Multi-level multiple Logistic regression analysis of the multi-value ordinal data collected from the unpaired design

DOI：10.11886/scjsws20191119003

 作者 单位 E-mail 凤思苑 天津医科大学公共卫生学院卫生统计学教研室 李长平 天津医科大学公共卫生学院卫生统计学教研室 胡良平 世界中医药学会联合会临床科研统计学专业委员会 lphu812@sina.com

本文目的是介绍多值有序资料多水平多重logistic回归分析方法。此法是在层次结构数据的基础上，构建多值有序因变量随一组自变量变化而变化的回归模型。具体的做法如下：①先介绍有关的基本概念；②呈现待分析的数据结构；③扼要介绍回归模型的构建与求解；④详细介绍如何使用SAS的GLIMMIX和NLMIXED两个过程来拟合此回归模型，并对相关结果进行解释和比较；⑤讨论多水平结构数据下拟合累积logistic回归模型时需注意的问题。

The purpose of this study was to introduce the multi-level multiple logistic regression analysis of the multi-value ordinal data. This method was based on the hierarchical data to build a regression model with the multi-value ordinal dependent variables changing with a group of independent variables. The specific methods were as follows. First, introduced the basic concepts. Second, presented the data structure to be analyzed. Third, briefly introduced the construction and solution of the regression models. Fourth, introduced in detail of how to use the procedures of GLIMMIX and NLMIXED of SAS software to fit the regression model, and explained and compared the relevant results. Last, the problems that should be paid attention to in cumulative logistic regression model fitting under the multi-layer structure data were discussed.
附件