Repair Records for ES&S InkaVotePlus Precinct Ballot Readers

for the Nov. 2010 Election, Los Angeles County

Judy Alter, Protect California Ballots

 

Rationale for this study:

            In early 2009 I studied and analyzed the poll observers’ reports of the Nov. 2008 election collected by volunteers for Work the Vote.  In those 201 reports the observers reported that 13% of the InkaVotePlus Precinct Ballot Readers (pbrs) failed to work or broke down in the precincts they observed.  Almost one in eight is a high rate of failure.  Since the number of poll observers’ reports represented such a small part of the total number of precincts in LA County I decided to study the coordinators’ reports of all the precincts for the Nov. 2010 election to understand the breakdown issue in a more accurate manner

            The InkaVotePlus Precinct Ballot Readers provide voters “second chance voting”:  a way to check their ballots for over-votes – voting for two instead of one candidate in a contest – and to tell a voter that his/her ballot is blank, caused by a dry pen or misaligned ballot in the vote recorder into which the voter slips a ballot when marking his/her ballot. 

            The taxpayers paid $25 million from HAVA funds in 2007 for these pbrs.  This money enabled the county to acquire more than 4780 (the approximate number of precincts in LA County) of these electronic scanners. I say “more than 4780” since, in this study I discovered that replacement pbrs are available during election day for the coordinators to swap out/replace broken ones for a pbr that should work. 

Sources for this report:

1.      The coordinators’ reports from the Nov. 2010 election.

2.      The exchange sheets for the replaced InkaVotePlus precinct ballot readers and the audio ballot booths, (ABB), signed by the coordinators at the distribution/collection sites.

3.      The Maintenance Referral Reports from LA County for April, June, and November 2010.

I will summarize and compare the information from each of these sources although the numbers in them do not match; thus, though informative, in this analysis I can only generalize about the many reasons these pbrs breakdown.

 

#1.  The coordinators’ reports from the Nov. 2010 election

            Coordinators travel on election day among their 7-17 precincts at least three times to make sure that everything at each poll is set up and running, that the voting is going smoothly, and that the poll workers are closing the polls correctly in a timely manner.  I read reports that the coordinators filled out as they travel from one precinct to another. In reading these reports, I saw that the coordinators do their jobs in very different ways; their use of those forms demonstrates this wide variation.  The total number of coordinators reports is 375; 14 were missing and 4 contained no information beyond the number and location of the assigned precincts. This study includes 357 reports.  260 of them included reports of precinct ballot readers’ breakdowns/replacements (73%) and 99 stated no breakdowns, (28%).

            Each coordinator inspects between 7 to 17 precincts. (I only recorded the number of their precincts for 178 reports (50%).) The most frequent number of precincts each coordinator must inspect is 14 (46), 13 (45), and 12 (34). 

            When coordinators bring a malfunctioning unit from one of their precincts to their particular distribution/collection site to exchange a precinct ballot reader or ABB, they receive an exchange sheet.  These sheets are pink, blue, or yellow.  The coordinators attach the units’ exchange sheets to their packets of reports. In some of the reports they attached no sheets and in others the coordinator did not note a precinct ballot reader or ABB breakdown or replacement but included these sheets.

 

Breakdowns or Replacements from the Coordinators’ Reports

1.      Coordinators indicated that 32 times the precinct ballot readers were fixed at the precincts by unjamming or rebooting them.  They also indicated that the readers stopped and then started again on their own.

2.      241 precinct ballot readers were replaced once.  13 were replaced twice and one was replaced three times.  21 pbrs broke down and were not replaced: total 291.  Problems included:  not starting, no zero tape, screen smashed, it froze, no header card, card stuck, chord broken, printed blank paper, and did not accept ballots.

3.      The Audio Booths (ABB) were replaced 17 times and in 5 cases they did not work and were not replaced: total 22.

 

#2.  The Exchange Sheets for the replaced InkaVotePlus precinct ballot readers and the audio ballot booths.

            Exchange sheets record that each inspector picked up his/her precinct supplies. The page has a place for the seal from each precinct ballot reader and a place for the seal of the audio ballot booth assigned to that poll inspector. When the inspector returns the equipment to the collection site at the end of election night, the election official receiving the supplies and equipment puts a second seal just below the first one on that page showing that the same or another piece of equipment was returned.

1.      The number of sheets showing exchanges for the precinct ballot readers: 175 and in addition, 10 had no exchange or return label.

2.      The number of sheets showing exchanges for the ABBs: 42 and, in addition, one had no return or exchange label.

3.      The number of sheets showing exchanges for both the precinct ballot reader and the ABB: 16; one had no exchange or return label.

4.      Four sheets showed the assigned seals but had no return or exchange labels.     

The sequence of numbers on the replacement labels end with an “R” or “T” and sequence of numbers on the distribution labels (those on the seal of the assigned precinct ballot readers all end in “P.”

A few of the sheets differed from the others:  one had a return label on top of the distribution label instead of below it; one sheet included a note that the label was missing; one had a note, “had machine that was broken; received a 2nd machine”; one had the labels in the wrong places.

