidelines, for all electronic equipment, and so 
you have double and triple redundancy there to where you can 
pull that out there. Yes, sir, I think you can show with great 
veracity that it is exactly what the voter voted.
    Mr. Holt. I think a key point is your statement that this 
shows how every voter voted, and I think you have missed, if I 
may say with respect, a basic point that every computer 
interface organization has addressed that the machine cannot 
verify itself. It is a fundamental principle with computer 
science.
    Ms. Lofgren. Our time has expired, and indeed, the panel 
has been very helpful, and we thank you for the time invested 
here with us. It is important. We will keep the record open for 
5 days. We may have additional questions for you, and--yes, 
sir.
    Mr. Lewis. I want to clear up, on page 8 of my testimony, a 
technical--I misstated one number incorrectly.
    Ms. Lofgren. If you will get it to us, we will make that 
correction, sir.
    Mr. Lewis. Okay. Thank you.
    Ms. Lofgren. Thank you, again, very much for the time you 
have spent and the expertise that you have shared with us.
    The hearing is adjourned.
    [Whereupon, at 4:35 p.m., the subcommittee was adjourned.]
    [Information follows:]
    [GRAPHIC] [TIFF OMITTED] 35804A.212
    
    [GRAPHIC] [TIFF OMITTED] 35804A.213
    
    [GRAPHIC] [TIFF OMITTED] 35804A.214
    
    [GRAPHIC] [TIFF OMITTED] 35804A.215
    
    [GRAPHIC] [TIFF OMITTED] 35804A.216
    
    [GRAPHIC] [TIFF OMITTED] 35804A.217
    
    [GRAPHIC] [TIFF OMITTED] 35804A.218
    
    [GRAPHIC] [TIFF OMITTED] 35804A.219
    
    [GRAPHIC] [TIFF OMITTED] 35804A.220
    
    [GRAPHIC] [TIFF OMITTED] 35804A.221
    
    [GRAPHIC] [TIFF OMITTED] 35804A.222
    
    [GRAPHIC] [TIFF OMITTED] 35804A.223
    
    [GRAPHIC] [TIFF OMITTED] 35804A.224
    
    [GRAPHIC] [TIFF OMITTED] 35804A.225
    
    [GRAPHIC] [TIFF OMITTED] 35804A.226
    
    [GRAPHIC] [TIFF OMITTED] 35804A.227
    
    [GRAPHIC] [TIFF OMITTED] 35804A.228
    
    [GRAPHIC] [TIFF OMITTED] 35804A.229
    
    [GRAPHIC] [TIFF OMITTED] 35804A.230
    
    [GRAPHIC] [TIFF OMITTED] 35804A.231
    
    [GRAPHIC] [TIFF OMITTED] 35804A.232
    
    [GRAPHIC] [TIFF OMITTED] 35804A.233
    
    [GRAPHIC] [TIFF OMITTED] 35804A.234
    
    [GRAPHIC] [TIFF OMITTED] 35804A.235
    
    [GRAPHIC] [TIFF OMITTED] 35804A.236
    
    [GRAPHIC] [TIFF OMITTED] 35804A.237
    
    [GRAPHIC] [TIFF OMITTED] 35804A.238
    
    [GRAPHIC] [TIFF OMITTED] 35804A.239
    
    [GRAPHIC] [TIFF OMITTED] 35804A.240
    
    [GRAPHIC] [TIFF OMITTED] 35804A.241
    
    [GRAPHIC] [TIFF OMITTED] 35804A.242
    
    [GRAPHIC] [TIFF OMITTED] 35804A.243
    
