1∶r配对设计二值资料一水平多重Logistic回归分析

One-level multiple Logistic regression analysis of the dichotomous choice data collected from the 1∶r matched pairs design

本文的目的是介绍“1∶1”“1∶2”和“1∶r”配对设计二值资料一水平多重Logistic回归分析方法。统计分析的前提条件是资料值得分析,而确保配对设计二值资料值得分析的重要基础是“配对设计”科学合理。为此,本文从六个方面阐述“配对设计的要领”,不仅介绍了构建“1∶1”“1∶2”和“1∶r”配对设计二值资料一水平多重Logistic回归模型的方法,还通过三个实例,介绍了采用SAS软件实现统计分析及结果解释的详细步骤。

The paper introduced the approach of multiple Logistic regression analysis for binary data collected from the"1∶1","1∶2"and"1∶r"paired design.The prerequisites of the statistical analysis were that the data were worth analyzing,in order to meet the demand mentioned before,the paired design should be scientific and reasonable.Therefore,this paper emphasized the essentials of paired design from six aspects.This paper not only introduced the approach of constructing the one-level multiple Logistic regression model of the binary data collected from the"1∶1","1∶2"and"1∶r"paired design,but also introduced the detailed methods of using SAS software package to realize statistical analysis and the result interpretation through three examples.