是吗?我说说我的理解,你看看是我哪里理解错了
1. the power of a test is the prob. of correctly rejecting a false null.
2. When we reject a null, there are 2 possibilities: the null is true or false.
3. Thus, there are 2 outcomes: we reject a true null or we reject a false null.
4. These 2 outcomes consititute the whole of this event (the event is "we reject a null hypothesis"), so taking the sum of their prob., we will get 1.
Where is the problem??
我自己找出毛病了。。
The event is "when the null is false, what measures do we take?"
Accept or reject. That is, accept the false null (Type II error) and reject the false null (power of a test).
So the answer is obvious...
呵呵,discussion还是挺有意义的嘛!让我自我纠正了
[此贴子已经被作者于2009-6-8 0:56:18编辑过] |