Electronic Voting Problems
Electronic voting machines are mysterious black boxes. When
you cast your vote on a direct-recording electronic (DRE) voting
machine,
the ballot appears on a display screen and votes are captured
and stored electronically. You cannot see that the vote you intended
to cast was recorded correctly inside the machine and you cannot
watch votes being added up inside the machine. Such
systems violate the principles of openness and transparency in
democratic elections.
A Common Cause report, “Malfunction
and Malfeasance: A Report on the Electronic Voting Machine Debacle,” includes
Indiana among 17 states are at high risk of having election results
compromised
due to the use of DRE voting machines that do not provide voter-verified
paper ballots in many counties and due to the lack of automatic
random audits in those counties that do provide paper ballots.
(Click
here to see the type of voting system your county uses.)
What could go wrong? Computer information security
specialists agree that no electronic voting system is infallible.
The Brennan
Center report, “The
Machinery of Democracy: Protecting Elections in an Electronic
World,” says, “Computerized voting
systems… have shown themselves vulnerable to error….
votes have been miscounted or lost as a result of defective firmware
(coded instructions in a computer system’s hardware), faulty
machine software, defective tally server software, election programming
errors, machine breakdowns, malfunctioning input devices, and pollworker
error.” The Brennan report also describes how one person
could secretly program the machines to throw an election to a
particular candidate, and election officials and voting machine vendors
might not ever know.
Is there evidence that voting machine errors have occurred
in Indiana? Yes. In Franklin County, in the 2004 elections,
straight party Democratic votes were counted for the Libertarian
candidates on
DRE voting machines. The result of one race was overturned when
the error was corrected. The error was detected because the Libertarian
candidate had an unexpectedly high number of votes – if the
error had instead counted straight party Democratic votes for the
Republican candidates (or vice versa) it might not have been detected.
In Pennsylvania, on the same DREs used in Monroe County, Indiana,
votes were lost when power surges caused the machines to go into
power-fail mode. In two races in North Carolina on the same equipment,
errors in tallying software resulted in incorrect vote totals – the
result of one race was overturned when machine totals were added
by hand.
Computer information security experts understand
that all technology is vulnerable to attack. A hacker
could secretly program the machines to throw an election to a
particular candidate
and honest election officials and voting machine vendors might
not ever know. In fact, the Brennan
Center Report indicates, ominously,
that one person could reverse the results of a statewide
election - overcoming efforts of the vendor to prevent such an
attack, eluding independent testing authority, avoiding detection
during
pre-election
machine testing, and avoiding detection through records kept
on event and audit logs.
How do we need to change Indiana voting systems
to make them secure and reliable?
---------------------------------------------------------------------
E-Voting Problems:
Many Different Systems, Different States,
Different Elections
Sources: VotersUnite.org: “Facts
About Electronic Elections” & "Vote
Switching Provided by Vendors” and Common Cause: “Election
Reform: Malfunction
and Malfeasance”
1. New Elections Needed after Electronic Voting Failures.
• A memory limitation on paperless Unilect Patriot voting machines
caused 4,438 votes to be permanently lost in North Carolina (2004).
• AVS WINVote computers at some polling places failed to start
up, others overheated and broke down during the election in Mississippi
(2003).
2. “Phantom” Votes Added by Electronic Voting
Machines.
• After the 2004 General Election, phantom votes (more votes than
voters) were reported in Florida, Nebraska, New Mexico, Ohio,
South Carolina, North Carolina, and Washington. In North Carolina
Microvote DREs showed nearly 3,000 more votes than voters; in
New Mexico Sequoia AVC Edge DREs showed over 2,700 early voting
phantom votes.
• Hart Intercivic machines in Tarrant County, Texas recorded an
additional 100,000 votes that were never actually cast in 2006
Primary Election (Formal challenge to results has been filed).
3. Bugs!
•
ES&S vote-tallying software counted to 32,767 and then counted
backwards in November 2004 elections: 70,000 votes temporarily
disappeared in Broward County, Florida; 8,400 votes in Orange
County, Florida; and 22,000 votes in North Carolina.
4. Votes Jump to the Opponent on the Screen
• In the November 2004, hundreds of votes jumped from Kerry to
Bush in New Mexico (Sequoia), Maryland (Diebold) and elsewhere.
In Monroe County, several voters reported to a poll worker that
when they pushed the button for Bush, the Kerry light came on.
(Personal
communication.)
5. DRE’s Present Incorrect Ballots to Voters
• In March 2004, the US Senate contest in Maryland was omitted
from ballots in three counties.
6. Negative Votes Added to Tally
•
In Volusia County, Florida in 2000, Al Gore’s count dropped
by 16,022 votes, while an obscure Socialist candidate picked
up 10,000 votes at 10:30 PM on election night. Global Election
Systems explained that two memory cards had been uploaded; there
should have only been one memory card uploaded; the second card
caused the problem. (BlackBoxVoting.org).
