Murphy’s Law and the Voter Interface of the Automated Election System

Introduction

In my experience with introducing and institutionalizing new technology at the National Statistics Office (NSO), I always had to contend with Murphy’s Law.  In dealing with information and communications technology, I have learned that everything has to be perfect for it to work like a human being at top speed.  A system has to be 100 percent accurate for it to work; less than 100 percent, what you have are highly expensive and decorative electronic equipment that serve no one and occupy office space.

Murphy’s law is an adage or epigram that is typically stated as: “Anything that can go wrong will go wrong”.  It is used as either a purely sarcastic musing that things always go wrong, or, less frequently, a reflection of the mathematical idea that, given a sufficiently long time, an event which is possible (non-zero probability) will almost surely take place; although, in this case, emphasis is put on the possible bad occurrences. http://en.wikipedia.org/wiki/Murphy’s_law

Front End à Voter Interface

I intend to discuss the possible pitfalls awaiting the automated election system (AES) to be implemented in the May 2010 elections.  But I will not deal with the contract, the source code, the transmission and all other matters that have been extensively discussed and argued.  I write these concerns arising from the front end process of the AES not to spite COMELEC but to alert them of mostly seemingly trivial matters but which could spell big trouble on Election Day with no immediate on-the-spot solutions.

1. Functional Literacy of Voters

To begin with, COMELEC has a rather big challenge, – functional literacy.  At least 15 percent of the population 10-64 years old is not functionally literate. A functionally literate person is one who can read, write and compute or one who can read, write, compute and comprehend[i].   Table 1 below shows the functional literacy rates among Filipinos, aged 10-64, by region and sex.  Admittedly these percentages include those aged 10-17 who are not voters and exclude those aged 65 years or more who are voters.  I am not sure if these groups would pull the rate up or down or cancel each other but nevertheless a sizeable amount of voters may not be capable of filling out the scan-able ballot correctly.  Assuming 80 percent of the 50 million registered voters will vote, thus 40 million. a good 6 million, or 15 percent, will have difficulty with the ‘new’ ballot.

Table 1. Functional Literacy of Filipinos Aged 10-64, 2003

Region and Sex Male & Female Male Female
PHILIPPINES 84.1 81.9 86.3
NATIONAL CAPITAL REGION 94.6 94 95.2
CORDILLERA ADMINISTRATIVE REGION 85.4 83.9 87
I – ILOCOS 88.6 88.1 89.2
II – CAGAYAN VALLEY 84.4 82.9 86.1
III – CENTRAL LUZON 86.9 86.5 87.4
IVA – CALABARZON 90.4 88.8 92
IVB – MIMAROPA 82.3 80.2 84.4
V – BICOL 80.1 76.6 83.8
VI  – WESTERN VISAYAS 81.5 77.7 85.2
VII – CENTRAL VISAYAS 81.7 79.8 83.6
VIII – EASTERN VISAYAS 76.7 71.7 82.1
IX – ZAMBOANGA PENINSULA 74.8 69.8 79.8
X – NORTHERN MINDANAO 83.7 80.5 86.9
XI – DAVAO 77.8 73.7 82.2
XII – SOCCSKSARGEN 77.1 74.5 79.7
XIII – CARAGA 81 77.3 84.6
ARMM 62.9 63.6 62.1

For more statistics please refer to http://www.census.gov.ph/data/sectordata/2003/fl03tabE.htm .

Voters’ education should not be approached in a general catch-all strategy of training and orientation since functional illiteracy is not uniformly or equally distributed among the population.

2. The Ballot

The size of the ballot, according to COMELEC, is “8 ½ inches by 26 inches [lately trimmed down to 25 inches] and some 50 million.   The ballots would be color-coded and printed back-to-back with the names of the candidates. The national candidates will be found on the front, while the local candidates will be on the back.

In a recent resolution, the Commission en banc listed 10 candidates for president, eight for vice president, 61 for senator, and 187 party-list groups, the names of which will appear in the ballots in acronym form.  The names of the local candidates would be unique for every local legislative district in the country since voters in each area would be choosing provincial officials with 80 different sets and city or municipal officials with 1631 different sets.

