r*******e
发帖数: 510 | 1 算Relative risk (RR)
不用normal approximation, 算WALD CI （这个应该是BASED ON CHI-SQUARE的吧）。
哪位同学解释下， 或给点参考资料。
谢谢 |
r*******e
发帖数: 510 | 2 我自己顶。 希望有大牛看到。
知道有些大牛每天来看的。 |
r*******e
发帖数: 510 | 3 为什么这么多人浏览， 没人帮助呀。
谢谢大家了。 |
s*******1
发帖数: 146 | 4 First, calculate the Standard error, then calculate 95%CI.
95%CI= mean +/-1.96*SE |
a*****n
发帖数: 8 | 5 我怎麽覺得Wald interval 不是 BASED ON CHI-SQUARE的吧 ？ |
a*****n
发帖数: 8 | 6 這個是based on normal distribution 吧？
【在 s*******1 的大作中提到】
: First, calculate the Standard error, then calculate 95%CI.
: 95%CI= mean +/-1.96*SE
|
D******n
发帖数: 2836 | 7 I have a fundamental question. What is a CI for a ratio? Does it necessarily
exist in frequentist sense? how about bayesian |
b*******r
发帖数: 152 | 8 if it's NOT a normal dist, how can use 1.96 for 95%?
【在 s*******1 的大作中提到】
: First, calculate the Standard error, then calculate 95%CI.
: 95%CI= mean +/-1.96*SE
|
j*****e
发帖数: 182 | 9 First, the square of a standard normal r.v. is chi-squared with df=1.
Second, in SAS, CI of relative risk is constructed based on the asymptotic
normality of log(estimated RR).
It is impossible to have a Wald CI without some sort of normality assumption
. This is a concept issue. |
g******n
发帖数: 339 | 10 Support this.
WALD/SCORE/likelhood ratio tests are asymptotically equivalent, all are
based on asymptotically normality of the test statistics.
assumption
【在 j*****e 的大作中提到】
: First, the square of a standard normal r.v. is chi-squared with df=1.
: Second, in SAS, CI of relative risk is constructed based on the asymptotic
: normality of log(estimated RR).
: It is impossible to have a Wald CI without some sort of normality assumption
: . This is a concept issue.
|
x***x
发帖数: 3401 | 11 CI for ratio 是有的啊. 为什么和是不是frequentist有关系?
necessarily
【在 D******n 的大作中提到】
: I have a fundamental question. What is a CI for a ratio? Does it necessarily
: exist in frequentist sense? how about bayesian
|
