Significance tests with Restricted Randomization Design
研究了在受限随机化分配规则下，将实验对象依次分配到两种处理中时产生的显著性检验，并证明了基于瓮模型的检验在温和条件下渐近正态。
Experimental subjects are assigned to one of two treatments sequentially by restricted randomization assignment rules. Significance tests generated from these rules are discussed. Under rather mild conditions on the test scores the asymptotic null permutational distribution of tests based on the urn design (Wei, 1977, 1978a) is shown to be normal.