OMR technology is used to detect shaded ovals or boxes that correspond to answers (or votes for a particular candidate). But shadings should completely cover each distinct oval or box and should not stray to the adjacent ones; otherwise the machine will reject the form or ballot.  A more serious error is incorrect printing such that all ovals are read as filled, even if this is not the case. This happens if the outline of the ovals is too thick, or is irregular, as was the case of 19 thousand absentee ballots in the county of Gwinnett, Georgia during the 2008 U.S. presidential election. This was discovered during a test run made in late October and after around 10 thousand had already been returned. The slight difference was not apparent to the naked eye. For more details, see http://en.wikipedia.org/wiki/Optical_answer_sheet.

The printing of 50 million ballots with more than 1600 varying candidate sets [combination of national and local candidates] is a momentous challenge even for the National Printing Office.  Can you imagine the distribution system that would deliver the right ballot forms to the right legislative districts, with no room for error?  99 percent accuracy in the delivery of ballots is not acceptable since voters in one district or municipality cannot vote for their candidates in ballots intended for another place.  In past elections, the only challenge for COMELEC is the number of ballots to be delivered to municipalities since there was only ONE form for the whole country.  For the May 2010 it is at least 1600 times more complicated.  During the past population censuses that I directed, the forms were only in English, even if theory dictated the use of dominant dialects in census taking.  Distributing the right forms to the right places was a logistical nightmare.  The following examples show the challenge of producing and handling the numerous ballots to be filled up by voters.

>>In Metro Manila, the following table shows the distribution scheme required; each district will have its own ballot.  In the past elections there was only ONE ballot for the whole country.  In Metro Manila alone, there would be 40 different ballots.  Fortunately if there would be errors in the delivery of ballots, or ‘switching’, within-the- day ‘corrective’ deliveries can be carried out.

Table 2. Registered Voters, as of 17 March 2009

No. of Registered Voters City Districts
CALOOCAN CITY
429,899 1ST DISTRICT
206,356 2ND DISTRICT
LAS PIÑAS CITY
177,092 1ST DISTRICT
162,993 2ND DISTRICT
MAKATI CITY
206,741 1ST DISTRICT
180,849 2ND DISTRICT
MALABON CITY
87,808 1ST DISTRICT
84,178 2ND DISTRICT
MANDALUYONG CITY
114,095 1ST DISTRICT
89,403 2ND DISTRICT
MANILA
208,779 1ST DISTRICT
134,474 2ND DISTRICT
155,708 3RD DISTRICT
152,670 4TH DISTRICT
212,330 5TH DISTRICT
156,183 6TH DISTRICT
MARIKINA CITY
94,626 1ST DISTRICT
110,418 2ND DISTRICT
MUNTINLUPA CITY
132,416 1ST DISTRICT
110,426 2ND DISTRICT
126,659 NAVOTAS CITY
68,328 1ST DISTRICT
58,331 2ND DISTRICT
PARANAQUE CITY
119,497 1ST DISTRICT
152,445 2ND DISTRICT
PASAY CITY
154,630 1ST DISTRICT
130,656 2ND DISTRICT
PASIG CITY
147,240 1ST DISTRICT
221,412 2ND DISTRICT
PATEROS
22,017 1ST DISTRICT
17,843 2ND DISTRICT
QUEZON CITY
205,245 1ST DISTRICT
535,432 2ND DISTRICT
145,570 3RD DISTRICT
205,909 4TH DISTRICT
SAN JUAN CITY
38,369 1ST DISTRICT
38,678 2ND DISTRICT
TAGUIG CITY
98,774 1ST DISTRICT
171,957 2ND DISTRICT
VALENZUELA CITY
119,861 1ST DISTRICT
140,068 2ND DISTRICT

>>In another instance, note that there are 6 San Fernandos in the Philippines: La Union, Pampanga, Romblon, Camarines Sur, Masbate. Bukidnon.  There are 9 San Joses: Dinagat Islands, Negros Oriental, Antique, Camarines Sur, Mindoro Occidental, Romblon, Batangas, Nueva Ecija and Tarlac. There may be similar examples of ‘favorite’ names for cities or municipalities.  The procedure for labelling of area-specific ballot shipments should be standardized.

>>The case of Batanes is also interesting.  There would be different ballots for each of its 6 municipalities, with 4 having registered voters of about 1 thousand each.  Compare these with the 535 thousand registered voters in the second district of Quezon City and the 430 thousand in the first district of Caloocan City.

No. of Registered Voters Municipalities
BATANES
4,019 BASCO
1,669 ITBAYAT
858 IVANA
1,059 MAHATAO
1,016 SABTANG
910 UYUGAN

Perhaps COMELEC is aware of all these and will come up with the efficient system to send the right ballots to the right province/municipality/city at the right time.

3. Filling Up the Ballot

The length of the ballot (25 or 26 inches) exceeds the length of a broadsheet/newspaper (22 inches).  Have you tried reading a newspaper without crumpling it on the desk in a public school classroom?  However, COMELEC Metro Manila regional director Michael Dioneda advised voters: “Take good care of the ballot. Be sure it doesn’t get crumpled, wet or smudged because every voter will only have one chance to get an official ballot…..”  Following COMELEC guidelines, a ballot that is rejected twice by the Precinct Count Optical Scan machine will be considered spoiled. http://politics.inquirer.net/view.php?db=1&article=20100201-250627

Once a filled-up ballot is considered spoiled because the PCOS rejected it twice, do you think the voter will go away meekly and leave it to the heavens for his vote to be acknowledge?  And if the PCOS jams up due to dirt, heat, short of the ideal conditions in which it can operate, will the voters in the precinct accept this sitting down?.

Since 2000 the National Statistics Office has added a step before feeding forms to the scanners; it calls this process as grooming. Creases are straightened out; folds are sometimes ironed [pinaplantsa], forms are dusted.  After all it would be a great waste to throw away a filled up questionnaire.  Or in this case, a citizen’s choice subordinated by a finicky machine.

COMELEC should likewise be prepared to groom the ballot for the PCOS rather than declare a filled up ballot as spoiled.

Voters will be asked to shade ovals placed beside the candidates’ names, which would be printed in 10-point font.  This font size may be difficult for the senior citizen voters. Some guidelines on desk top publishing point to the inadequacy of the COMELEC-selected font size.

  Choose a font size that is easy to read.
The younger and the older your target audience, the larger the type size should be for comfortable reading. Getting older doesn’t necessarily mean poor eyesight for everyone; however, declining vision is a common side effect of aging.

  Use larger font sizes for most older audiences.
For a predominantly older readership of 65 and over or for audiences with known visual handicaps, set body text in sizes from 14 to 18 points. Avoid tiny print.

  Use an 11 to 12 point type size for readers in the 40-65 age range.

http://desktoppub.about.com/cs/finetypography/ht/type_size.htm

To win the bidding for AES2010, SMARTMATIC presented to COMELEC a certification from the New York State Election Board that it had successfully used Dominion Voting System (DVS). http://www.vote-ny.com/english/transcripts-sequoia/ The COMELEC accepted the certification in awarding the automation project to Smartmatic.  Is PCOS the same as the DVS?

Obviously not.  The Dominion system shown in the link is voter friendly.  They even give the voter an audio way of verification (presumably for blind voters).  This system is nicely complete, in contrast with the truncated system offered by Smartmatic [in which the voter has no chance to see and verify his choices before casting his ballot] in the contract with COMELEC.  http://foolawecon.wordpress.com/2009/08/20/the-ballot-goes-blowing-in-the-wind-where-are-you-bob-dylan/

COMELEC should make available magnifiers for voters who are vision-impaired.

I understand too that a felt-tipped pen will be used by the voters to shade the ovals.  During storage and transport the ink supply of the pens might dry up or be limited due to the intense April and May heat and humidity.  Shading of the ovals has been identified as an important concern during the briefings. 

