Why I Pursue The Issue of 12/4 Numbers - Part 1

This is no doubt the first part of many more to come.

The first reason is professional in nature.  

In the extended post The Numbers for 12/4 Hong Kong, I listed a number of newspaper reports on the numbers.  There are a number of different sources at this time (the Civil Human Rights Front, the Hong Kong police, and several university research teams).  The purpose of the listing is to show which newspaper reported which sources (and also not reported the other sources).  That may be interesting about media accuracy, fairness and balance, but surely the media depended on the information provided by those sources.  That should be the question.

Which are the sources and what have they got to say?  As of December 5, here are the major sources and their respective estimates:

For disclosure, I have only corresponded with Thomas Lee by email in 2004 concerning the aerial photo analysis methodology that year.  I don't know anyone else at any of the other organizations and I have never met with or spoken to any of the named individuals in those organizations.  In The Standard, I was included in Justin Mitchell's article A Question of Survey Abuse along with Robert Chung and Michael DeGolyer.  I was mentioned in that article alongside those two eminent Hong Kong survey researchers because of my professional standing as a survey researcher for a major media research company in the United States and I also happen to live in Hong Kong presently.  Justin Mitchell interviewed us separately.

Thus, this blog post was done completely on my own initiative and Robert Chung, Yip Siu-fei, Michael DeGolyer and Thomas Lee have no pre-knowledge about this beforehand.  So, forget about conspiracy theories from the cabal of survey researchers!

Look at the 12/4 numbers.  

There are clearly two completely different sets: Civil Human Rights Front at 250,000 is very much out of line with the others: 63,000; 89,500; 72,000; 72,500 (100,000?); and 92,000.

What are the consequences?  As in years past, heaps of abuse have been launched at both sides, less for technical reason than partisan loyalties.  Here, it is worthwhile to read about what happened to Yip Siu-fei in 2004 (see previous post):

[in translation]  Speaking about the recent criticism, Yip Siu-fei looked as if he did not know where to begin.  "Actually, you could classify me as a pan-democrat ... but maybe I would not say that because this is going to draw criticisms such as 'How can you possibly be a pan-democrat?'  Anyway, I support the 7/1 march and I was a member of the march."

At a time when emotions are surging, Yip Siu-fei published a smaller estimate of crowd site.  He foresaw that he would be regarded as being on the wrong side: "My thirteen-year-old son said, 'Daddy, aren't you afraid of being slammed()?'  But I feel that I am not trying to be contrarian and I am only telling the facts according to the data."  He believes that those who support democracy also believe in justice and honor the facts.

Moreover, he thinks that an "exaggerated number" is not necessarily good for those who marched for democracy.  "It creates a mistaken hope to let people feel that it would be a big letdown if there were fewer than 500,000 the next time.  If we accept 530,000 this time, then are we going to have to get 1 million next year?" said Yip Siu-Fei.

On the day of the march, he and his students set up two counting stations at Victoria Park and Admiralty, working from after 2pm to after 7pm.  "The numbers were 125,000 at Causeway Bay and more than 110,000 at Admiralty.  Even if all the 125,000 at Causeway Bay left before the finish and all the 110,000 at Admiralty joined in the middle, the sum if still less than 530,000.  His method is to obtain counts of the flow every 15 minutes, and then interviewed a sample of marchers to estimate the number that joined in the middle.  The estimated number of marchers was 165,000 with a margin of error of 25,000.

At last year's 7/1 march, his estimate on the number of marchers was 250,000, which was criticized by academics.  "Last year, I did not account for the number of marchers who joined in the middle.  Therefore I included this category in my new calculations and the number is now more complete."  Yip Siu-fei's voice showed the happiness of a child and he was obviously delighted with having found a more accurate method.  He said that he will publish a paper on this method.

"I also feel that I have to say something about the police.  Why do people feel that the police underestimated the number of marchers?"  Yip Siu-fei said that the police estimate was closer to scientific data.  "Someone said to add the police and civil numbers together and divide by two to get the right answer.  This is unscientific."

