I've spent the past week looking over the voting equipment problems
captured by the Election Protection Coalition's 25 nationwide
call centers into the Our Vote Live database. There were
around 1900 such incidents in the database, although that number is
probably closer to 1700 taking into account duplicates.
Before I launch into the analysis, a few caveats:
This is voter-reported data, which means it can be inaccurate.
There has been no attempt to control for multiple reports from a
single precinct.
In many cases it is hard to tell what exactly happened as the
incident reports were taken by mostly non-technical legal volunteers
from mostly non-technical voters and volunteers in the field.
Given the unbelievable popularity of the 1-866-OUR-VOTE hotline, OVL
was unable to capture all incidents that people wanted to report.
There are undoubtedly incidents that were not reported to the OVL
hotline for a variety of reasons.
Bottom-line: This is useful for qualitative notions of what went
wrong on election day.
While I've worked to make this post accessible to an audience that
may not be familiar with the vagaries of voting technology, I just
don't have enough time to explain everything. In that sense, I
encourage you to ask questions () and I can
amend this document to clarify as needed.
The Big Picture
Somewhere close to 85-90% of all voting equipment incident reports
from the OVL database are very simple and report some combination of:
broken equipment, long lines, and/or emergency ballots being handed
out and/or auxiliary bins on optical scan systems being used (many
optical scan systems have a bin incorporated into the design of the
machine where ballots can be placed in the event the system ceases to
function or the power remains out for hours). If we can do anything
to improve the experience of the average voter facing a machine
problem, it should be reduce the amount of time they spend in line.
Another curious feature of the data is the voters' uniformly negative
attitudes toward contingency or back-up plans. Whenever the primary
mode of voting is affected, whatever the contingency plan to keep
people voting, it appears that voters are often upset and mistrustful.
They seem to think that there is a possibility that their vote will
not count if cast via a contingency plan. This is unfortunate as with
any critical technical system, there should be a contingency plan in
place that covers what to do if that system fails and how the system
may continue or recover from that failure. For example, when an
optical scanner goes down, the standard contingency plan is to place
ballots in an auxiliary bin or container. Voters reported concerns
with this kind of accommodation and were worried that their votes might
not count. In a number of cases, voters refused to place their
ballots into auxiliary bins and waited for hours before the scanner
could be serviced or a new replacement scanner brought in. It's clear
from this that we need to do a better job of educating voters as to
what to do and what to expect in case of emergencies.
Ok, enough of the big picture; now on with the details!
Broken Machines, Long Lines, Long Waits
As I said, the OVL database contains mostly complaints about broken
machinery, long lines, long waits to vote and reports of emergency
ballots being used instead of the normal mode of voting. These
incidents are so numerous that it doesn't really make sense to say
much more. However, there are some interesting features from these
reports.
Machine breakdowns and poll book bottlenecks lead to some voters
waiting in line for a long time. Machine breakdowns ranged from
simple problems to reports like one where all 15 voting machines had
stopped working in an Atlanta polling place (61728). In some
cases, the registration process caused the bottleneck leading to long
lines, including reports of only one pollbook for hundreds of voters
in New York City (68936) and e-pollbook problems in Georgia
and Maryland (48481, 86413). We have reports of
people waiting in line for 3 hours in New Jersey (63356), 3.5
hours in Georgia (55674), 5 hours in Ohio (21500), 6
hours in Missouri (82785) and a poor voter who fainted due to
exhaustion in NYC while waiting in line (83379). In many
cases, long lines were exacerbated by voters insisting on feeding
their own ballot into an optical scan machine, despite it taking a
long time to service or replace the affected equipment (Ohio:
52729). In a number of cases, precincts began to run out of
paper ballots (Virginia: 43073).
Problems with Emergency Ballots
Emergency ballots are paper ballots used in case the main voting
equipment in a polling places ceases to function properly. They're
typically placed in a sealed ballot box or in an auxiliary bin
incorporated into the design of an optical scan system. Some
jurisdictions in the past have not planned accordingly and either not
provided enough emergency ballots, so that polling places run out, or
not provided them at all. I was very encouraged to see that in most
cases, emergency ballots were available. However, what I didn't
count on was that voters consider voting via an emergency ballot to be
fundamentally suspect; that is, most were worried that their vote
wouldn't count if cast via emergency ballot. We also saw some
confusion between provisional ballots and emergency ballots, cases
where pollworkers were refusing to hand out emergency ballots and a
number of complaints about decreased voter privacy when emergency
ballots were being used.