COMELEC should determine how long the pens can be used after taking off the protective cap in the harshest of conditions and ensure the availability and functioning of these pens during Election Day.

4. Not the First Time

No one gets it right the first time.  But COMELEC says it is not the first time.  Let us go back to history.

ARMM had been used as a testing ground for computerized elections.  In September 9, 1996, 42 optical scanners ($ 15,000 each), from Nebraska-based American Information System (AIS) were subject of the Senate Committee experiment led by Miriam Santiago. It found the machines defective.  Again in the May 11, 1998 ARMM elections, the same machines were tested and were found defective, resulting to manual counting next day.

0n July 22, 2008, however, the COMELEC, using these machines, conducted a successful “mock poll” and simulated the actual processes and procedure for the August 11 ARMM election.

The Philippines’ first-ever electronic voting (automated) polls had some 1.6 million registered voters, who elected a governor, vice governor and regional legislators for the ARMM. The COMELEC used 3,300 Smartmatic-Sahi Direct Recording Electronic (DRE) voting machine in Maguindanao and 156 “automated counting machines” (Optical Mark Readers from AVANTE International Technology, Inc.) in Basilan, Sulu, Tawi-Tawi, Shariff Kabunsuan, and Lanao del Sur.

DRE or “touch screen electronic voting,” allows voters to “simply touch the pictures of candidates they wish to elect,” and were especially for disabled and illiterate voters.

“No writing of names” is the principal feature of OMR machines. Voters had to simply shade the circles beside the candidates’ names. …A 24-digit randomly generated bar code identified the ballot as emanating from a specific precinct, and the OMR was programmed to count only ballots from designated precincts.

The final official COMELEC canvass of the 11 August 2008 Autonomous Region in Muslim Mindanao general election results showed Zaldy Ampatuan capturing 93 percent of the votes and his 6 other opponents sharing 7 percent. http://en.wikipedia.org/wiki/Autonomous_Region_in_Muslim_Mindanao_general_election,_2008

It is my opinion that the ARMM experience does not satisfy the basis to go all out and implement automated elections nationwide.  Moreover there is an interesting legal case that arose from the earlier ARMM automation exercise.

5. Legal precedent on the automated elections

In the Loong v. Comelec case (April 14, 1999) cited by former Chief Justice Panganiban, the COMELEC ordered the stoppage of the automated counting in the entire province of Sulu when the machine did not count three ballots wherein the COMELEC officials noticed that the oval space opposite the name of a candidate in Pata, Sulu (a part of the ARMM) was shaded but the ovals were not properly aligned to the candidate’s printed name.    As a result, all the ballots cast in the entire province were brought to the COMELEC headquarters in Manila for manual counting.

Chief Justice Panganiban dissented from the majority decision in Loong v. Comelec to uphold the COMELEC’s manual count of the automated ballots.  He cited as his main reason that, “there were no rules on how to manually count electronic ballots. The rules on manual counting could not be used in appreciating automated ballots…They were good, for instance, in determining whether a ballot was written by one hand or two hands, but not in appreciating the voter’s intent via the mere shading of pre-printed ballots.”

With this legal precedent and all the other points presented above, and all the issues and concerns raised during the past months, should the COMELEC not seriously consider a parallel manual count?

6. Parallel Manual Count

COMELEC is preparing for a manual count in 30 percent of precincts in case the PCOS fails.  It posits that this is in preparation for, not an acceptance of that the machines will fail.  http://politics.inquirer.net/view.php?db=1&article=20100128-249863

However when the staunchest automation advocate prepares for a 30 percent failure rate, it can be deduced that COMELEC can only be 100 percent sure for the automated election system, not only the machines but including the voter interface, the polling precinct, the transmission, the software and everything else, to function according to plans in 70 percent of the precincts.

