Manual audit shows May polls 99.97% accurate – Comelec
The random manual audit for the May 13 elections proved that the precinct count optical scan (PCOS) machines were 99.97 percent accurate, Commission on Elections (Comelec) Chairman Sixto Brillantes Jr. said yesterday.
This was the report of the Random Manual Audit Committee (RMAC).
“With 99.97 percent, it is practically perfect. It’s not perfect but it’s definitely accurate. This validates the results of the 2013 elections,” Brillantes told a press conference that followed the release of the RMAC report.
The RMAC, which includes the Comelec, Parish Pastoral Council for Responsible Voting ( PPCRV) and National
Statistics Office ( NSO), was convened by the poll body to do a manual count of 1,016,860 votes from 212 of the 234 clustered precincts chosen randomly by the committee.
According to RMAC chair Henrietta de Villa, they were not able to include 22 ballot boxes due to various reasons.
Four of the ballot boxes were subject of an electoral protest while one was allegedly tampered.
De Villa said the audit took 43 days and involved 1,522 individuals, including 330 personnel from the Comelec, 1,170 teachers from the Department of Education, 21 employees from the NSO, and one from PPCRV.
“The RMA, in many ways, is the result of painstaking investigative work. Matching the automated and manual count – two parts of the same process: elections – will seldom, if ever, be 100 percent the same. Variances or discrepancies will always occur,” claimed De Villa, also the chair of the PPCRV.
She said a “plausible explanation could be that machines – in the automated election system, the PCOS – will only read what it is programmed to read while the manual auditor is moved by human appreciation of what might be the intent
of the voter or human error.”
The results
In her presentation, NSO chair Carmelita Ericta reported that covered by the random manual audit were the votes for senators, congressmen, and mayors.
The votes for senators got the highest accuracy rate of 99.9775 percent after only 184 variance or discrepancy from automated and manual counts were found in 817,729 votes.
Twenty-seven variances, on the other hand, were monitored in 96,155 votes audited for congressmen, leading to 99.9719 percent accuracy rate.
The report showed that an accuracy rate of 99.9748 percent was reached in the votes for mayors as only 26 variances were found in the 102,976 votes.
Ericta said that based on 99 percent accuracy rate, the allowable margin of variance was set at one percent, which is equivalent to less than 10 votes’ dierence in absolute value for every 1,000 votes counted.
An NSO report showed that “since most of the ‘variance’ can be attributed to human errors or clerical errors, aiming for a higher accuracy rate to as high as 99.995 percent (i.e. one vote dierence in absolute value for every 20,000 valid votes counted) could be statistically improbable.”
“I think speaking from experience because we also underwent automation at the NSO, we have a similar case where the threshold was so tight for scanning – when we digitize the birth certificates. It was totally impossible to achieve,” Ericta added.
Based on the manual audit conducted by RMAC on the results of the country’s first automated polls in 2010, the accuracy rate was 99.96 percent.
According to Comelec Commissioner Christian Robert Lim, Republic Act No. 9369 or the poll automation law does not specify any target for accuracy rating, contrary to perception that the requirement was 99.995 for the PCOS machines to be considered accurate.
“It was under the request for proposal of the bidding documents ( for PCOS machines that the accuracy rate should) be 99.995. It was never said in RA 9369. It was just in the bidding document in 2009,” he said.
Lim added it was impossible to achieve a 100-percent accuracy rate unless the Comelec “cooks” the audit.
But despite the accuracy rating of the PCOS machines, Brillantes does not expect the critics of the Comelec and the automated election system to be pacified.
“The critics will always look for the negative. So I think they will remain critics for the next 100 years. We don’t silence the critics. That is their category: as critics, they will always look for the faults and negatives,” he said.
“We cannot resolve the problems with the critics. That is their problem, not ours,” Brillantes maintained.