JUST like any other elections in the past, opinion or pre-election polls (or “surveys”) are released to help provide voters a clear picture of how people perceive candidates for national or even local posts, and help aid voters in their decision-making when choosing the country’s next leaders.
But many view surveys as a platform open to misuse or abuse, and can have somewhat of a conditioning effect, though empirical studies in other countries have been divided on this. A research firm head said other studies argue that while survey results may lead to a bandwagon (supporting the one who leads), the same survey may also have an underdog effect (supporters of the trailing candidate will then work hard to increase support for their favored candidate).
Still, many believe surveys to be a more accurate and scientific representation of the true sentiments of the electorate depending on the framing of the right questions and other considerations to help cull accurate data.
Recently, three representatives from some of the popular research firms in the country gathered to provide answers to burning questions on how surveys are done via a virtual roundtable discussion organized by the Philippine Press Institute (PPI) titled “Na-Survey Ka Na Ba? How Opinion Polls are Shaping the May 9 Elections,” and supported by the Hanns Seidel Foundation Philippines.
Are surveys paid?
DR. Ronald Holmes, president of Pulse Asia Research Inc., said there are two types of surveys—commissioned and non-commissioned. He said what they usually release to the public are the non-commissioned ones that Pulse Asia pays for, done quarterly in a year and, as elections draw near, every month from January to April. For the commissioned surveys, which is funded by a separate entity, Pulse Asia decides exclusively on the survey’s structure, the sampling, the questions. The only difference perhaps, Holmes said, is the source of funds. “All research requires funds, but the source of funding does not, in any way, affect the integrity of the survey,” he stressed. “Whatever the results, whether good or bad for the one that commissioned the survey, the results should not be changed.”
He added that what the clients can only do with regard to the survey is state their objective, but the framing of the questions lies on Pulse Asia’s academic team, Holmes added. “If they don’t agree with the framing of the questions, we just return the funds.”
Dr. David Yap Jr., chief data scientist at Publicus Asia Inc., agrees. He emphasized that it is important that survey firms stand their ground on funded surveys. He said that there are many perils when it comes to formulating or constructing questionnaires. “You can have a loaded question, a double-barreled question, all leading questions that we as statisticians would like to avoid. While the client has objectives on what information they need to see from the survey, at the end of the day, we will stand our ground that this is how the questions should be phrased.”
Yap said they are also watchful of other factors, though trivial, when conducting surveys, especially during interviews, and make sure that their field personnel are trained to avoid the pitfalls of conducting them. “They just don’t ask the questions; everything from the appearance of the interviewer, their tone of voice when asking questions, their intonation, how they respond to the interviewees, all these factors are considered and we are very watchful of.”
Respondents’ size
FOR Dr. Guido David, OCTA Research fellow, that’s a matter of statistics. He said that based on statistical design, if survey questions are designed properly, then that is still representative of people’s sentiments. “If the survey is random and has sufficient distributive characteristics, then we can make the inference that it is representative of the sentiments of the voting public. Of course, there is a margin of error but that depends on the sample size. How much is sufficient depends on how little margin of error is desired.”
As for surveys done via social media, Dr. Yap said this depends on the randomness of the sample, and the selection of the sample allows the survey firms to invoke the “central limit theorem,” which he said is critical to the sampling since it allows survey firms to assume certain distributional properties of the sample that can be extended to the greater population. “This allows us to indicate the margin of error. The crux here is the randomness implicit in the sampling design. If the survey collected respondents in a non-random manner, then all these assumptions will not hold. That is where the skew, the biases, will come into play, which we try to minimize through the randomness of the sample.”
Should survey firms be regulated? Do they condition the minds of people?
As regards regulating survey firms, Dr. Antonio Gabriel La Viña, chief of Jurisprudence and Legal Philosophy Department of the Philippine Judicial Academy of the Supreme Court, said that regulation applies to any industry, but in terms of suppressing, not allowing to publish survey results, this has been settled by the Supreme Court. That’s a basic academic freedom and basic right to information of the people, he said. “I’d rather have a situation where survey firms publish the results of non-commissioned surveys, and when they have commissioned ones subject to their rules, it should be published, so it can be critiqued by the people.”
La Viña believes, however, that through the years, people have shown that they are not much affected by surveys on who they decide to vote for, but surveys done nearer to the elections do shape the decision of politicians.
On the question of mind-conditioning, Yap pointed out that reputable survey firms are not media entities or influencers but are engaged in polling, not mind conditioning, and do not compel people to believe their results. The firms in the roundtable, he said, provide the public with all the information necessary to make their own assessments. “Interpretations of other people of content we release are beyond our control. The numbers released are precisely that: numbers. Just like the other reputable survey firms, we take our work and commitment to statistics bravely. More importantly, it is against our interests to falsify our reports, given that our continued survival as a firm is contingent to our reputation.”
Image credits: Sudarsan Thobias | Dreamstime.com