What are the origins of opinion polling? How trustworthy are such polls? What sort of attitude should we normally take towards poll results published in the media? What affect will Gallup's arrival in Taiwan have on the development of domestic opinion polling? In this issue of Sinorama we interview Professor Hsu Chia-chih of the Department of Journalism at National Chengchi University.
Q: The arrival of the American Gallup Poll organization in Taiwan has prompted discussion of whether popular opinion polls are credible or not, and we see a host of opinion polls in the papers every day; what is the origin of opinion polling, and what function does it serve?
A: Opinion polling was developed by western democracies to assist the representative system of government and the media. We generally refer to democracy as a representative system of government, and we tend to think of newspapers as representing public opinion, so we assume the stance of public representatives or newspapers represents the will of the people. Ideally this supposition is correct, but even in advanced western countries they have found that there seem to be considerable discrepancies between the views of lawmakers or the media and those of the average man in the street. It was thought that plutocratic influences or the pressure of interest groups might prevent real "grass-roots" opinions from being expressed; so opinion polls were developed.
In western countries opinion polls do in fact serve as a very important means of testing public opinion, very much like an election. In fact the "yes/no" style of questionnaire is similar to voting. The question is how should this procedure be carried out.
A good opinion poll provides a valuable guide for everyone in politics. Although politicians represent the popular will per se, they need to know what the electorate really thinks. Popular opinion may change, and such trends can only be seen clearly through constant polling.
Q: Many people say that despite the popularity of opinion polls in Taiwan in recent years many subtle problems still remain under the surface; what do you think?
A: One problem with opinion polls is that they are often used for propaganda purposes or as arguments for your own particular policy or proposal. Politicians often use opinion polls to bolster their political platform. When Taiwan's largest opinion poll organization was first set up it was supported by politicians; we didn't know their political aims, but their motives were a cause for concern. There are only a few opinion poll agencies in Taiwan, and currently the most respected polls are conducted by individual polling agencies. These are widely taken up in the press, which is naturally a big problem. We can't just have one or two polling agencies, there must be competition; just as we need competition between newspapers to achieve a more progressive press.
Polling organizations must be financially independent, preferably run by academics, and maintain a politically neutral stance. Their polling methods must inspire widespread confidence, in other words be objective, fair and as scientific as possible. So those involved must be experts with training in social science, statistics, surveying, sampling and analysis.
Q: The local press often carries out its own opinion polls--what do you think of these?
A: Press opinion polls are called "accuracy reports" in journalism terminology, and of course these are better than the usual fluctuating reports, as long as they are done fairly. However as such reports usually concentrate on topics in the news or issues of general interest, they can rely on interviewing passers-by in the street.
For example, ranking top government officials in order by asking members of the Legislative Yuan to give them marks may result in a topical news item, but do legislators really have a better appreciation of ministerial qualities than the general public? Government officials may score lower marks in such a poll if their speeches don't go down well in the legislative assembly or if they tend to work hard but keep their mouth shut. In that case, what do such marks and placings really represent?
In addition, newspaper opinion polls often don't follow social science principles. Things like sampling procedure, model size, content of the questions and margin of error are vital indicators when deciding whether a poll is trustworthy. But polls can be distorted by neglecting certain aspects, either by cutting them out altogether or by burying them away in a brief comment near the end of the article. This prevents the result's true significance from emerging, and readers may easily be misled.
Q: What factor will most determine the future trend of opinion polls in Taiwan?
A: The worst scenario would be for the public to harbour doubts about opinion polls, and think they were subject to government manipulation. If the image of opinion polls is spoiled it would be disastrous, because if it was spoiled right at the beginning they would never get public support.
Q: What are the limitations of opinion polls? Are there any limiting factors which are particularly difficult to avoid in our culture?
A: I think they have quite a number of limitations. For instance, public opinion is a fickle thing and different results keep emerging; to counteract this, opinion polls must be taken regularly so as to provide a general trend, which is expensive in terms of effort and financial cost. The polls we are seeing today aren't done like this, they're just a joke. A one-off poll conducted all of a sudden has little bearing on reality.
Another consideration is whether members of the public are willing to express their opinion or not. This is a big problem in our culture, especially with certain political issues where people may not actually think the way they say they do. Also, we Chinese tend to want to polite and anticipate the answer the questioner wants. This only adds to the limitations.
Of course the questioning procedure is itself an artificial product, and any of a number of factors including questionnaire design, questioning method, sampling and analysis may play a part in influencing the outcome of the poll. It's near impossible to come up with a faultless methodology. Psephology is still in its infancy in Taiwan. Large numbers of social scientists have been drafted into designing questionnaires and conducting poll analyses, but unless they research deeply into improving their methods of questioning, the truth about Taiwan's public opinion will never be known. This is a serious failing. In particular, most of our polling methods are western in origin, such as the parameters of a recent Gallup poll on government ministers which asked the public to rank them according to factors such as courage, approachability and far-sightedness. These are fundamentally western standards, and it would be worthwhile asking social scientists to consider more whether Chinese people don't have their own particular standards of assessing public figures. Of course this is not simply a question for opinion pollsters, but a wider question affecting the social science community or even the academic community as a whole.
Q: What influence has the Gallup Poll organization setting up an office in Taipei had on the opinion poll market in Taiwan? What do you think of the idea that this is the first step in internationalizing the market?
A: It's not a question of Gallup internationalizing Taiwan! Rather, Taiwan is now a fully-fledged member of the international community, which cannot afford to ignore our political customs, values and opinions. This is why Gallup has come here.
[Picture Caption]
Opinion polls abound on all sides in Taiwan, but how should we regard their conclusions? Professor Hsu Chia-shih explains the long and the short of it.
Good quality researchers are a key element in ensuring the accuracy of opinion polls.
Telephone surveys are the easiest and most widely used method of conducting opinion polls. Shown here is the Opinion Poll Foundation.
Collecting, processing and analysing questionnaires requires considerable time and manpower.
These piles of data are going to be processed into an opinion poll result for use by the public.
Good quality researchers are a key element in ensuring the accuracy of opinion polls.
Telephone surveys are the easiest and most widely used method of conducting opinion polls. Shown here is the Opinion Poll Foundation.
Collecting, processing and analysing questionnaires requires considerable time and manpower.
These piles of data are going to be processed into an opinion poll result for use by the public.