(UPDATE) If the 2022 elections were held today and votes were cast through Facebook reactions, the only son and namesake of the late dictator Ferdinand Marcos would easily win the presidential race as he led the online polls of two mainstream media outlets.
Manila Bulletin kept its poll online to “get the pulse of the nation at a particular moment in time,” while Rappler took down its survey, citing a surge in reactions from accounts with foreign-sounding names voting for Marcos and his rival, Vice President Leni Robredo.
Former Sen. Bongbong Marcos denied that his presidential bid was backed by an army of trolls.
Opinion polls, done carefully, give a reliable reading of the public’s stand on issues, and in the case of election surveys, determine the candidates preferred by voters.
“The problem with social media election polling is that it does not take advantage of statistical techniques to make sure na yung results ng poll would be representative of the population of Filipino voters,” political scientist and WR Numero Research head Cleve Arguelles told reportr.
“The results therefore are only true for those who participated in the survey. That means the data that you generate from those results can't really be generalized for the entire population of Filipino voters,” he added.
Social media polls vs. old school surveys
Election surveys require different statistical techniques to make their results reflective of the opinion of all Filipino voters.
While traditional election polls like the ones by Social Weather Stations and Pulse Asia use samples to equally represent each member of a population, social media surveys do not use the same approach, University of the Philippines political science professor Alicor Panao said.
“Social media polls conducted via Facebook and Twitter typically utilize a non-probability approach to solicit responses,” Panao said, which does not accurately represent all members of a population.
Anyone can also simply participate in social media polls as long as they have a Facebook account or access to the internet, and this means that not every population is represented proportionately in the survey, he added.
But using a non-probability approach doesn't necessarily make surveys completely unscientific, Panao said.
"Although it may be difficult to make a generalization about the national population using these polls, they may still be useful in describing a non-demographic subpopulation," which refer to groups defined by behavior or circumstances, like a news outfit's subscribers or Facebook users.
"These limitations should be made clear to whoever will be consuming the information," he added.
Social media polls can be ‘hijacked’
Social media election polls are also prone to “hijacking” by trolls or certain groups who want to project that they are gaining a level of public support, according to Arguelles.
Like how anyone can just participate in these polls, those who intend to manipulate social media surveys can simply create multiple accounts to tip the scales in their favor.
“If you do social media polling, those who are in the disinformation industry, because they have so many accounts, they are well-placed to unfairly influence the results,” Arguelles said.
“The result of that is you may end up with results that may mislead the public about the state of public opinion,” he added.
Rappler deleted its Oct. 7 Facebook poll after it noticed a surge of "wow" reactions, indicating voters for Marcos, from Arab and Pakistani-sounding accounts. There were also foreign accounts choosing "hearts" assigned to Robredo.
How to deal with social media election polls
Participating in social media election polls may be tempting, especially because it allows netizens to show their support to their preferred candidates. But considering its possible dangers, Arguelles and Panao shared some tips on how to deal with online election polls.
Watch out for the red flags
Surveys should be transparent about their methodology. Before jumping in, evaluate first if it’s using statistical techniques to ensure that it attempts to represent all sectors of the population.
“Try to look at how it generates its sample. Usually, the easiest mark of a scientific survey is that they're quite public about its methods. You see the sample size, you see how they generated their samples, what the questions are, etc.,” Arguelles said.
Try to avoid the temptation
Social media surveys can be a fun and convenient way to exercise your right to vote. But remember that when done carelessly, these online polls can fall in the wrong hands.
“Netizens should try to avoid the temptation because at the end of the day, no one will really benefit from the very unscientific and unreliable poll,” Arguelles said.
“So instead of making these social media election polls popular and engaging them, parang isipin nila that they will mostly go against the disinformation industry as well,” he added
You are not required to participate
There’s no harm if you just simply skip online surveys for now. After all, your actual vote will be counted come May 9, 2022.
“People should remember that even if sampling for social media election polls is purposive, participation is voluntary. Hence, their worth as a citizen is not at all diminished when they do not participate,” Panao said.
While you can do your part in avoiding social media election polls, news outlets and other groups should also realize that they have the responsibility to ensure that their surveys are scientific and reliable.
Rappler apologized after taking down its poll, saying it was “not aligned” with their objective. Manila Bulletin technology editor Art Samaniego said that while their survey remains online, the publication would "clean” the data to remove suspicious accounts.
“Readers should be critical of news outfits for it is also the news outfit’s responsibility to report the polling design and its limitations,” Panao said.
“Ethically, the responsibility of informing the public of these caveats falls—not on Facebook, not on the candidates indirectly or directly benefiting from the survey, nor the citizens who participated—but the media outfit which conducted them,” he added.
(Editor’s Note: This update includes the clarification from UP political science professor Alicor Panao that using a non-probability sampling doesn’t necessarily make surveys “unscientific.”)
MORE ELECTION EXPLAINERS: