对王颉修同志的“对宋哲和著《农药药效试验的设计与分析》一书的若干意见”(以下简称《意见》)一文中的若干论点笔者尚有不同的看法,特提出与王颉修同志商榷。 1.《意见》一文中指出:“迄今,没有发现否定它(笔者注:指t测验法)的作用的文献。而《分析》一书作者没有根据的用臆想的说法否定了t测验的作用。”笔者认为这种说法是完全不符合事实的。t测验法应用于两个相互独立的样本平均数间差异显著性测验,其作用在过去和现在并没有人否认。但多个处理间互相比较是否适宜于应用t测验法,自五十年代以来,这个问题已受到了普遍的怀疑。为了克服t测验 Comrade Wang Jiexiu’s “Some Opinions on Song Zhe and His Opinion on the Design and Analysis of Pesticide Efficacy Test” (hereinafter referred to as “Opinion”) I still have different opinions in this article, and I hereby put forward the proposal of comrade Wang Jiexiu Discuss. 1. In the Opinion article it was stated: “To date, no literature has been found to undermine the validity of the t-test, which the author of the book in Analytical does not base on conjecture ”I think this is totally out of line with the truth. The t test was applied to two independent significance tests of mean difference between samples, and no effect was denied in the past or in the past. However, whether or not multiple treatments are more appropriate for applying the t-test has been a subject of widespread skepticism since the 1950s. To overcome the t test