a****g
发帖数: 8131 | 1 不好意思打搅, 如果一个实验性药物是beneficial的by hypothesis
比如, 发病率比对照组降低１０%
如果从conservative的角度出发,做２-tailed test, 这个sample size应该用怎样的
assumption, 这个proposal应该如何写呢? thanks
感谢各位 | s******h
发帖数: 539 | 2 Sounds like a two-sample proportion test. http://www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/binotest.htm
★ 发自iPhone App: ChineseWeb 7.8
【在 a****g 的大作中提到】
: 不好意思打搅, 如果一个实验性药物是beneficial的by hypothesis
: 比如, 发病率比对照组降低１０%
: 如果从conservative的角度出发,做２-tailed test, 这个sample size应该用怎样的
: assumption, 这个proposal应该如何写呢? thanks
: 感谢各位
| a****g
发帖数: 8131 | 3 i know it's a proportional test
the question is the hypothesis is superiority
with a conservative approach, we decide to make it 2-sided, instead of 1-
sided
how to justify that the sample size calculation hypothesis is different from
the main hypothesis
【在 s******h 的大作中提到】
: Sounds like a two-sample proportion test. http://www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/binotest.htm
:
: ★ 发自iPhone App: ChineseWeb 7.8
| w*******9
发帖数: 1433 | 4 the sample size based on the two-sided test is a "conservative" sample size
(larger than what is necessary when one sided test is needed). In practice,
the FDA requires two-sided test in superiority test.
from
【在 a****g 的大作中提到】
: i know it's a proportional test
: the question is the hypothesis is superiority
: with a conservative approach, we decide to make it 2-sided, instead of 1-
: sided
: how to justify that the sample size calculation hypothesis is different from
: the main hypothesis
| a****g
发帖数: 8131 | 5 这个sample size如何计算呢？
比如trt是20％ （by hypothesis）， control 是30％
it's easy to calculate 1-sided
when it comes to 2-sided, should i calculate the sample size?
can i do it in two calculations
trt 20% vs control 30%
and then trt 40% vs control 30%
and then pick the larger one?
thanks a lot
size
practice,
【在 w*******9 的大作中提到】
: the sample size based on the two-sided test is a "conservative" sample size
: (larger than what is necessary when one sided test is needed). In practice,
: the FDA requires two-sided test in superiority test.
:
: from
|
|
