Randomization Analysis of Experimental Data: The Fisher Randomization Test
本文探讨了费希尔随机化检验的两种变体，指出该检验在逻辑上不可行，对理解实验数据的随机化分析有重要参考价值。
Abstract R.A. Fisher's classic text on the design of experiments is the principal source of inspiration for a mode of data interpretation that is usually characterized as randomization analysis. In Chapter III of this text, Fisher briefly commented on how to make a randomization test on some data generated by a Darwin experiment. Two variants of this randomization test are discussed in this article. The variant that is discussed in Section 4 may be regarded as the forerunner of all nonparametric tests. The original variant of the test is discussed in Section 6. The author concludes that the Fisher randomization test is not logically viable. Key Words: PrerandomizationPostrandomizationSufficiency principleLevel of significanceReference set