[Outlook]Surveys: good and bad

During this year’s presidential election campaign, surveys were conducted a lot more often than in the past and their influence has grown. Media outlets made candidates’ approval ratings headline news, and polls played a decisive role in political primaries. But critics have raised questions about the accuracy of surveys.
In fact, election polls face a more rigid evaluation than other polls. Unlike surveys for marketing, those for an election can be compared to election results so their accuracy is revealed immediately.
There are other problems with surveys. Response rates are lower than 20 percent. Some people do not have home telephones or their numbers are not listed in directories, and others are only home for fairly short periods of time during the day.
In some cases, surveys were conducted as soon as a certain political issue arose so public opinion was hardly formed. This resulted in inaccurate results.
Nevertheless, polls have proved to be quite accurate.
Skeptics even ask if polls are really needed. They say that rather than reflecting public opinion, surveys actually define it, if too many of them are conducted.
Surveys emphasize certain issues, preventing new ones from coming to the forefront.
Last weekend, a video clip showed Lee Myung-bak saying he owned BBK in a lecture at Kwangwoon University. After it was publicized, I wanted to know how that incident affected his approval ratings.
But the election law prohibits publicizing polls in the six days immediately before an election, so there was no way to measure changes in public opinion after the video.
Politicians estimated approval ratings to suit their purposes.
For instance, rumors swirled that Lee Myung-bak supporters became a more close-knit community after the video. Others said approval ratings for Chung Dong-young and Lee Hoi-chang went up.
Too much information blocks public opinion from changing freely, but last week’s case demonstrated that too little information can distort public opinion.
In a broad sense, election results capture public opinion on the day the election is held. If elections are valid, there is also no reason to be skeptical about surveys.
The argument that surveys must not be allowed to affect voters is only about norms. To allow people to make their own decisions, factors that could influence them to vote in a certain way should be removed.
It doesn’t mean that factors that only reflect a situation as it stands at a certain point in time, such as a survey that measures public opinion, should be ignored.
When a person loses the ability to hear, his pronunciation will likely become less articulate because he has difficulty listening to his own voice and comparing his pronunciation to other people’s.
Likewise, knowing what others think helps people make up their own minds about issues.
It is not wrong that individuals form opinions under the influence of surveys. People need to compare their own thoughts with others’ and reflect on others’ opinions, if needed.
Some suggest that we should hold debates to reach a consensus in order to overcome the problems of a representative democracy.
That means we should have a process whereby we listen to other people’s opinions and work out differences in opinion in a debate format.
In a broad sense, surveys show different opinions in the same way.
The problems during the presidential election campaign were the way survey results were delivered and used.
The media was interested only in the race in approval ratings and did not examine why a certain candidate was winning more support than another. It is important that the media not only deliver statistics but also provide information on the reasons and meanings of the numbers because the people need to know other people’s thoughts about candidates.
It was wrong for political parties to directly reflect survey results when they nominated their presidential candidates.
Survey results should be used as references to make decisions, not as deciding factors.
In surveys, researchers ask only a part of the entire group about an issue and estimate the opinion of the entire group based on the sample.
During the presidential campaign, the problem was not that there were too many surveys, but that survey results were delivered only as numbers or misused.

*The writer is a professor of political science at Sogang University. Translation by the JoongAng Daily staff.

by Lee Hyun-woo