    [GRAPHIC] [TIFF OMITTED] 35804A.244
    
    [GRAPHIC] [TIFF OMITTED] 35804A.245
    
    [GRAPHIC] [TIFF OMITTED] 35804A.246
    
    [GRAPHIC] [TIFF OMITTED] 35804A.247
    
    [GRAPHIC] [TIFF OMITTED] 35804A.248
    
    [GRAPHIC] [TIFF OMITTED] 35804A.249
    
    [GRAPHIC] [TIFF OMITTED] 35804A.250
    
    [GRAPHIC] [TIFF OMITTED] 35804A.251
    
    [GRAPHIC] [TIFF OMITTED] 35804A.252
    
    [GRAPHIC] [TIFF OMITTED] 35804A.253
    
    [GRAPHIC] [TIFF OMITTED] 35804A.254
    
    [GRAPHIC] [TIFF OMITTED] 35804A.255
    
    [GRAPHIC] [TIFF OMITTED] 35804A.256
    
    [GRAPHIC] [TIFF OMITTED] 35804A.257
    
    [GRAPHIC] [TIFF OMITTED] 35804A.258
    
    [GRAPHIC] [TIFF OMITTED] 35804A.259
    
    [GRAPHIC] [TIFF OMITTED] 35804A.260
    
    [GRAPHIC] [TIFF OMITTED] 35804A.261
    
    [GRAPHIC] [TIFF OMITTED] 35804A.262
    
    [GRAPHIC] [TIFF OMITTED] 35804A.263
    
    [GRAPHIC] [TIFF OMITTED] 35804A.264
    
    [GRAPHIC] [TIFF OMITTED] 35804A.265
    
    [GRAPHIC] [TIFF OMITTED] 35804A.266
    
    [GRAPHIC] [TIFF OMITTED] 35804A.267
    
    [GRAPHIC] [TIFF OMITTED] 35804A.268
    
    [GRAPHIC] [TIFF OMITTED] 35804A.269
    
    [GRAPHIC] [TIFF OMITTED] 35804A.270
    
    [GRAPHIC] [TIFF OMITTED] 35804A.271
    
    [GRAPHIC] [TIFF OMITTED] 35804A.272
    
    [GRAPHIC] [TIFF OMITTED] 35804A.273
    
    [GRAPHIC] [TIFF OMITTED] 35804A.274
    
    [GRAPHIC] [TIFF OMITTED] 35804A.275
    
    [GRAPHIC] [TIFF OMITTED] 35804A.276
    
    [GRAPHIC] [TIFF OMITTED] 35804A.277
    
    [GRAPHIC] [TIFF OMITTED] 35804A.278
    
    [GRAPHIC] [TIFF OMITTED] 35804A.279
    
    [GRAPHIC] [TIFF OMITTED] 35804A.280
    
    [GRAPHIC] [TIFF OMITTED] 35804A.281
    
    [GRAPHIC] [TIFF OMITTED] 35804A.282
    
    [GRAPHIC] [TIFF OMITTED] 35804A.283
    
    [GRAPHIC] [TIFF OMITTED] 35804A.284
    
    [GRAPHIC] [TIFF OMITTED] 35804A.285
    
    [GRAPHIC] [TIFF OMITTED] 35804A.286
    
    [GRAPHIC] [TIFF OMITTED] 35804A.287
    
    [GRAPHIC] [TIFF OMITTED] 35804A.288
    
    [GRAPHIC] [TIFF OMITTED] 35804A.289
    
    [GRAPHIC] [TIFF OMITTED] 35804A.290
    
    [GRAPHIC] [TIFF OMITTED] 35804A.291
    
    [GRAPHIC] [TIFF OMITTED] 35804A.292
    
    [GRAPHIC] [TIFF OMITTED] 35804A.293
    
    [GRAPHIC] [TIFF OMITTED] 35804A.294
    
    [GRAPHIC] [TIFF OMITTED] 35804A.295
    
    [GRAPHIC] [TIFF OMITTED] 35804A.296
    
    [GRAPHIC] [TIFF OMITTED] 35804A.297
    
    [GRAPHIC] [TIFF OMITTED] 35804A.298
    
    [GRAPHIC] [TIFF OMITTED] 35804A.299
    
                   ELECTION REFORM: AUDITING THE VOTE
                            ELECTION REFORM:

                           AUDITING THE VOTE
