• Fri
  • Dec 19, 2014
  • Updated: 3:26am
NewsHong Kong

Opinion poll techniques 'could lead to distorted results', says think tank boss

Shiu Sin-por joins row over surveys, saying some techniques used by public opinion programme could lead to distorted results

PUBLISHED : Monday, 10 March, 2014, 11:05am
UPDATED : Tuesday, 11 March, 2014, 9:09am

The head of the government's in-house think tank has waded into a row over opinion polling by describing some polls conducted by the University of Hong Kong as "problematic".

Shiu Sin-por, head of the Central Policy Unit, cited certain techniques used by HKU's public opinion programme, headed by Dr Robert Chung Ting-yiu, which he said could lead to a distorted view of public opinion.

"Chung loves to ask respondents to assign marks [to government officials] … but by taking averages, the findings would easily be affected by the extremes," Shiu said in Beijing, where he is a delegate to the Chinese People's Political Consultative Conference. "I think it is problematic but I am not saying it is wrong.

"Foreign polls do not ask people to give a score on [US President Barack] Obama."

HKU's polling has been under the spotlight since CPPCC Standing Committee member and businessman Peter Lee Ka-kit accused HKU of "publishing poll results that are unfavourable to the central and local governments at critical moments".

He suggested "patriots" set up their own polling organisation.

While Shiu refused to say if polls had become a political tool, he said: "If one masters certain statistical skills, one can manipulate public opinion."

Wording could also affect the findings of a poll. "There is a huge difference between asking 'do you support' and 'do you approve'," Shiu said.

HKU's polls on government performance ask respondents to rate the extent to which they support Chief Executive Leung Chun-ying on a scale of 0 to 100 and whether they approve or disapprove of his performance.

One of the Central Policy Unit's key functions is monitoring public opinion, and Shiu said its latest venture would be to go online to gather views.

The project includes monitoring major internet discussion boards. Websites such as the infamous Golden Forum - which has more than 600,000 registered users - have become increasingly popular spots for Hongkongers to discuss issues of the day.

"We want to observe the trend of internet opinions and study the nature of internet activities. People can speak on unanimous grounds and some of them are paid. What does it reflect?" Shiu said. "We are six months into the study but have not yet reached a concluding view."

The study would not include social media. After it was completed in the next few months, the unit would look at how the government used the internet.

Meanwhile, former Institute of Education chief Professor Paul Morris hit out at Lee's proposal for the business sector to form its own polling organisation as an alternative to HKU's research.

"To set one up to explicitly provide results which favour either the pro- or anti-establishment camp is very suspect," he said.

Morris, who left HKIEd in 2007 amid accusations a government official tried to get him to fire academics critical of the administration, warned that universities risked damaging their reputation if they engaged in biased or politically driven polls.

"That members of [the CPPCC] make such proposals reflects a worrying level of authoritarianism," said Morris, now working in London, of Lee's idea of a new polling body.

Meanwhile, Dr Tom Smith, director of the General Social Survey at the University of Chicago, said Chung's work was well respected by the international polling community. "His surveys have a strong track record of reliability and validity," Smith said.



For unlimited access to:

