Back

Note: This article represents the view of the author and not the University of Hong Kong.

According to the latest survey findings of the Public Opinion Programme of the University of Hong Kong, CE Tung Chee-hwa's popularity rating was 45.7 marks, just above the 45-mark crisis level. The author has written before that 50 marks could be regarded as a dangerous level in terms of popularity, while 45 marks could be the threshold of a credibility crisis. The relation between credibility crisis and governance crisis will be further discussed in future. This article focuses on the meaning of the 0-100 rating scale, and its connection with other popularity indicators.

Twelve years ago, after serious investigations and considerations, the author decided to introduce into Hong Kong the 0-100 rating scale, which was occasionally used in democratic societies. In the West, the popularity of political leaders are normally measured in terms of "approval rates" or "support rates", mainly using two broad types of survey questions:

(1)

Do you approve so-and-so's way in handling his/her job (say, as President or Prime Minister)?

(2)

If a general election were to be held tomorrow, would you vote for so-and-so?

Normally, respondents can answer positively, negatively, or "don't know". Positive and negative answers are normally bi-polar, or on 4-point scales. Bi-polar answers may include "approve" or "disapprove", "satisfied" or "dissatisfied", "yes" or "no", and "would" or "would not", while answers on a 4-point scale can include "very satisfied", "quite satisfied", "quite dissatisfied" or "very dissatisfied", or "definitely yes", "probably yes", "probably not" or "definitely not".

In the West, persons being rated using a 0-100 scale are normally those not popularly elected, or those yet to be elected. The interviewer would normally ask the respondent to imagine a Celsius thermometer, with 0 degree as the freezing point, 100 degree as the boiling point, and 50 degree as half and half. The respondent is then asked to rate one's feeling towards some person using that "thermometer". Obviously, such an exercise does not require any imagery of elections, or comparison among different persons.

However, data obtained by this method cannot be directly converted into vote shares, and is therefore of little help in predicting election results. In democratic societies, people's mandate determines the winner, whether during or after elections. No wonder opinion surveys are more often than not called "opinion polls". For this reason, there is a strong tendency in the West to use bi-polar answers to measure the popularity of political leaders, in terms of "approval rates" or "hypothetical vote share", rather than thermometric figures.

Democracy in Hong Kong is still in its infancy. Except for the popularity of the directly elected legislators, using "hypothetical voting" questions to measure the popularity of political figures seems to be of little meaning to the respondents. That's why the author has adopted the scale of 0-100 as the mainstream test for popularity ratings, 12 years ago. Nevertheless, the idea of "accountability" has now become loud and clear, and people's call for democracy has become increasingly fervent, popularity measurement will have to be readjusted accordingly. What's more, some local officials have learnt to twist the meaning of, say, a support rating of 40 something marks into a 40 something percent support rate. They even claimed that such a degree of popularity could rival that of any political leader in the world. These remarks might well have cooled down the public anger, but they are simply betraying science, unfair to public opinion, and harmful to our society.

As an opinion researcher, the author has been testing a variety of indicators which could be used to measure the public mood. On the one hand, we need to investigate the relationship between different indicators. On the other hand, we need to cater for the need of the changing public. Take the opinion surveys on CE's popularity as an example, aside from the 0-100 support rating exercise, the following three questions have also been used:

(1)

On the whole, how satisfied are you with Tung Chee-hwa's policy direction? (5-point scale)

(2)

Do you think Tung Chee-hwa is doing a good or bad job as the Chief Executive? (5-point scale)

(3)

If a general election of the Chief Executive were to be held tomorrow, and you had the right to vote, would you vote for Tung Chee-hwa? (bi-polar scale)

Question (2) is actually a local version of "approval rate", except that we have given up the 2-point or 4-point scale. Instead, a 5-point scale is used, which includes an additional option of "half-half", such that public opinion would not be artificial polarized. On this front, we have tried different scales, question wordings and probing techniques, such as "Do you approve or disapprove Tung chee-hwa's way of handling his job as the Chief Executive?" We finally decided that the above-mentioned questions were most appropriate, although we would not be able to compare the results directly with the "approval rates" used in other countries.

Should a direct comparison be needed, Question (3) can serve as a bridge. Although the public does not have the right to vote for the Chief Executive, and may not know what to consider when making such a "decision", this question can connect us to the international community, and would one day belong to the mainstream. We have, therefore, decided to apply different methods alternately and concurrently, so as to enhance the analytical power of the data.

All in all, our study has shown that average ratings on the 0-100 scale do not fluctuate too much, usually between 45 and 65 marks. However, taking into account the relatively smaller deviations of the values, the normal distribution of the figures, the applicability of statistical calculations, this is still considered to be the best option, if applied appropriately.

As for the local versions of "policy direction satisfaction rates" and "job performance approval rates", the statistics are easy to understand. However, using 5-point instead of 4-point scales like the West will normally compress the figures. Fairer they may be, they may "dwarf" the popularity of the government and political leaders.

If international comparison is what we are after, then we should better use "hypothetical vote shares". Concurrent tests showed that a support rating of 55 marks was more or less equivalent to a "vote share" of 45%, 50 marks could be converted to round about 30%, 45 marks to 20%, and 40 marks to 10% to 15% (please see table and chart). The "HKU POP Site" managed by the author has already released all figures related to CE Tung Chee-hwa's popularity since 2002, for public reference and scrutiny.

Exactly how low is low and when would it trigger a "governance crisis"? The author will discuss this in future. In late 1990, after the "approval rate" of Margaret Thatcher sank to 25%, she withdrew from the election for the leader of the British Conservative Party, thereby gave up her job as the Prime Minister of UK, a post which she held since 1979. In early 1997, John Major lost his post of Prime Minister to Tony Blair, after his "approval rate" hovered around the level of 30% for a long time. As for former USA President Bill Clinton, his lowest ever "approval rate" within his 8-year terms of office was as high as 37%.

These figures are presented with a hope that our Hong Kong society would keep up its pace to develop itself, after it has emerged from crises. At least, the media and our local officials should use opinion figures fairly and objectively, sticking strictly to what is and what is not.