Voters Uncomfortable with Using Emergency Ballots
Many voters seemed worried or uncomfortable casting emergency ballots.
In New York City, a caller reported that people were leaving rather
than casting emergency ballots (68315) and, as mentioned
above, there is evidence that some people refused to leave until they
could cast their ballot normally in the machine (Virginia:
43392, 45800). In one case from Ohio, a voter
reported that one of his ballot pages wouldn't scan and that he was
told by a pollworker to place his ballot in a folder labeled "Oh Crap"
with his name, phone number and address on it (47805).
We saw cases in at least two states where poll workers were refusing
to hand out emergency ballots despite significant machine failures.
Despite a ruling in Pennsylvania that required emergency
ballots to be given out when 50% of machines in a precinct were down,
OVL received a number of reports that pollworkers were not handing out
emergency ballots (e.g., 45436, 45545). In one case
a caller claimed that voters in line were "fighting with poll workers"
over emergency ballots (46401). There were also reports of
polling places in New York not offering emergency ballots
(43072, 43116, 43152). We also had a
number of cases from New Jersey where pollworkers were confusing
provisional ballots with emergency ballots (e.g., 43302)
Privacy Concerns with Emergency Ballots
Numerous voters called in to say that the privacy afforded to them
while filling out their ballot was severely impacted by emergency
balloting procedures. A pollworker in California decided to take over
the role of the broken optical scanner and manually review the marking
of each ballot, presumably looking for overvotes and undervotes
(57719). Voters in Ohio and Pennsylvania reported feeling
very uncomfortable filling out their emergency ballots as others in
the polling place could see how they voted (44091,
52199). Finally, the length of the lines in Michigan lead to
a shortage of privacy sleeves, without which voters were not being
allowed to cast their ballots (47325, 48018,
49671, 53785). (A privacy sleeve is essentially a
legal-sized Manila envelope, about $1 apiece at Staples.)
Problems with Disability Access Equipment
One of the big promises of the Help America Vote Act of 2002 (HAVA)
was that voters with disabilities would finally be able to cast votes
privately and independently. Unfortunately, while all jurisdictions
in the US now have disabled access technology in their polling places
(or some other accommodation), this promise has fallen short.
In many cases, the disability access equipment simply didn't work or
was not set up properly on election day (AZ: 54531; CA:
82162, 82751, 93171; NY: 43261,
88576; MO: 84713; MN: 92806).
Pollworkers frequently did not know how to use these systems (NY:
45978, 43372, 46992; MD: 49457).
And despite laws like HAVA, we still see reports of basic
accessibility problems such as building and polling place
accessibility (51922, 84713, 88353),
complicated ballots (85857) and voters with serious
conditions such as walkers and diabetes having to wait in long lines
(49457, 88512)... the diabetic left without voting,
by the way.
One particular case warrants special attention (88483). In
this case, a blind voter found that the machine was not working. When
the voter brought this to the attention of the poll workers at her
polling place, they "complained about trying to make machine operable
when there was a line and no one else needed to use it." This is
truly terrible: a poll worker complaining about doing their job and
doing so in a way that appears intended to make the disabled voter
feel guilty about how their disability affects the polling place.
This pollworker at Long Lane Court in Delaware County, Pennsylvania
should be ashamed.
Finally, in New York there was an interesting case where a voter of
Haitian descent had trouble figuring out what was going on when the
machines broke down due to language issues (92905). This
case gives a hint at what voters with limited English proficiency face
when things go wrong.
Machine Malfunctions, Fixes
In the realm of machine problems, the data shows a variety of
interesting things, from questionable technical fixes to broken and
unstable machinery to unexpected behavior from machines.
Improper Technical Fixes
The most serious cases concerning machine malfunction involved what I
call improper technical fixes or inappropriate servicing of equipment.
In South Carolina, individuals removed a voting machine from the
polling place and took it out to a car to tinker with it
(89533). In New York, there were three striking cases of
inappropriate service. A man who used to be a poll worker or election
technician, but was no longer, was resetting the lever machines in one
precinct after each vote (55057). In another case, a police
officer reportedly fixed a machine (70424). Finally, a
voting machine needed to be "reset" and the pollworkers required one
voter to "waste" their vote as it was reset with the voters ballot
still activated but not cast (88142). In all of these cases,
it is hard to tell if the person "fixing" the machines is an
authorized agent of the state or local election official.
