MANILA, Philippines - It may have been high-speed, but was it accurate?
A report to be released by the poll watchdog National Citizen’s Movement for Free Elections (NAMFREL) showed that some Precinct Count Optical Scan (PCOS) machines failed the accuracy test.
It may have no bearing on the national races, particularly the presidential race, but the questionable accuracy may have had an impact in hotly-contested races in the local front, and in the some cases, neck-and-neck races in the last spots for the Senate.
The Commission on Elections (Comelec) has specified that the minimum accuracy rate for the China-made PCOS machine should be 99.995% or a maximum error rate of .005%. PCOS machines that failed to meet this standard should have been replaced as stated in the contract.
The poll body is supposed to test the accuracy of the PCOS in the final sealing and testing a few days before the elections. But the mess caused by the wrong configuration in the memory cards a week before the polls has sidelined any honest-to-goodness monitoring whether the accuracy of the PCOS was followed in the final test run.
As the Comelec has not released any report on the final testing and sealing, there is also no way of knowing how many machines that failed the accuracy test were actually used in the May 10 polls.
For the May 10 polls, the Comelec leased 82,200 PCOS from supplier Smartmatic-Total Information Management for the 76,000 clustered precincts nationwide. The rest of the PCOS machines were intended to be used as back-up.
Accuracy issues
NAMFREL executive committee member Damaso Magbual told abs-cbnnews.com/Newsbreak that the group will present its initial findings this Friday on their observations as to the accuracy of the PCOS.
The report will indicate that there were accuracy issues on the PCOS.
For instance, in Precinct 1212 at the UP Integrated School in Quezon City, the PCOS counted 602 ballots. But the random audit showed that there were 612 ballots cast.
“This is more than the maximum margin or error of .005%,” Magbual observed.
In that precinct, the PCOS also failed in the accuracy test in the counting of the votes cast. Quezon City Mayor Feliciano Sonny Belmonte, who is seeking a congressional post in the 2nd district, got 501 votes in the PCOS count, but the election return showed that he got 498.
Another congressional candidate, Don Emil de Castro, got 95 votes in the PCOS count. But the ER showed he actually got 102 votes.
In one precinct in San Juan City, the PCOS counted 616 ballots but the manual audit showed only 614 ballots were cast.
Still, in another precinct, the PCOS counted 607 ballots, but the manual audit showed there were 613 ballots cast.
‘Not alarming’
Magbual said the discrepancies in the PCOS count “may not be alarming,” but “this does not mean you should sacrifice accuracy.”
Asked if there was a pattern of inaccuracies on how the PCOS machines functioned, he clarified that “we are not yet seeing a pattern.”
Namfrel is conducting its own manual audit using ERs shared with other groups. The group sought accreditation to conduct the random manual audit but was blocked by its erstwhile ally, Parish Pastoral Council for Responsible Voting (PPCRV)
Namfrel had wanted to expand coverage of the random manual audit from 1 precinct for every legislative district to 5 precincts. As Comelec denied Namfrel’s accreditation, the poll body passed that responsibility to the PPCRV.
The PPCRV is not expected to release the report of its random manual audit anytime soon.
