Words by Samantha Delis
With public opinions being ever-changing, the questions are never-ending.
As Samuel L. Popkin quoted in his book entitled The Reasoning Voter, “Public opinion is an illusive commodity. Attempts to measure it will perforce reveal inconsistency and change”.
Public opinion constantly changes from time-to-time. The realm of politics cannot escape from these ever-changing public opinions. Before the votes are shaded, it is necessary to let the voices be heard first. With the upcoming 2022 election nearly approaching, the election polls continue to proliferate. These polls are said to be the “preferences of the masses” that strengthen the democracy of the voters by providing a platform of their opinions. With a total of 18,180 elective positions at stake, conducting election poll surveys is an ideal starting point to get to know each candidate. This, however, is not always the case. A multitude of questions arise every time an election poll is released in the mainstream or online media, especially when the results do not reflect their “truth”.
For some reason, even the candidates themselves disproved the survey results. A tandem even claimed that these poll surveys are used to “condition the minds of the public”. Their campaign experience on the ground is allegedly different from the survey results.
These doubts and questions are all wrapped up due to the public’s and the candidates’ claims of how these polls can be manipulative. While these election poll surveys provide a snapshot of the voters’ preferences, it remains dubious to some due to its alleged mind-conditioning capability, bias, and inaccuracy.
How election poll surveys work
Poll surveys remain a critical aspect of our democracy. While the elections declare the victor, poll surveys answer how a candidate won. The election poll releases have always been the subject of media interest. This interest focuses only on the poll results instead of how these polls/surveys are carried out.
There are several types of polls used in political reporting – all of which aim to collect information based on the opinions of a group of people. During the election period, tracking polls and exit polls are the most frequently conducted. Tracking polls refer to a series of individual surveys conducted at key periods to track the behavioral and attitudinal changes of the respondents. This can be done on a regular basis (e.g., every month, every quarter). Meanwhile, the exit polls are conducted after the voters cast their ballot on election day. This kind of poll provides an immediate gauge of the possible election results. Thus, it is important that these polls should be conducted by reliable firms, especially in this surge of disinformation period.
In the Philippines, the primary mainstream political polls that conduct tracking polls and exit polls during the election period are Pulse Asia Research Inc. and Social Weather Stations. Pulse Asia survey releases election 2022 poll results starting as early as June 2021. Meanwhile, the SWS survey also has its regular release starting October 2021, shortly after the filing of candidacies for the upcoming election.
Pulse Asia survey polls conduct survey fieldwork using face-to-face interviews based on a sample size of 2,400 representative adults (18 years old and above). The poll survey covers subnational estimates for geographic areas such as Metro Manila, the rest of Luzon, Visayas, and Mindanao. It also includes socio-economic classes ranging from class ABC to class D and E. However, their poll surveys for January to March 2022 excluded people’s opinions from class A and B representations, claiming that these top classes only comprise 1% of the voter’s population. This class exclusion raised questions as to whether these surveys are reflective of the voters’ choices.
On the other hand, SWS surveys also use face-to-face interviews of 1,200 registered voters nationwide, with 300 respondents each per geographic area (i.e., Metro Manila, the rest of Luzon, Visayas, Mindanao). Unlike the Pulse Asia survey, SWS surveys do not consider socio-economic classes in conducting poll surveys, or at least not mentioning it in the results. In a webinar discussion by The Legal Network for Truthful Elections (LENTE Philippines) last March 18 with Dr. Mahar Mangahas, Chair Emeritus of SWS, he emphasized that the questions and answers for poll interviews are all recorded for archival purposes. The names of the respondents, however, remain anonymous.
Aside from the mainstream polls, the current 2022 election period is also proliferated by poll surveys conducted by various local universities and colleges. This type of poll survey amplifies the voices of the student sector, which comprises a large portion of the country’s voter population. The latest data from the Commission on Higher Education (CHED) recorded about 1,975 higher education institutions in the country. In line with this, Commission on Elections (COMELEC) Director James Jimenez believes that young Filipino voters play a significant role in the upcoming election since they comprise 52% of the total registered voters in the country.
With this information on how an election poll survey works, can they be considered reliable? Do these poll surveys impact election outcomes?
