d*******7 发帖数: 118 | 1 大家帮忙看看这个问题哈:
What is the minimum sample size to detect a change in sensitivity from 85%
to 80% with type I error less than or equal to 0.05 and power larger than 95
%?
Thanks! | s*****r 发帖数: 790 | 2 试着回答一下。只有思路。
1)假定population是所有true positive, then sensitivity is the detection
probability. So the same size needed is the same as in a case that you need
to detect the difference between 85% and 80%.
2) now you need to convert to all population, i.e, you need to know the
disease prevalance rate.
To answer your question, it is uncliear in what kind of setting you want to
detect a change in sensitivity. I mean, how the experiment is set up?
95
【在 d*******7 的大作中提到】 : 大家帮忙看看这个问题哈: : What is the minimum sample size to detect a change in sensitivity from 85% : to 80% with type I error less than or equal to 0.05 and power larger than 95 : %? : Thanks!
| c****g 发帖数: 156 | 3 test for proportions, binomial distribution, effect size=5% | d*******7 发帖数: 118 | 4 I have the same thoughts as you. But could you please calculate the actual
number because somehow I can not get the right answer and I do not know why.
Thanks!
【在 c****g 的大作中提到】 : test for proportions, binomial distribution, effect size=5%
| s*r 发帖数: 2757 | |
|