The country's poll body and statistics experts said that a seeming irregularity regarding the margin of difference between the votes of frontrunner Ferdinand "Bongbong" Marcos Jr. and Vice President Leni Robredo is not anomalous and may just be basic math.
While election returns from different parts of the Philippines were being transmitted into the Commission on Election's transparency server, social media users observed the difference in votes between frontrunner Marcos Jr. and Robredo remained consistent with every update.
The constant 68:32 pattern or 47% ratio had netizens thinking: are these consistent numbers normal? Or is this evidence of possible voter fraud?
same observation from previous retweets: consistent yung 67-68% increase ni babyem per update ng server while 32-33% kay VPL. is this normal?
sorry, wala na ako sa tamang disposition to further analyze these numbers. kayo na bahala. i'm just putting these data out here. pic.twitter.com/xSSaVCpAwe— 4SIS4LENI 🌸 (@4sis4leni) May 9, 2022
As of writing, the camp of Marcos Jr. have claimed victory in the elections after securing over 31 million of the votes, which is more than double that of Robredo's.
'Hard to prove'
On May 10, Comelec said that allegations of cheating will be "hard to prove" based on the percentage.
"Mahirap i-prove yan, pag sinabi mong may cheating with regards to that percentage," commissioner Marlon Casquejo said, adding that their system underwent certification.
"If they really insist, pwede naman natin i-check yung election returns in each province," he continued.
Additionally, the Parish Pastoral Council for Responsible Voting (PPCRV) stated they have not seen anything of concern.
"We've seen those reports. We've been receiving the same information. Right now, it's no cause of concern but the PPCRV is looking into it," the poll watchdog's spokesperson Atty. Vann Dela Cruz told reporters on May 10.
He also said that the PPCRV is coordinating with the poll body regarding the matter to verify if there are any anomalies. "We've actually communicated a request from Comelec for system logs from the server," Dela Cruz said.
In a separate interview, the poll watchdog's chairperson Myla Villanueva echoed that no irregularities in the alleged pattern have been seen. The group has partnered with experts from Ateneo de Manila University and the University of Santo Tomas to review the data.
"At the moment, since their time this afternoon [May 10], they have reported no irregularities in that pattern. Let's give them a little more time," she said.
Likewise, the National citizens' Movement for Free Elections (NAMFREL) is not seeing any anomaly with the 68:32 ratio.
"It’s consistent with what people say about surveys, it’s consistent with what the survey standings are," secretary-general Eric Alva said.
What do statistics experts say about the alleged pattern?
OCTA Research fellow and mathematician Guido David told GMA News that the consistent 47% ratio claim is mathematically probable and not evidence of possible cheating.
“Kung per region, mag-iiba talaga yung pattern ng numbers. So ito, may consistency kasi sabay-sabay silang pumapasok from different parts of the country. And dahil sabay-sabay, more or less, ang distribution niya is the same," he said.
OCTA Research Tugon ng Masa April 22 to 25, 2022 Presidential Preference Survey Results compared with Comelec unofficial tally (more than 98% election returns). pic.twitter.com/0n4hYuEoUi— Dr. Guido David (@iamguidodavid) May 11, 2022
For his part, University of the Philippines associate professor Dr. Peter Cayton told One News PH that it's "difficult to discern" whether the changes in voting could be anomalous.
"When we look at the aggregate counts and the change of the number of votes with each transmission, mahirap kasing ma-discern kung yung changes in the voting are something that could be anomalous," he said.
"It’s because it is a confluence of a lot of things happening. But when you aggregate them, the differences, the variations fade away because there’s the law of large numbers working for us.”
I think this work shows that in the new added votes for Marcos and Robredo, it is not a constant stream.— Peter Cayton, the Stats Guy (@PJACaytonPhD) May 10, 2022
Marcos ranges from 66.2-68.7%
Robredo varies around 31.3-33.8%.
The trends and new votes are not evidence for cheating. The already large numbers drowns variations. https://t.co/v793uJ43Pl
The law of large numbers states that a given sample, especially a small one, is not guaranteed to reflect population characteristics.
Instead, Cayton suggested investigating precincts for possible anomalies.
“If we are looking maybe looking out for possible discrepancies or flaws or anomalies, it is best to [not look at] the aggregate of total votes. The total votes would drown out any possible changes, variations, flaws that comes from the individual precincts. Oftentimes, we either look at the precincts or how the voters have been interacted with before they get in the precinct," he said.
"A lot of possible anomalies could be caused with the voter being affected by for example vote-buying or at the precincts some level of tampering before they get into what we see as the total or aggregate counts,” Cayton added.