"7/1 is an important event and many people pay attention, in Hong Kong and internationally.  It has a huge impact.  Actually, any march over 100,000 people is already a significant event.  Just think, even if the anti-Article 23 march had 6,000 or 10,000 people, it would not have been ignorable.  Once the number gets past a certain point, it cannot be ignored."

He remembered that at the time of the June 4 incident in 1989, he was teaching at Melbourne University.  He heard about the suppression of the students at Tienanmen Square and that the local Chinese community was going to hold a demonstration.  On that day, he "cut class" and left his students to march.  "The local students were very understanding, because they felt that this was a very important matter."

A supporter of marching to express demands, he frankly admits that there were pressures about publishing the estimate of the marchers this time.  "Some students participated in this research.  Prior to publication, we discussed this.  Our final decision was to proceed to publish, because we believe that this number is closer to the facts and the citizens should know that democracy means respecting the facts" said Yip Siu-Fei.

Yip Siu-fei grew up in an ordinary middle-class family.  He went to attend university in Australia.  In 1991, he began to teach at Hong Kong University.  Every time that some new statistics appear, his name would appear in the newspapers.  He uses his knowledge of statistics to analyze the social phenomena contained in population growth rates, suicide rates, birth rates, population aging and other statistics.  As for estimating crowd sizes in marches, 7/1 last year was his first time and 7/1 this year was his second time.  Two months ago, he discussed the method of calculation with academics in Taiwan.

"This number is important and it is in my area of specialty.  That is why I did it."  Yip Siu-fei said that he disagreed with the 530,000 number, but he resolutely believes that the democrats did not engage in 'exaggeration.'  The estimation of crowd size needs specialized knowledge and experience.

This interview was conducted on Monday.  That afternoon, he learned that the Civil Human Rights Front was going to hold a press conference about the number of marchers.  At that time, he hesitated and asked our reporter to delay the publication of this interview.  He said that he does not doubt the scientific basis of his data, but he was concerned that the timing was too sensitive because it creates the impression of intentional opposition: "We are only academics.  We don't want to engage in political fights.  I only want to find out what the data say."

As an ordinary Hong Kong citizen, my heart breaks for Yip Siu-fei (whom I have never met with or spoken to).  What kind of world are we living in or will live in?  If you thnk that you are a Hong Kong pan-democrat, do you feel proud and righteous?

As academics, Yip Siu-fei, Robert Chung, Michael DeGolyer and Thomas Lee may feel that it is not appropriate to conduct a discourse about data analysis and statistical methodology on a Commercial Radio call-in program or at the Yahoo! forum.  But why should they take all that abuse in silence for simply wanting to "find out what the data say"?  If so, then this is a truly miserable and hopeless society that we live in.

It so happens that I am a professional statistician like them, I don't have an academic appointment and I don't have a professional position in Hong Kong.  I also run a blog.  Therefore, I speak out in a way that they may not want to (and this is not a criticism of them at all: my favorite all-time quotation is "I learned long ago never to wrestle with a pig. You get dirty and besides, the pig likes it.").

The essence of the current problem is this: Yip Siu-Fei, Robert Chung, Michael DeGolyer and Thomas Lee have disclosed the details of their data, methodology and assumptions on one side.  On the other side, the Civil Human Rights Front has only said that they used a "head count" method and refused to discuss their adjustment techniques.  Why?  Civil Human Rights Front spokesperson Lee Cheuk-yan said, "It is too complicated."

No.  Nothing is ever too complicated, especially when so much grief is thrown on serious academic scholars who only want to seek the truth.

That is one reason why I pursue this issue.

Maybe the designated Civil Human Rights Front spokesperson Lee Cheuk-yan really knows about an ingenious way of estimating the true attendance at demonstration marches.  It would seem that the academic scholars have got it all wrong previously.  Shouldn't the academic scholars learn "The Way" so they can follow this methodology and do all future surveys correctly?  And then we will live in peace and harmony forever afterwards without the contentiousness.

How about that?

Lee Cheuk-yan, how about telling us what you counted and how you made the adjustment to arrive at 250,000 plus?