Election poll surveys vis-à-vis public mass
The approximately 2,000 poll survey respondents did not even reach 1% of the 65.7 million registered voters on the 9th of May. With that ratio, the impact of election polls is technically negligible. However, this less-than-a-percentage is believed to echo the votes of the vast majority. This, admittedly, is crucial. Public opinion answers why a candidate won; thus, it is essential to be reflective of the voter’s preferences.
Most studies present two hypotheses to analyze the effect of poll surveys on public opinions: the bandwagon effect and the underdog effect. The bandwagon effect claims that a popular option tends to become more popular due to the poll outcomes. Meanwhile, the underdog effect suggests the tendency of potential voters to choose the unpopular option in the poll surveys.
In contrast to the studies suggesting how these hypotheses impact public opinion, SWS director Leo Laroza refuted these claims in his presentation at the Asian Conference for Political Communication in September 2017. Upon examining the post-election effect of election surveys on voters, only 5% of the respondents claimed the bandwagon effect influenced them in the 2016 elections. Meanwhile, only 3% of them claimed they were influenced by the underdog effect. However, the same presentation also released SWS poll surveys on tactical voting. On quoting the poll survey’s question: “If the most desired candidate has no hope of winning, then vote for one who might win in order not to waste one’s vote.”, 56% of the respondents agreed – thereby admitting the bandwagon effect would have influenced them. These two poll surveys were done by the same organization and almost within the same time frame. Hence, it depicted contrasting results due to different delivery of questions.
According to statistician and senior research fellow of the Philippine Institute for Development Studies (PIDS) Dr. Jose Ramon Albert, wording matters in the election polls. Split-ballot experiment refers to dividing the sample into two or more sub-samples to distinguish if differences exist between different versions of the survey. In this type of design, there is a tendency that these survey takers may acquire a subconscious bias based on how the questions were asked.
Pulse Asia and SWS poll surveys have similar questions asked differently to the respondents. In the Pulse Asia poll surveys, the question was delivered: “Of the people on the list, whom would you vote for as president/vice president/senators of the Philippines if the May 2022 elections were held today and they were candidates?”. Meanwhile, in the SWS survey polls: “Here is the list of the possible candidates for presidential/vice presidential/senatorial post(s). If the May 2022 elections were held today, who would you most likely vote for?”. In addition, Pulse Asia surveys also gather the second choice voter preference, asking: “If the one you chose for president/vice president of the Philippines does not continue his/her candidacy, whom would you vote for if the elections were held today and the rest on the list continued to be presidential/vice-presidential candidates?”. From these questions, the manner of asking questions was delivered differently. Could it affect the responses? We will never know, that is for sure.
From this, can we say that the public opinions influence polls or the polls influence public opinions?
A throwback to the two-decade snapshots
Poll surveys always have room for surprises. Most people, including the candidates, treat these polls as a gospel, thereby leading to rejection of the election outcome when it did not favor them. We are accustomed to treating these polls as highly accurate as they hold long histories of picking the election winner.
However, the poll survey results are not permanently fixed – they are variable. In the 2016 presidential elections, then-vice president Jejomar Binay and Senator Grace Poe led the poll for both Pulse Asia and SWS surveys in the early months. However, Duterte kept a wide lead in the poll surveys months before the election. This is an evident manifestation of how poll surveys are dynamic.
Similarly, the 2016 vice-presidential surveys showed how the voters’ preferences changed over the course of time. Along with Binay’s lead in the presidential surveys, Senator Chiz Escudero dominated the vice-presidential polls for the early months. The early months’ poll outcomes were far different from the final polls, wherein Escudero and Cayetano were statistically tied for third place. In contrast to the early polls, the final polls showed a close race between Ferdinand Marcos Jr. and Leni Robredo for the top spot.
The case was also the same for the 2010 elections, wherein then-senator Manny Villar topped both Pulse Asia and SWS surveys for August 2009. The tables turned when Liberal Party endorsed then-senator Noynoy Aquino as the standard-bearer for the presidential race, a month after the death of former President Corazon Aquino. The poll surveys after this endorsement showed the domination of Aquino, which lasted until the exit polls along with then-Makati mayor Jejomar Binay for the vice-presidential race. Similarly, 2004 election surveys had the same scenario. Fernando Poe Jr. topped the SWS surveys until March 2004. However, Gloria Macapagal-Arroyo took the lead for April and May SWS surveys which is reflective of the national elections.
