摘要
本文目的是介绍一种特殊的处理多因素试验设计一元定量资料差异性分析方法,即结合分析法。通常情况下,处理多因素试验设计一元定量资料应采用方差分析。但是,此方法无法给出各对影响因素“重要性”的评价,也无法给出因素各水平的“效用值”,更无法给出“理想试验点”。本文通过对一个实例的全面解析,显示了SAS中的TRANSREG过程具有很强且多样性的变量变换能力,它集方差分析、回归分析和结合分析于一体,能够很好地处理不符合传统统计学要求的复杂资料,能够实现前述期望达到的目的。
The purpose of this paper was to introduce a special approach of the difference analysis to process the univariate quantitative data collected from the multi-factor experimental design, which was called the conjoint analysis. Normally, ANOVA should be used to deal with the quantitative data of multi-factor design. The variance analysis, however, could not give the evaluation of the "importance" about each of the influence factors and could not output the "Part-Worth Utility" of each level of every factor as well as could not produce the "ideal experimental point". The paper showed that the TRANSREG procedure in SAS had powerful and various abilities of the variable transformation. The procedure could achieved the previous aims, since it assembled the variance analysis and regression analysis and conjoint analysis together, and it could process the complex data of noncompliance with requirements of the traditional statistics.
作者
胡良平
Hu Liangping(Graduate School, Academy of Military Sciences PLA China, Beijing 100850, China;Specialty Committee of Clinical Scientific Research Statistics of World Federation of Chinese Medicine Societies, Beijing 100029, China)
出处
《四川精神卫生》
2019年第3期203-208,共6页
Sichuan Mental Health
基金
国家高技术研究发展计划课题资助(2015AA020102)
关键词
方差分析
回归分析
结合分析
BOX-COX变换
理想试验点
Analysis of variance
Regression analysis
Conjoint analysis
BOX-COX transformation
Ideal experimental point