Possible Broken Lights/Buttons, Levers, Etc.
Voting machines, like all machines, break (in fact, another way of
stating the 2nd Law of Thermodynamics is simply: "Everything breaks").
The OVL data has numerous reports of lights and buttons not working on
machines in New Jersey and Pennsylvania (NJ: 46008,
56364, 59330, 85342, 87092; PA:
66514, 69144, 80924). In New York, we saw
many reports of broken levers and broken lights that should be
illuminating the lever ballot (NY: 61117, 62013,
46658, 43649, 45963, 84727). In
one case, a voter tried to cast a write-in vote, but the write-in
wasn't working; when the voter was allowed to use a different machine,
she reports voting for someone else on the ballot because she was
afraid that using the write-in slot on the machine would cause the new
machine to break (70962).
Optical Scanners Not Incrementing Ballot Card Count
When a voter inserts an optical scan ballot into an optical scanner,
usually a serial number counter increments to show that a new ballot
has been scanned. In a number of cases---from Ohio, Virginia,
Minnesota, Texas and North Carolina---we see reports of voters saying
that the number did not increment when they inserted the ballot (OH:
70443, 84413; VA: 46089, 58330,
93670; MN: 67120; TX: 65937; NC:
59196, 61940). What might cause this? Typically,
if a ballot page has an overvote or write-in, that page will be
redirected (or "diverted") to a smaller, separate ballot box in the
machine for ballots that need human inspection. Pollworkers usually
have to manually hit an "override" switch to allow these ballots to be
fed in and the counter does not increment in these cases. In many
jurisdictions, these ballots are examined and then "remade" to
correctly reflect the voter's intent (as long as they can tell what
the voter intended to do).
Machine Stability
In a number of cases, machines appeared to be unstable or simply
didn't work as they were supposed to. The data shows machines that
keep rebooting (60554), frozen or hung machines
(43698), machines that shut down (63059), machines
that would work only after periodic shaking and kicking
(43430), problems with networked eSlate voting machines
(57580) and a machine where the optical scan memory card was
not working with the machine it was suppose to work with
(81997).
Ballot Navigation, Missing Races
We saw a number of issues with voters having difficulty understanding
how to navigate a digital ballot, including "fleeing" voters,
"premature" voters, missing races and issues with the user interface
for voting machines.
"Fleeing voters" and "Premature voters"
A "fleeing voter" is a voter who leaves a voting machine without
having cast their voted ballot. A "premature voter" is one who
accidentally casts their ballot (or has it cast for them) before they
are finished voting their ballot. I have typically associated both of
these kinds of voters with first-generation DRE voting systems like
the Sequoia AVC Advantage and the Danaher ELECTronic
1242; both these systems have large "VOTE" buttons that are
easily missed by some voters---"fleeing" voters---but that are
activated and can be pressed anywhere during the voting
session---"premature" voters.
In terms of reports of user-interface issues, we saw a few cases where
the iVotronic interacted strangely or unexpectedly when voting a
straight-party option (56027, 57497). There was
confusion where a voter saw arrows next to her choices as well as the
"VOTE" option (58627). In one case the voter reports hitting
the "cast ballot" button and the machine responded that her vote was
canceled (66279). Finally, in Washington DC, a voter
reported that the review screen on a DRE had reported that he hadn't
voted in all races, but when he went back in the ballot to check, he
found that he had voted in all races (87927).
Missing Races
A seemingly abnormal amount of voters called in to report missing
races off their ballot (e.g., 62077). In Virginia, these
came in the form of ballots where the Presidential race was the only
contest available (45852, 47895, 49416,
55472, 56160) and vice-versa---ballots where the
Presidential race was the only race missing (45157,
45272, 48455, 49819, 53575,
55802, 57468). There were also reports of
incomplete ballots where the presidential race was missing in
Pennsylvania on the AVC Edge and iVotronic DREs respectively
(59545, 63224).
Paper Record Problems
There were a number of problems reported with voter verified paper
record (VVPR) printers, optical scan machines and bleeding pens.
In Ohio, three concerning problems with VVPR printers on the Premier
AccuVote-TSx cropped up. Voters reported seeing blank paper
when they went to check the VVPR (58266), a paper jam
affecting 36/50 ballots (84679) and a problem where the last
voter's VVPR was left exposed to the next voter (93742). In
other states, voters reported a non-printing paper trail in Missouri
and Illinois (88648, 90876) and a paper trail
printer error in California (63150).
