weakest

Sorry, don't really get the meaning of "inconclusive"...Do you mean that no result could be reached because there are many reasons could cause "肝風", as such, even if "肝風" cause cancer, we can't conclude that believing Christianity is the cause of having cancer?

weakest

原帖由 Jennifer 於 2008-8-13 00:26 發表
大概係咁，你要有兩個人背景、年齡、性別、生活習慣一樣，但係一個返教會，一個唔返，
黎比較佢地邊個容易患癌，結論才會比較 conclusive。其實現今的癌症研究，根本好難鎖定
一個成因，好似吸煙可能引致癌症，但有些人不吸煙，生活習 ...

那些"吸煙引致癌症"的警告乃inconclusive?

weakest

"Correlation does not = causation"
---> I get this concept. However, what method is used to check that A cause B?? In your example:

"To see wether he is right, we will need to see the change, if any, when people enter and leave the religion.  A special phenomenon of hosptital is that people tend to leave the hospital as a healthier person.  This is not applicable to religion as they claim that only those who are sick will leave.  

The other phenomenon is that people who is entering the hospital tends to get heathier. Now THIS, is observable."

This is observable, but could it be interpreted as correlation also?


And back to the beginning question: "conclusive" means a certain hypothesis have a large enough sample that supports it?

weakest

謝賜教.

Then, can i say correlation (+ve or -ve) is the necessary condition for causation?

[ 本帖最後由 weakest 於 2008-8-19 02:27 編輯 ]