BMJ 统计问题(3)：关于P值，哪种说法正确？

《英国医学会杂志》（BMJ）自2008年9月开始至2015年由两位流行病与统计学专家不间断地出了300多期statistical question系列。在这个系列中，两位学者每次出一道统计学选择题，进行选择并解释。现在我精选300道Statistical Question，形成中文版，请有兴趣的朋友们进行回答。

Which one statement is true of the P value?

a) It is the probability that the null hypothesis is true

b) It is the probability that the alternative hypothesis is true

c) It is the probability that the null hypothesis is false

d) It is the probability that the alternative hypothesis is false

e) It is the probability of the given result or more extreme results if the null hypothesis were true.

Answer：

Statistical testing doesn’t quite do what clinicians would wish for. We are interested to know how near the truth a study result is likely to be. Standard statistics start from a the oretical probability distribution and give answers to questions such as: “If this study were based on completely random data, how unusual would these actual results be?”

Answer a) is frequently assumed to be the practical interpretation of a P value. It is not strictly correct, but for most people (except under exam conditions) that will not matter much when working to this assumption.

中文解释：

统计检验并不能完全满足临床医生的期望。我们有兴趣知道一项研究结果有多么接近真实情况。标准统计方法从理论概率分布开始，并给出以下问题的答案：“如果这项研究是基于完全随机的数据，那么这些实际结果有多异常？”

选项A通常被认为是P值的实际解释。但是，这并不是严格正确的，但是对于大多数人（在考试条件下除外），按照这一假设进行操作并没有什么问题。

答案是选择 E，我们假设H0是对的，然后我们有一个样本，看在H0成立的时候，样本以及更极端样本的发生概率。选项A,正如这篇文章所言，考试的时候不能这么理解，但对于医务工作者（公共卫生除外），这么理解也是可以的。