[原创]哪位知道怎么用SAS寻找配对样本啊？

而proc univariate用的秩和检验，是非参检验，具有更强的势能，不一样的。

I doubt that signed rank test has more power than t test since t test is based on some assumpations. Moreover, many people think that, signed test points to median but t test points to mean.

问题的关键不是用哪种检验（用哪种检验好，看分布，和检验的假设），请问如何实现配对？