- 在線時間
- 31 小時
- 最後登錄
- 12-3-28
- 國民生產力
- 2
- 附加生產力
- 4
- 貢獻生產力
- 0
- 註冊時間
- 11-3-11
- 閱讀權限
- 10
- 帖子
- 211
- 主題
- 3
- 精華
- 0
- 積分
- 217
- UID
- 713414
 
|
回復 eviepa 的帖子
Eviepa
I've been extremely busy these two days with my daughter's coming back and my mother-in-law's birthday. But as my family is going on a holiday starting tomorrow, I have no choice but to rush out some points for you. I do not have time for some detailed arguments your questions warrant but I am sure some other friends here would be happy to correct or supplement them. I am also thrilled to see here Judy whom I have never ‘met’. Your question, like the one from stccmc and Judy certainly will find no easy answer unless we can get hold a copy of Tsang's report. However, based on the two statistics courses I attended at university and the information in the links other friends provided here, I would have been extremely surprised that the statisticians posted at EDB which sponsored the study had failed to pick up some obvious pitfalls in the sampling design as some people here suspect. First, I am not sure the purpose of the study was an ambitious multi-regression model to find out all the factors contributing to better academic results. It was said, he purpose was just to “分析語文分流對五項升學階梯的「勝算率」時，學生的五項背景因素均透過統計方法加以控制，包括(1) 學生小六派位成績、(2) 學生家庭社經地位、(3) 學生性別、(4) 同屆同學的平均小六派位成績，以及(5) 同屆同學的平均家庭社經地位。根據這些因素，研究探討相同背景的香港學生，被分派到中文或英文中學就讀對他們在多個升學階梯的「勝算率」所產生的影響。
It is only natural to suspect the purpose of EDB’s sponsorship was somewhat political, i.e. to look for statistical proof for the merit of using Chinese as the teaching medium. Frankly, this should have been the outcome expected by a reasonable man like Eviepa, or even me.
Regarding the questions on the sampling raised by stccmc and Judy, here is my two cents. The total sample is very large but it does not mean all the samples were used in every part of the statistical analysis. As I suspect, the study was more a comparison exercise by classifying the students by academic standing, socio-economical background and sex, and then tracing them from the point they are separated into CMI and EMI schools to the point they go into universities. The large total sample size was just convenient as EDB has them in their database. But the large sample enabled an adequate size of each sub-sample for each control factors. That is why he could track the paths of 起點相同的學生。I suspect some simple correlation analysis was also included as I saw some comments from the interview saying socio-economic background did not affect the value-adding during the secondary school stage as Tsang suspected the factor was already built into a student’s academic standing before the secondary stage.
Eviepa's concern whether Tsang’s report was an apple to apple comparison as top students were relatively few at CMI school. I tend to agree but I suspect the researcher would have been aware of this as well. I can only presume the really big fish were left out in the study which could only tackle with smaller big fish and medium fish. By the same token, the really small fish should have been excluded as well. However, the tracking of those really big fish and really small fish are only academic, no matter what the academics say. Few parents, as Eviepa put it, would have the wisdom to forgo the chance of placing the really big fish in SPCC or DGS or other top schools and even fewer parents would have the capability to place the really small fish in an EMI school other than schools like 嘉道理。
Let’s have some faith in our academics who are responsible for teaching people like ourselves and our kids, at least in the handling of some basic sampling design. Our faith was also built on the reasonable assumption that the government was only more than happy to find fault with the study and don’t forget they have tons of very highly paid statisticians and some other friendly academics with them. And last, my husband told me that, thinking like a government official, he would never sponsor a so-called neutral study. There was always an agenda behind such a study. And the secret of obtaining a favourable result is not to interfere with the researcher, like a former vice-chancellor of HKU stupidly did, but to pick a researcher who has a similar position like your own and think similarly. Tsang’s report was divided into 3 stages and it went well initially for the government. Actually, one can tell Tsang did feel sad to see his own research result about the final stage. He said,「中中學生在初中雖享受到『輕鬆易學』的優勢，但卻同時大大 局限了他們學習英文的相對機會。面對大學語文門檻，更只能望門興嘆，對他們實在很不公平。」
You said ,“如果越大的塘，有越高的培養學生能力”-- 有點牽強；and you also said, ”平均來說，塘越大，增值指標越高；塘越小，增值指標越低” – of course not. As I said, we can have different value added scores for ponds of the same size. For example, 格仔裙 seems to have a higher value score than most other girls’ schools but I wouldn’t say their pond size is bigger than some other girls’ schools, for example DGS.
We can just stick to the seemingly credible results of Tsang’s research and rest our arguments unless someone among us has the means of getting hold of a report copy.
Finally, 李天命is not my idol. 勢利小女人如我係唔會封一個淨係講吓邏輯入門、寫幾句唔多會流傳後世嘅新詩，到退休都仲係講師嘅人做偶像嘅。
雙儿
[ 本帖最後由 雙儿 於 11-6-29 20:03 編輯 ] |
|
發表於 11-6-2 00:40