SCMP.com SCMP Tablet Edition SCMP Mobile Edition 10-year news archive



This article is now closed to comments

Dear Dr Chung,
Transparency is what HKU needs for its polling? How the questions asked then the conclusions drawn, and also how the polled individuals are selected? I am from middles class retired only very recently, me, my family members, my relatives and my friends, none of them I know have ever been invited to a poll conducted by Dr Chung. Besides, we often find that the polling results of Dr Chung gone out of line of what we see and feel. We are part of HK, and we also want our views to be part of HK's view!
Yours Sincerely!
Poll, I sincerely think can be a good tool if it is used properly with highest and fairest standard. Anything short of that I believe is a tool for falsehood. It will be ridiculous to bracket a poll I suspect for Dr. Chung’s or the proposed ‘patriotic' poll is will serve the public well.
I will call on Dr. Chung to make his poll question known to all public prior conducting his polling. Time, date and location of his poll should also be standard information revealed to public after the poll. The public has the right to make judgment on the conclusion of the poll presented.
The greatest fallacy of Dr, Chung's frequent polling is that Hong Kong people has least interest in political matters and therefore mostly have very short span of political memory towards the polled individual. Any unfavorable issue proposed inevitably yield unfavorable popoularity result at the moment of poll. The issue overrides the larger ability of the individual. It is highly questionable to do polling that frequently.
Dr. Chung may get a high mark from abroad which I suspect the high mark is given without awareness of the polls working in the local context in Hong Kong:
1. How frequently the polls in Hong Kong are being conducted that the objective of the polls are a suspect.
2. Good poll costs money and time which the frequent polls can be a hardship to uphold international standard.
3. The poll results always concluded as approval rating by the public of an individual in an office without public awareness in what being polled.
Such deficiency, perhaps mislabeling of result and frequency have been gone on for over a decade without any improvement noticed by the public.
Dr. Chung needs to defend his work here in SCMP or other publications for his future polling.
Dai Muff
By the way, "patriots" already have their own polling organisation. The DAB is releasing the results of its own polls all the time.
Dai Muff
"Shiu Sin-por, head of the Central Policy Unit, cited certain techniques used by HKU's public opinion programme, headed by Dr Robert Chung Ting-yiu, which he said could lead to a distorted view of public opinion."
Well, he would, wouldn't he?
It is always good to have a separate view to balance out an otherwise a one-sided observation. Never good to have only one analysis which can be manupulated for political purposes.
For CCP mouthpieces like Peter Lee, it seems no idea is too stupid to say out loud. The polls are bad because they're critical of the central and HK governments? So the respondents who rated those governments poorly are bad too then? Yes, the quality of a poll has everything to do with the results, and nothing to do with the methodology. Here's the best part: a guy stupid enough to say (or even think) something like that apparently has the "merit" to be on a committee that supposedly guides China. That is indeed a great system.
Now, Chung's methods differ from the norm. Usually, questions don't ask for a random numerical rank, but instead offer quartile or quintile-form choices. Instead of rating CY's performance out of 100, people might instead be offered choices like terrible/poor/mediocre/good/excellent. Chung's way isn't wrong, but maybe isn't typical. The point about wording is legitimate, and that's probably where good pollsters earn their keep.
But I think Peter Lee is over-thinking it. If he doesn't like the results of a certain poll, he doesn't need "patriots" to come up with their own. He just needs "patriots" to make stuff up. Don't actually do anything, and just publish a "poll" that says CY and the CCP have 100% approval among HKers. After all, he's not concerned with methods, or reality; he just wants results that he likes. Best way to guarantee that is to make it up. If NKorea can claim 100% voter turnout, why can't CY claim 100% approval?
To shu....
Better still is to have a real poll and not contest of many ‘polls’ which are not only waste of money but waste of our time to learn the results. That is, much ado of nothing – which one you trust?
So you have missed the whole real purpose of poll. It is not for balancing a view but to know about a prevailing view.
Have you ever heard about biased polls? It is the questions and opinion asking about subject which defines the direction of the poll and survays. I stopped taking part in HKU questionares when I was asked as I experienced several times that they wanted to lure me into one direction which does not really reflects my point of view. I couldn't helped as the questions were created in this way to have this effect.
I already said "The point about wording is legitimate". That concern would be true of any poll. Do you ever read before you respond? BTW, your personal anecdote is summarily useless. Quote us the questions you were asked, and the selection of responses you were offered, then we can decide for ourselves if it was the poll which was "biased", or if it was you.
I'm big on methodology. It would be useful if Dr. Chung released his sampling methods, as well as the list of survey questions, in addition to the results of his polls. And again, that would be useful information to have about any poll.
Anyway, I imagine the only polls you would take to heart will be the fake ones that Peter Lee and his fellow mouthpieces make up.



SCMP.com Account