When it says manual count, does it mean that votes are counted by hand at the precinct level, then precinct returns canvassed at the municipal level, municipal returns at the provincial level, and then added up at the COMELEC in Manila?  Who decides that there will be a manual count?  In which of the precincts will there be a manual count system prepared?  When can the decision for a manual count be made?  Can a decision to use a manual count be made on Election Day?  Will the ‘non-automated election system’ ballots and all other forms be used?  Are these being printed now?

I attended a Harvard Forum where I did get the declaration from the COMELEC person that when all else fails, there is a ‘continuity’ provision in the law that provides for a manual count.  And since neither the COMELEC nor anyone can foresee in which precincts the PCOS will fail, it has to prepare for manual counts in ALL precincts.

Why not have the teachers count the filled-up ballots at the close of voting on Election Day? Their counts can be compared with the count from the PCOS. If the counts are the same, or if different by not more than a COMELEC-set percentage, then the results can be transmitted onward.  Otherwise, the COMELEC can decide what to do with this discrepancy.  Would you not agree that the manual count will be the more credible figure in this case?

For consideration of the COMELEC – A Summary

Voters’ education should not be approached in a general catch-all strategy of training and orientation since functional illiteracy is not uniformly or equally distributed among the population.

The procedure for labeling of area-specific ballot shipments should be standardized.  Perhaps COMELEC is aware of all these and has come up with the efficient system to send the right ballots to the right province/municipality/city at the right time.

COMELEC should likewise be prepared to groom the ballot for the PCOS rather than declare a filled up ballot as spoiled.

COMELEC should make available magnifiers for voters who are vision-impaired.

With the legal precedent and all the other points presented above, and all the issues and concerns raised during the past months, should the COMELEC not seriously consider a parallel manual count?

Why not have the teachers count the filled-up ballots at the close of voting on Election Day?


[i] In the 2003 FLEMMS, a self-administered functional literacy questionnaire designed by the National Statistics Office, the Literacy Coordinating Council and the Department of Education was accomplished by persons 10 to 64 years old in order to determine their literacy status. Literacy was categorized into four levels based on the individual responses in the self-administered questionnaire. Persons who cannot read and write are considered illiterate.  Persons who can only read and write are considered basically literate (Level 1). Persons who can read, write and compute (Level 2) and persons who can read, write, compute and comprehend (Level 3) are considered as functionally literate. Hence, Persons who graduated from high school or completed a higher level of education were automatically considered functionally literate in the tabulations.  For more information see http://www.census.gov.ph/data/sectordata/fl03_lsf.html

Advertisements

3 Comments

Filed under election, Philippines

3 responses to “Murphy’s Law and the Voter Interface of the Automated Election System

  1. monetarist

    At this point of the game, may I ask, ahhh, what is your proposed alternative, Lagman’s Open Election System or back to the manual system? Nothing is perfect in this world. That said, let’s just go with what we’ve got, which I believe you too, would say is much better than the previous manual system. Yes, Murphy’s Law will happen this coming elections but I would rather be hopeful that what we get is an acceptable and credible performance and outcome of the automated system. As a statistician, I know you will understand that. As to your wishlist on a perfect system, I also wish that our government would have unlimited funds to pay for all these but unfortunately, they don’t. Peace out.

    • makuhari

      Yes, it is too late for Lagman’s OES. COMELEC though has to prepare for the previous manual system or a variant since the law says that when all else fails, it has to resort to the manual count. I have also posed a question to a few lawyers as to what should a voter do in case the PCOS machine rejects his/er ballot twice and the BEI declares his ballot as ‘spoiled’. It would be ironic that everyone around could read his vote (an exercise of his right of suffrage), except the PCOS which is the only thing that matters and ‘counts’. Thanks for your comments.

  2. I like your suggestion that the teachers would count the ballots. This makes the 2010 elections a giant pilot test, which is what it should perhaps be to begin with. It shouldn’t be that hard to do, especially if you do it for a representative sample of precincts (say, 10 percent). It could be separate from or part of the so-called random manual audit.

    Of course, there’s this pesky thing of pay. Is there a budget? I believe so, since the automated elections appropriation was an add-on to the default appropriation for manual elections.

    So is it a matter of “political will”?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s