7. DREs Pass Pre-Election Testing, Fail on Election Day
• In Mercer County, Pennsylvania all 250 UniLect Patriot machines
had been checked and rechecked. On election day some machines
never operated, some offered only black screens.
8. Programming Errors Give Votes to the Wrong Candidate
•
In the 2006 Pottawattamie County, Iowa primary when vendor ES&S
incorrectly programmed the ballots, the early electronic vote
tally showed a popular 23-year Republican incumbent losing to
a 19-year-old college student. Highly suspicious, the auditors
stopped the electronic count and started counting by hand. The
hand count revealed that the incumbent had won by a large margin.
9. Voting Machines Present a Default Candidate (Electronic Version
of a Pre-Marked Ballot)
•
In Travis County (Austin) Texas, Hart Intercivic eSlate DREs
operated such that voters who voted straight party Democratic
ticket and then pressed ‘enter’ on the next screen
caused their Kerry/Edwards vote to be changed to the default
candidate -> Bush/Cheney.
10. Voting Machines Do Not Count Some Votes
• Voters claimed that machines failed to register votes for incumbent
school-board member, Rita S. Thompson ( R ), who lost an election
in Fairfax, Virginia by 1,662 votes. In post-election testing,
election officials observed that one of the questionable machines
appeared to subtract a vote from Thompson for about one out of
every 100 attempts to vote for her.
E-Voting Problems in Indiana
(All material from VoteTrustUSA & VotersUnite.org)
•
Franklin County, Indiana (2004). Straight Democratic Party votes
were awarded to the Libertarian candidate on Fidlar direct-recording
electronic voting machines. The result of one race was overturned
when the program was corrected.
•
In March 2004, it was discovered that ES&S had installed
an uncertified version of firmware in the iVotronic machines
in four counties. The older certified version did not tabulate
votes correctly so the uncertified version was used in the election.
Marion County Clerk Doris Anne Sadler is quoted as saying that
ES&S “has willfully and purposely deceived me and the
Marion County election board…”
•
In April 2006, ES&S was supposed to have Johnson County iVotronic
machines set-up and in use for “early-voting”. However,
contrary to their contract, and in violation of state law, ES&S
failed to program the counties’ ‘smart cards’ with
ballot definitions. Also missing: Paper ballots for Johnson and
nearly two-dozen other counties around the state.
•
In 2006, Indianapolis-based Microvote had installed uncertified
software in the voting machines in 47 counties. They filed for
federal certification just two weeks before the primary election.
Tom Wheeler, chairman of the IEC, said, “The most disturbing
thing I heard here was these guys knew it wasn’t certified
and they went out and installed it. That can’t be tolerated.”
In September 2006, Wheeler discovered that in order to get Infinity
software certified for the May 2006 primary election, Microvote
had disabled the straight-party voting function. As of early
September, Microvote had not informed County Clerks of this situation
and was seeking approval from the State to install new software
for the General Election. The new version of the software has
now been approved for installation. Citizen confidence in the
certification process and Microvote corporation is sinking. (Wheeler's
9/13/06 letter as PDF).
E-Voting Problems with Microvote
Please note: The information
below is intended to illustrate that no voting system is perfect
and that is why we need VVPBs and MMRAs. All
information is from “Microvote
In the News” posted at VotersUnite.org.
•
Montgomery County, Pennsylvania, 1995/96. The scroll motors emitted
power surges that caused the machines to go into power-fail mode
and shut down to protect the circuitry. The DREs shut down, causing
the current voter’s vote to be lost.
•
Putnam County, Tennessee, August 2002. “As Putnam County
officials learned last night, when computers fail, the electoral
process slows to the pace of a 20th-century hand-held calculators.
That’s how counters there had to proceed when a glitch
caused the computers to produce wrong totals. “Nothing
you have is good,” Administrator of Elections Nancy Boman
explained to a reporter. After first reporting 20 of 44 precincts
were in, Boman said officials noticed there were problems. “A
write-in candidate received 1,000 votes, and we knew that it
just didn’t sound right, so we started looking closer.
The computer had shifted all the numbers down a line. Nothing
was right,” she said.
•
Boone County, Indiana, November 2003. Tabulation software failed
to properly initialize (zero out) the vote totals before starting
to tabulate the votes from the data cartridges causing more votes
than voters to be initially reported.
•
Jasper County, North Carolina, June 2004. Results from the central
tabulator showed candidate Jones in the lead with 1,661 votes
to Gregory’s 1,470. When the County printed tapes from
the individual machines and added the results by hand, Gregory
bettered the incumbent, 1,139 votes to 1,079 votes. In one precinct,
there was a race in which all three candidates received exactly
111 votes.
•
Mecklenburg, North Carolina. Unofficial results from the central
tabulator showed more votes than voters. Adding results from
individual machine tapes by hand changed vote counts.
