论文部分内容阅读
We propose new,optimal methods for analyzing randomized trials,when it is suspected that treatment effects may differ in two predefined subpopulations.