Refer Report
Recently, the results of a public opinion poll showed that President Yoon Seok-yeol’s approval rating reached 40%, causing excitement not only in the political world but also in the opinion poll industry. Among pro-partisan supporters, suspicions were raised about past attempts by the representative of the polling company to run for the ruling party, as well as questions about the content of the survey and response rate. Since then, the controversy seems to have subsided as a similar trend was captured in a public opinion poll using a completely different research method.
The Democratic Party of Korea filed an objection to the National Election Opinion Survey Deliberation Committee regarding the public opinion poll and announced that it would consider ways to legislate the standards for public opinion polls. There is a backlash from the opinion poll industry, calling it “nonsense.”
Issue 1. Is the representative’s past history a problem?
The Korea Public Opinion and Reputation Research Institute (KOPRA) conducted a public opinion survey of 1,000 adults over the age of 18 nationwide on the 3rd and 4th at the request of Asia Today (ARS using wireless RDD, response rate of 4.7%, sampling error at 95% confidence level) ±3.1% points) was announced on the 5th. As a result of the survey, President Yoon’s approval rating was found to be 40%.
As a result of the unexpected result, some people claimed that KOPRA CEO Hyun Kyeong-bo had a history of being ‘right-leaning.’ Some groups criticized him, claiming, “The current representative is someone who has attempted to run as a candidate for the Saenuri Party and the People Power Party in the past.” In a media interview, he said, “Jeju Island is my hometown. It is true that I thought about running for office because there was an invitation to run there,” but refuted, “But I have never had any political affiliation.” Representative Hyeon drew the line that there was no problem with the investigation and emphasized, “I didn’t know that it would come out that much.”
One political official pointed out, “By that logic, most public opinion polls are problematic.” It is pointed out that most opinion poll companies are somehow connected to the political world. For example, Realmeter CEO Lee Taek-soo is a middle school classmate of broadcaster Kim Eo-jun, one of the key progressive figures, and is well known to have entered the public opinion polling industry at Kim’s suggestion. Research View CEO Ahn Il-won is also a former administrative officer in the Roh Moo-hyun administration. Park Si-young, CEO of ‘Park Si-young Co., Ltd.’, who ran a public opinion polling company in the past and was involved in planning public opinion polls, was also a former Blue House administrator during the Roh Moo-hyun administration and is currently a member of the Democratic Party’s Political Innovation Committee.
It is true that even in the research industry, there are unfavorable views on the history of the CEOs of these companies, including the current CEO, but they do not seem to be making a big issue of it. This is because it is viewed as a separate issue.
Issue 2. Is the survey content biased?
As time went on, criticism was raised that the survey questions were biased. However, opinions are divided even among experts. There is a conflict between the opinion that “it is not very appropriate, but surveys are inherently subjective” and the opinion that “there is a bias.”
It is also strange to measure the approval rating of a president who has been suspended from office, such as ‘What do you think about the Corruption Investigation Office’s forcible arrest of the incumbent president despite the illegal controversy over President Yoon’s arrest warrant?’, ‘The central government cited by President Yoon as the reason for declaring martial law’ There is criticism that expressions in questions such as ‘Do you think public verification of the National Election Commission’s computer system is necessary to resolve suspicions about the possibility of hacking and election fraud in the National Election Commission’s computer system lack objectivity?’
Pro-Yeoman netizens summoned a survey regarding former People Power Party leader Han Dong-hoon (then Minister of Justice) conducted in December 2022 by Flower, headed by Kim Eo-jun. At that time, in questions 1 to 3, Flowers asked about political party support and evaluation of state administration, and in questions 4 to 9, all questions were about former representative Han. It includes questions about personal favorability, answering attitude, political qualifications, and whether it was appropriate for the Minister of Justice to file a criminal complaint and a 1 billion won damages lawsuit for defamation by false information in relation to the suspicions of a drinking party in Cheongdam-dong raised by former Democratic Party lawmaker Kim Eui-gyeom and an official of The Tamsa. Done. At the time, it was pointed out that “Isn’t this a Han Dong-hoon stalking style questionnaire?”
The results of a survey commissioned by MBC from the Korea Research Center in 2018 showed that the level of trust in North Korean Chairman Kim Jong-un was 77.5%. At the time, Hong Joon-pyo, leader of the Liberty Korea Party, raised the question, “The world has become pathetic. Should I blame the people of my country or the fake opinion polls based on which broadcaster’s opinion polls?”
However, these public opinion polls are difficult to consider as legal issues. This is because under relevant laws, the subject of public opinion poll deliberation is limited to ‘election-related questions.’
Moreover, the survey questions are not designed by the opinion poll company, but are provided by the party that requested the survey. Park Sung-min, CEO of political consulting firm Min, a representative figure in the world of criticism, also raised issues with the survey questions, saying on CBS Radio Kim Hyun-jung’s news show on the 8th, “The person requesting the survey must have created and sent the survey,” and “I cannot agree with the survey method.” “It’s the same for both sides,” he said, apologizing.
Issue 3. Is the research method the problem?
It is true that there are many skeptical views about the automatic response system (ARS) method based on the random dialing (RDD) sampling frame used by the public opinion polling agency. RDD has high cost-effectiveness because it can lower survey costs, but it has the disadvantage that the sample may be biased and the response rate is low. Since there is a high possibility that highly politically involved groups will participate enough to listen to and respond to ARS until the end, it is considered to be a relatively suitable use for election prediction during the election phase.
Kim Eo-jun’s Flower also presented the RDD method and the Telephone Investigator (CATI) method. It is said that ARS-based surveys are not being conducted these days. However, in his broadcasts, he continues to quote the results of Realmeter surveys that only use ARS. Realmeter surveys consistently show that the Democratic Party is ahead of the People Power Party by about 10 percentage points. The opinion poll process commissioned by Daily An/Pen & Mike, media research commissioned by Newspim, and Chowon C&I commissioned by Straight News all use ARS. And although there are differences in degree in all of these surveys, an upward trend in party approval ratings is detected in all of them.
One survey industry official criticized, “If the survey method is the real problem, the National Election Commission will have to close down most opinion poll companies,” adding, “All opinion polls have a degree of bias, and ARS is just a methodology among many ways to view public opinion.” did it
Lee Geon, a professor of public administration at Hanyang University, said in a 2023 Korea Institute of Public Administration publication, “It can be said that the ARS method was relatively more accurate than the CATI method in predicting candidate vote rates in public opinion polls conducted ahead of the last presidential election. “This result is somewhat different from the general belief that it is biased,” he said. “We cannot say that ARS is always a superior investigation method to CATI, but at the same time, we cannot always disparage it as an inferior investigation method.” mentioned.
Issue 4. Is response rate a problem?
Some pointed out that the response rate to the KOPRA survey was only 4.7%, indicating that it was not representative. However, many opinion polls use ARS, and it is difficult to find ARS polls exceeding 5%.
Rather, the opinion poll industry responded that “it is a very high level for an ARS method using wireless RDD.” However, it is pointed out that it is not a positive surprise and that there are some aspects that need to be clarified. Kim Chun-seok, head of Korea Research, said, “Realmeter barely meets the response rate by adding landline calls. The wording of the questions is also quite long, and it is difficult to properly digest questions of this size with ARS. That raises questions.” The RDD-based ARS method uses both wired and wireless phone surveys to increase the response rate, but the survey was conducted 100% wirelessly only.
Director Kim said, “The fact that there were almost no non-responses is a clear indicator that it does not represent the entire electorate,” and analyzed, “Although this is the nature of ARS, it is highly likely that the so-called highly politically involved class participated more in the survey.” In the poll, the positive evaluation rate for President Yoon was 40%, the negative evaluation rate was 60%, and the negative evaluation rate was 1%. As a result of decimal rounding, the sum exceeds 100%. Typically, in public opinion polls, 3-5% don’t know or don’t respond.
Strengthening prior regulations?
The Democratic Party initially considered filing a complaint against the polling company, but lowered the level after filing an objection. It is known that a separate legal review is being conducted. However, he recently announced, “We will actively consider ways to legislate the specific method of public opinion polls, the method of calculating response rates, and the standards for publishing survey results.”
There are also voices of opposition from the opinion poll industry. The National Election Opinion Survey Deliberation Committee’s preliminary review has already been announced, but additional regulations are said to discourage the research industry that monitors public opinion.
An official from a research company, who requested anonymity, protested, saying, “It is ridiculous to say that it is regulated in advance,” and added, “There is already a sufficient system in place to make an announcement after being reviewed by the National Election Commission, and to file an objection afterward if there appears to be a problem.” .
Another industry official said, “Public opinion polls can be subject to subjective judgment,” and added, “The problem of lack of representativeness in surveys related to current issues is serious, but in the social science field, there is a need to strengthen regulations on areas that need to be debated and held accountable after the fact.” “This is difficult to accept,” he said sharply.
Telephone investigators are also ‘quiet’ when a similar trend emerges.
Since then, even in the survey results with relatively high response rates using the CATI method, similar phenomena to the KOPRA survey results have been discovered, and the voices criticizing the survey seem to have quieted down.
Gallup Korea, which conducts a survey using only the CATI method, surveyed 1,004 people over the age of 18 across the country on the 7th and 9th, and found that 64% were in favor of impeachment and 32% were against it. Compared to just before the impeachment bill was passed, there was an 11% point change from approval to opposition. In fact, the public opinion in the survey, which is the same as President Yoon’s approval rating, is showing a sharp reversal.
The People Power Party’s approval rating rose 10 percentage points compared to the third week of December to 34%, and the Democratic Party’s approval rating plummeted 12 percentage points over the same period to 36%, making it a close race. The approval ratings of both parties were ‘reset’ to before martial law, and the People Power Party’s approval rating is higher than before martial law, although it is within the margin of error.
According to the National Base Survey (NBS) conducted by Embrain Public, K-Stat Research, Korea Research, and Korea Research, which is also a CATI method, on 1,000 men and women over the age of 18 from the 6th to the 8th, the People Power Party’s approval rating was at the previous level. The survey was calculated to be 32%, a 6 percentage point increase compared to the third week of December. It is the highest since the second week of August (32%), about four months ago. During the same period, the Democratic Party’s approval rating fell 3 percentage points to 36%. 62% said, ‘The Constitutional Court should use impeachment to remove President Yoon from office,’ while 33% said, ‘The impeachment should be dismissed and he should be reinstated in office.’
Experts agree that since public opinion polls are a partial survey rather than a complete survey, inclusiveness is needed. Shin Yul, a professor of political science and diplomacy at Myongji University, said, “It is true that some public opinion polls stand out,” but analyzed, “As polarization has become more severe, people are not accepting any public opinion polls that show results that are completely different from what they think.”
For more information regarding the public opinion poll mentioned in the article, please refer to the National Election Opinion Survey Deliberation Committee website.
Shin Hyun-bo, Hankyung.com reporter greaterfool@hankyung.com
Source: Korean