Why do polls show different results for presidential race?

A signboard installed inside the National Election Commission shows that there are 30 days left to the presidential election. Yonhap

Experts say polls can be wrong, as pollsters find it tougher to conduct surveys after mobile phones replaced landline telephones

By Kang Hyun-kyung

One poll finds that ruling Democratic Party of Korea candidate Lee Jae-myung is ahead of main opposition People Power Party candidate Yoon Suk-yeol within a margin of error. Another says the opposite: Yoon is comfortably leading the presidential election race, with the gap between the two outside of the margin of error.

Nearly 20 public opinion surveys of the presidential candidates are made public every week. These poll results share one thing in common: there's certainly a two-way race going on between Lee and Yoon, with Ahn Cheol-soo of the minor People's Party behind in third place.

Other than that, those polls share almost no other details, including who is leading and what qualitatively characterizes the competition between the two leading candidates.

Amid these confusing results, a closer look at these polls reveals a consistent pattern. Polls taken based on the so-called Auto Response System (ARS) ― a prerecorded autodialed survey ― as a polling method, tend to show results that are in favor of Yoon. Under this scheme, respondents are interviewed by a prerecorded voice in place of a live interviewer, and asked to press certain numbers to give their answers to the questions.

The latest UPI News poll conducted on Feb. 1 to 3 by the polling agency Research View, indicated that out of 1,000 eligible voters nationwide, Yoon earned 46 percent of the total support, followed by Lee who stood at 38 percent.

Considering the fact that the poll has a margin of error of plus or minus 3.1 percent, its results confirm that Yoon is the undisputable frontrunner. This poll was conducted using the pre-recorded autodialed survey method.

Another recent poll, conducted on Feb. 2 and 3 by the Korea Society Opinion Institute (KSOI), also found that Yoon was ahead of Lee within the margin of error. The KSOI poll was also based on the same polling method. Details of recent polls are available on the website of the National Election Commission.

Yoon Suk-yeol, center, the presidential candidate of the main opposition People Power Party, heads to the May 18th National Cemetery in the southwestern city of Gwangju on Sunday. Yonhap

As for the discrepancy between public opinion surveys employing the prerecorded voice method versus live interviewers, one pollster said that the relatively higher ratings for Yoon in the ARS-based surveys seem to be related to the characteristics of the polling method itself.

"The ARS-based polling method tends to overrepresent opinions from voters who are highly engaged in politics," he told The Korea Times on condition of anonymity, as he was not authorized to speak to the media.

Ahn Cheol-soo rejects unified candidacy, vows to finish race

2022-02-06 16:27 | Politics

"Imagine what would happen if you were to get a phone call from a polling agency conducting a survey based on the prerecorded autodialed survey method. You listen to questions and answer them by pressing numbers on your smartphone. To do this, you need to hold your smartphone up to your ear to listen to their questions, and then move it to place it on your palm, answering them by touching the numbers on the screen. You repeat this action several times until the interview is finished. Most people find this method very inconvenient, so what happens is that many people just hang up or just don't answer their questions. But if you are a person who is keenly following political news, you may think the entire process is not that annoying."

The pollster said that about a quarter of all eligible voters are highly interested in politics. However, due to the inconvenience of completing ARS-based surveys, these surveys tend to overrepresent the views of those individuals who are highly interested in politics, and so are less reliable overall than surveys conducted as live interviews by professional pollsters.

Others, however, disagree with this interpretation of the polls' results. Not only Yoon but also Lee has stalwart supporters who intently follow political news, and they rarely withdraw their support for the ruling party candidate, even in the event of new allegations against him. Hence, the claim that Yoon's supporters are more active in responding to ARS-based surveys compared with Lee's supporters doesn't quite hold water.

Some also claim that some members of the public who support opposition party candidates might feel uncomfortable about publicly naming their preferred candidate in live interviews by professional pollsters.

Ruling Democratic Party of Korea presidential candidate Lee Jae-myung, center, delivers an impromptu speech to a crowd gathered in Busan's coastal Haeundae District on Saturday. Yonhap

On top of the various polling methods used, an increase in the number of polling agencies tracking this presidential election seems to be another reason making it difficult to find a pattern in the election race.

As of Feb. 6, 89 agencies were registered as pollsters on the NEC website. This figure is more than double that of polling agencies registered during the 2017 election. Back then, there were some 40 polling agencies only. Experts say that poll results differ among the agencies because, among other reasons, they use different polling methods and samples.

"Each polling agency also has different capabilities, too," said Cho Jin-man, a professor of political science at Duksung Women's University in Seoul. "Given this, it's understandable that polling agencies and their surveys tell the results differently."

Cho noted that polls are simply polls, nothing more.

He went on to say that in the digital era, pollsters find it tougher to conduct surveys in general.

"In the past, when mobile phones were not widely used, public opinion surveys enjoyed a strong reputation as reliable indicators that could help observers forecast the winners of elections. But those days have gone with the disappearance of landline telephones," he said.

"Now, many people live without landline telephones, as smartphones have replaced them. Caller ID services and smartphone apps help users screen calls from unknown numbers, and so pollsters find it tougher to conduct polls."

Cho noted that polls serve a reference that can show public opinion at the moment, while discouraging people from viewing poll results as an indicator of the election's inevitable outcome.