Aside from the presidential and vice-presidential positions, the election poll surveys clearly provided good track records in the senatorial races. 2004, 2010, and 2013 senatorial final pre-election polls for both Pulse Asia and SWS show 100% accuracy between their surveys and the election outcomes. Meanwhile, 2007, 2016, and 2019 polls show 92% accuracy.
These previous election poll surveys are manifestations that serve as snapshots, not predictions. As the elections draw near, the voices of the public change on a regular basis. The poll surveys also show the significance of exit polls. Its accuracy is reflected in the election outcomes over the past years. While some claim these poll surveys condition the voters’ minds, it would be a lie if one would deny how these poll surveys reflect public opinions in many ways. Despite the small sample size for each poll survey, the accuracy of the exit polls and final pre-election polls are mirrored in the election outcomes, time and again.
To some extent, does accuracy reflect reality?
A glimpse of the current snapshot
The upcoming May 2022 elections have dominated the headline across all media platforms. Ten candidates will vie for the presidential post, while nine candidates will run for the vice presidency. For the senatorial posts, 64 aspirants will compete for the 12 slots. Meanwhile, there are 177 names in contention for the party-list groups.
The first Pulse Asia major poll was conducted on December 1-6, 2021, after parties changed candidates last November. In this survey, Marcos Jr. jumped from 15% in the September 2021 polls to garnering 53% in the survey for December 2021. Then-presidential front-runner (for September 2021), Sara Duterte, led the vice-presidential December poll survey with 45% of voters’ preferences. September vice presidential polls show that Sen. Vicente Sotto III is leading since [Sara] Duterte has no interest yet in seeking a national post. Meanwhile, the SWS survey for October 2021 was also dominated by Marcos Jr. with 47%, while Sotto topped the vice-presidential race with 44%.
The Marcos-Duterte tandem gained the upper hand for the subsequent releases for both Pulse Asia and SWS poll surveys. However, with just over a month to go until the elections, Vice President Leni Robredo gained the momentum for the latest poll surveys. The 9-point increase within a month (from February to March) in Pulse Asia surveys implies how surveys are constantly fluctuating.
For the senatorial post, the poll survey results for both Pulse Asia and SWS surveys within October to March revolve around these candidates: Tulfo, Legarda, Cayetano, Escudero, Zubiri, Binay, Villar, Villanueva, Gatchalian, Estrada, Padilla, Ejercito, Hontiveros, and Bautista. The poll surveys remain consistent with these 14 names vying for the 12 upper chamber positions. Meanwhile, nine party-list groups garnered at least 2% voter preference in Pulse Asia survey conducted February 18 to 23. Among these groups are: Anti-Crime and Terrorism Community Involvement and Support (ACT-CIS), Ako Bicol, Senior Citizens Partylist, Malasakit@Bayanihan, 4Ps, An Waray, Uswag Ilonggo, Gabriela, and Agricultural Sector Alliance of the Philippines (AGAP).
These poll survey outcomes from mainstream polls are largely different from university surveys. Most of which showed VP Robredo and Sen. Pangilinan dominated the presidential and vice-presidential race. Diokno, Hontiveros, De Lima, Escudero, Colmenares, Baguilat, Lacson, Gordon, Zubiri, Espiritu, Matula, and Trillanes topped the senatorial preferences of students, which is far different from the mainstream poll results. As per the party-list groups, Kabataan Partylist is in the lead. Compared to the results from mainstream polls, university-based polls gave different outcomes. This is a manifestation of how different sectors have different vote preferences. Indeed, poll surveys always have room for surprises.
As we enter the final stretch before the May 9 elections, the nuisance surrounding the elections is deeply-rooted in the long-standing dilemma of Philippine politics. For a long time, we have been deluged in the system of choosing our leaders – yet, little did we know that nothing has ever happened, time and again. Election poll surveys had served its purpose for the longest time in amplifying the voices and preferences of the public mass.
Whether or not these polls have unduly influenced the election outcomes, no one could disregard their power to pick the winner accurately. The truth behind these election poll surveys depends on how one would perceive it – either to accept or to reject the outcome of national elections.
If the polls answer the how’s, then the public opinions answer the why’s of the election outcomes. It is not mutually exclusive.
It has always been a long-overdue question whether public opinion is reflected on the poll surveys or if poll surveys influence these public opinions. To answer this question, would you rather take a poll or not?
UPLB Perspective 