In Virginia, we saw problems reported with wet ballots gumming up the
works of optical scan machines (92623, 90490). In
one polling place in California, a voter reported their optical scan
machine was "shredding" ballots (61864).
In a few cases, bad marking instruments were reportedly being used
with optical scan ballots: felt-tip pens or permanent "magic" markers
that can bleed through ballot paper. Felt tip pens were reportedly
being used in California and Florida (61899, 65847)
and in a few cases affecting races on the opposite side
(60169, 61636).
"Vote-Flipping"
Reports of "vote-flipping" were significant. "Vote-flipping" is where
a voter attempts to register a vote for one choice on the ballot only
to have a different choice selected. I'm not a big fan of this
terminology as I think it inappropriately lumps a bunch of design
problems, technical problems and voter behaviors into one over-broad
category. Regardless of what I think, voters reported it often.
Few voters understand what I think the dominant cause of vote-flipping
to be: touchscreen miscalibration. However, one voter from Chicago
recognized a miscalibration and had the pollworker recalibrate the
touchscreen (51219). Other interesting cases of note
include: one case of republican to democratic party vote-flipping
(87474), a couple of cases from Mississippi and Missouri where
the error was only noticed on the paper trail (70899,
69731) and a couple cases from Mississippi and Georgia of
vote flipping in a contest other than US President (70606,
84842).
We saw a number of reports of poor chain of custody (control) over
either ballots or voting machines or both. In California, these two
problems were mixed in one polling place when a pollworker decided to
periodically empty the emergency/auxiliary bin and place emergency
ballots in the main bin (88942). Not only is this way too
much handling of the paper ballots, but the seal on the door to the
main ballot bin was compromised.
When ballot boxes got full or auxiliary bins began to overflow, or
simply when there was no ballot-box-like container in which to place
ballots, we saw reports of poor ballot custody. Ballots in New York
were "just laying around" (85609), in Pennsylvania and
California in plastic bags (48028, 70442,
70670), in Florida in duffel bags, on the floor and in
folders (58347, 68859, 67857), stacked on
top of a machine in Minnesota (62835) and into an unsecured
and open ballot box in Indiana (45556).
There were reports of unsealed machines in New York and Pennsylvania
(56995, 47935, 52090).
Problems with Pollworkers Related to Machines
Reports of problems involving pollworkers were numerous. Staffing and
training a pollworker force can be a challenge and occasionally
problems with pollworkers can effect the voter's interaction with
voting technologies.
Improper Advice or Improper Casting of Ballots
In a few cases, pollworkers intentionally or mistakenly cast a voter's
ballot before they are finished voting or before they've had a chance
to revise their ballot. A pollworker in New York prematurely cast a
voter's lever machine ballot (85508). When a voter was
asking for instructions on how to vote in Virginia, a pollworker cast
the voter's blank ballot (48003). Pollworkers in Illinois
appeared to be improperly overriding optical scan ballots and casting
a ballot before a voter was finished (56655, 68580).
Finally, pollworkers in Arizona were also reported to be improperly
overriding overvoted ballots (86848).
Other Pollworker Machine Effects
There were a few other interesting pollworker-related cases. A lead
pollworker in Ohio suddenly got sick and had to be taken away in an
ambulance; this cast the polling place into chaos when the other
pollworkers were unsure how to compensate (63881). In
Pennsylvania, a voter reported a problem with a voting machine and the
pollworker requested that the voter swear on the Bible to the veracity
of her claims (58922). In one case, unrelated to voting
technology, a pollworker was forcing people to sign their names in the
pollbook next to names that were not the voters' (46764). In
Ohio, pollworkers were handing out the incorrect ballot style for a
precinct for a while (44347) and in Florida, one voter even
received a partially filled-out ballot (87761).
Miscellaneous
In the proverbial "kitchen-sink" category, I should mention a few
other interesting reports:
In Pennsylvania, a caller reported that the instructions in the
local newspaper for voting on the touchscreens showed a touchscreen
that said "No" and "Yes and No", but the voter just wanted to vote
"No" (25526).
A pollworker from Georgia called in to ask for the password to their
voting machine so that they could delete a ballot (58275).
Voters in Ohio and Pennsylvania noticed that their voting machines
were still on Daylight Savings Time and were concerned that their
votes (or some votes, at least) would not count due to this
(47027, 86539).