试验设计类型之无法考察交互作用的多因素设计:嵌套设计

摘要

This paper aimed to introduce a type of the multifactor experimental designs with no interaction: nested design.Two different kinds of nested designs were summarized by introducing four concrete actual cases: ①There was a nested relationship between the experimental factors in natural;②According to the influence of the experimental factors on the evaluating indicator, the experimental factors had a primary and secondary ones.Through the analysis of the example, an obvious characteristic of the nested design was revealed that the levels of the factors were "vertical linkage" rather than "horizontal combination".From the point of view of the proper and reasonable application, the matters need attention when using nested design are also pointed out in the paper.%本文目的是介绍一种无法考察交互作用的多因素设计类型,即嵌套设计.通过介绍四个具体实例,总结出两种不同的嵌套设计:①试验因素之间具有自然属性上的嵌套关系;②试验因素对评价指标的影响存在主次关系.通过对实例的解读,揭示嵌套设计具有一个其他多因素设计所缺乏的明显特点,即因素的水平是"纵向连接"而不是"横向组合"的;并从正确合理应用的角度考量,指出了使用嵌套设计时应注意的事项.