 

                                    Comparison of the Coordinators’ Reports and the Exchange Sheets

                                                            Replaced/broken pbrs            Replaced ABBs

 Coordinators’ Reports            241+26+3+21=291              17+5=      22

Exchange Sheets                        175+10+16=    201              42+1+16=59

 

#3. The Maintenance Referral Reports from LA County RR/CC

            Below is a list of the problems identified on the Maintenance Referral reports I reviewed from a digital list received from LA County RR/CC.  I studied the reports for the April 13 special election in the 43rd Congressional District (total 47), the June Primary (total 947), and the Nov. 2010 Guberna­torial Election (total 304).  Many more problems were repaired after the June primary; many fewer repairs were made after the November elections, perhaps because they repaired many of the problems of the precinct ballot readers that occurred in June.  I list the percentage of the total repairs after the raw numbers.

 

 

 

 

 

Problem

April 2010

June 2010

Nov 2010

Category

Frequency

47

947

304

 

handle

19 /  40%

118 /  13%

1

hardware

replace battery

0

243   26%

122   40%

electronic

missing pin

2   4%

123   13%

0

hardware

reader

4  .9%

80   8%

61   20%

electronic

printer

8   2%

15   2%

16   5%

electronic

touch-screen

1

48   5%

18   6%

electronic

monitor

2

222   23%

33   11%

electronic

ABB port

4

15   2%

2

electronic

ABB touchpad

0

11  1%

0

electronic

machine froze

0

2

3

electronic

white screen

0

1

0

electronic

machine hang

0

1

0

electronic

no election mode

0

1

0

computer

power-up failed

0

2

0

electronic

case cracked

0

2

3

hardware

hard drive upside down

0

1

0

electronic

keyboard not work

1

1

0

electronic

not close

0

1

0

hardware

not start

0

1

0

electronic

bad Ethernet port

2

4

6   2%

electronic

slow load

0

2

3

computer

boot failed

0

9   .1%

16   5%

computer

no diagnostic code

0

0

1

computer

thumb drive door not close

0

0

1

hardware

missing rivets

0

0

2

hardware

not load password screen

0

0

6   2%

computer

security seal missing

0

1

0

computer

error code 999

0

1

1

computer

error code 346

0

11

0

computer

error code 906

0

2

4   1%

computer

error code 939

0

0

1

computer

error code 905

0

0

1

computer

error code 349

0

0

1

computer

 

 

            The problems fit into these broad categories: the physical structure (hardware) of the readers:  the wheels, feet, handle, case, pins, rivets, lock; the electronic parts (electronic) of the computers:  the screen, keyboard, printer, reader, monitor, battery, ABB port and touchpad, the Ethernet port and thumb-drive ports; the functioning of the computer (computer): failing to power up (start), reboot, slow load, not get diagnostic code or load password, and six error codes. The need to replace the batteries was the most frequent problem in June and November and the monitor and reader are the next most frequent problems (which may be the same).

 

Observations

            Although the sources for this analysis of the problems with the ES&S InkaVotePlus Precinct Ballot Readers, pbrs, vary considerably and my ability to study some of the sources depended on different circumstances, I can safely conclude that these machines show numerous structural/hardware 44%, electronic 23%, and computer 32% problems:  total problems 33.  The   pbrs require ongoing maintenance and oversight especially for their batteries.  The maintenance reports revealed that these precinct ballot readers have Internet capacity that may be connected during an election, not legal in California—seen in the need for repair of the Ethernet and thumb-drive ports. It is not clear if these functions are necessary for the interconnection of the precinct ballot reader with the ABB unit.

             In my public records request to get copies and access to these documents, I asked for any analysis of these reports done by staff; I received none.  When I reviewed the coordinators’ reports, it appeared to me that no one had studied them.  If the reports had been read, then, perhaps, a few pink, yellow, and blue sheets would not have been attached to the wrong packets of these reports. The exchange sheets for the pbrs and ABB units show 28% fewer pbrs exchanged than do the coordinators’ reports, and, in contrast, they show that more than twice the number for the ABB units were exchanged. I believe proper repair and maintenance of these electronic machines require greater oversight and coordination than is evident from these sources.

 

Further Questions

            This report triggers further questions: 

¨      Who has the maintenance contract to repair and maintain the electronic equipment and how much does that cost the County? 

¨      Is the routine maintenance done by RR/CC mechanics as well as specialized outsourced maintenance?

¨       For what are the batteries used, how much do they cost, and are they rechargeable?

¨      LA County was one of the counties in California that failed the recent HAVA audit because the county employees could not find and pull out of storage a randomly selected number of these pbrs.  Record keeping is part of proper maintenance.

¨      Poll worker training needs to include some warning about the potential for this equipment to have problems.  Perhaps the Inspectors who pick up the equipment should inspect it for at least structural problems. 

 

            Finally many experienced poll workers have been able to prevent over-voting from occurring at their precincts by interacting with voters: asking them to inspect their ballots before casting them in the ballot box.  This electronic equipment, especially the precinct ballot reader, may not really be necessary.

 

Final note:

            On March 16, 2012 Tim McNamara arranged a meeting with me and the RR/CC staff members who work with the coordinators to discuss my findings from only the Coordinators’ reports.  At that meeting I learned that one staff member receives all the telephone calls on election day from coordinators about their equipment problems and breakdowns.  I did not know about that document when I did this study.  That document may help clarify the actual numbers of breakdowns of these pbrs and ABBs.  I believe it is important for at least one staff member to review all these records to assess how well the equipment functions during an election.