合理进行均值比较——含区组因素的两因素析因设计定量资料方差分析

Reasonably carry out mean value comparison:ANOVA of the quantitative data collected from the two-factor factorial design with a block factor

本文目的是介绍与均值比较有关的6个基本概念、计算方法、一个医学试验实例及其SAS实现。基本概念包括参数检验与非参数检验、前提条件、设计类型、交互作用以及方差分析的基本思想。计算方法涉及单因素两水平设计定量资料t检验、单因素多水平设计定量资料方差分析和含区组因素的析因设计定量资料方差分析。医学试验实例涉及如何考察可待因和针刺对男性受试者术后牙痛的影响。本文给出了采用SAS处理实例中定量资料的全过程,包括单因素分析、两因素分析和三因素分析,对输出结果作出解释,并明确给出合理选择统计分析方法进行定量资料差异性分析的关键点。

The purpose of this article was to introduce 6 basic concepts,calculation methods,a medical experiment example and its SAS implementation related to the mean comparisons.The basic concepts included the parametric test,non-parametric test,preconditions,design types,interaction and the basic ideas of the analysis of variance(ANOVA).The calculation method involved the t test of the quantitative data of single-factor two-level design,ANOVA of the quantitative data of single-factor multi-level design and factorial design with a block factor.An example of a medical experiment that examined the effects of codeine and acupuncture on postoperative toothache in male subjects.The paper gave the whole process of using SAS software to process the quantitative data in the example,including single-factor analysis,two-factor analysis and three-factor analysis,explained the output results,and the key points of reasonably selecting statistical analysis methods to analyze the differences of the quantitative data were clearly given.