Let’s design an automated election system…together



Part 5
I MENTIONED in my last article that shortly before the 2013 elections, a group of Filipino IT practitioners, all sufficiently knowledgeable about election systems, did a cost/benefit study of what could be a “most appropriate” automated system for the Philippines. Included in the study were the Direct Recording Electronic System (DRE), Optical Mark Recognition system (OMR, specifically PCOS), Central Count Optical Scan (CCOS), the Hybrid System (HB), and the Pure Manual system (PM). The major criteria used in the comparison were transparency, accuracy, vulnerability to cheating, cost, speed, auditability and verifiability, training and staffing requirements, and machine reliability.

To review the descriptions of each of the above-mentioned systems:

1. DRE, sometimes referred to as the “touch-screen system,” is one where the voter indicates his choices by tapping the names and/or photographs of the candidates that appear on a computer monitor.

2. OMR, under which falls PCOS, or Precinct Count Optical Scan, is where voters make their choices through pre-printed ballots, simply by shading the spaces opposite the names of the candidates.

3. CCOS is the same as PCOS, except that the uncounted ballots are brought to a central counting center, where they are fed into a high-speed OMR machine for counting and printing by precinct.

4. In HB, voters signify their choices by writing down the names of the candidates they want to vote for, in ordinary ballots. After the voting period, the ballots are counted manually. Henceforth, the automated canvassing will proceed similarly as in the DRE and OMR systems.

5. All voters are of course familiar with the pure manual system, where no machines, except perhaps calculators, are used.

I can no longer find a copy of that study, so I did the next best thing. I tried to reconstruct the matrix that we built then. I still remember the logic that was used in creating that matrix. Following is the result of the analysis, using a scale of 1 to 5, where 5 is the most favorable:

Transparency: DRE-1, OMR-2, CCOS-2, HB-5, PM-3
Accuracy: DRE-4, OMR-4, CCOS-4, HB-5, PM-3
Vulnerability: DRE-2, OMR-2, CCOS-2, HB-5, PM-3
Cost: DRE-1, OMR-3, CCOS-4, HB-4, PM-5
Speed: DRE-5, OMR-5, CCOS-4, HB-4, PM-1
Auditability: DRE-3, OMR-3, CCOS-3, HB-4, PM-4
Training/staff: DRE-2, OMR-3, CCOS-3, HB-4, PM-5
Machine rel.: DRE-3, OMR-3, CCOS-3, HB-4, PM-5
Totals: DRE-21, OMR-25, CCOS-25, HB-35, PM-29

For this study, the above criteria were given equal weights. If the more major ones (the first six) were, however, to be given higher weights, the margin of HB over the others could even be wider.

Some notes regarding the above rating:
Transparency – Transparency is lost in any system that automates precinct-counting. OMR and CCOS are slightly better than DRE because they, at least, have paper ballots. HB is manual precinct-counting and therefore completely transparent. With PM, while precinct-counting is transparent, such is lost because the weeks-long canvassing is also manual andwatchers do not stay till the end.

Accuracy – As experienced, even in the United States, automated counting has been found to have many inaccuracies, either because of software bugs, internal tampering, loose system controls, and other similar shortcomings. In addition, OMR has threshold problems. Disenfranchisement occurs when ovals are improperly marked. HB is a “5” because, precisely, manual counting is the basis of accuracy. While PM uses manual precinct-counting, it also uses manual canvassing, which is so vulnerable to cheating.

Vulnerabilities – DRE, OMR, and CCOS all feature automated precinct-counting. Cheating can easily be programmed into these systems. To unsuspecting election administrators, such will remain undetected.

Cost – HB uses only PCs and laptops and is therefore less expensive than the other automated options.These machines can be passed on to the Department of Education after each election, thus cost is shared, while also saving warehousing and repair costs.

Speed – CCOS is slower than DRE and OMR because the filled-up ballots will have to be physically brought to the city/municipal centers for counting (thus also exposing them to substitutions). HB which features manual precinct-counting will be slower by about five hours. That’s really all, especially with an improved manual counting system.

Auditability – In case of protests, HB’s and PM’s Election Returns are a second “trusted” document, having been produced through precinct-counting witnessed by the voters. They can therefore be used as bases for such protests.

Staffing – It’s easier to find staff who are familiar with PCs and laptops than those familiar with automated election systems. PCs are also slightly more reliable, as machines, and repair facilities are available almost anywhere. With the pure manual system, on the other hand, no additional staffing and training would be necessary.

As expected, the Hybrid System came out far, far ahead of the other technologies. It’s small wonder therefore that it is the most popular election system worldwide. Our Commission on Elections should go through this kind of exercise before making decisions on the technology to use.


Please follow our commenting guidelines.

Comments are closed.