双变量横断面资料的多水平模型分析  

Analysis of Multilevel Modeling for Cross Sectional Bivariate Data.

作　　者：李智文[1] 陶育纯[1] 刘钢[1] 

机构地区：[1]白求恩医科大学卫生统计教研室,130021

出　　处：《中国卫生统计》1999年第6期334-337,共4页Chinese Journal of Health Statistics

摘　　要：目的　分析双变量横断面资料两个反应变量之间的相关性随协变量变化的规律。方法　双变量多水平模型 ,最小二乘法。结果　双变量多水平模型可以估计各水平两个变量的方差 协方差阵 ,据此可以计算出相关系数随协变量变化的函数式。结论　双变量多水平模型是分析双变量横断面资料相关性的有力工具 ,克服了用简单直线相关处理这类资料的局限性。Objective To analyze the changing rules of correlativity for cross sectional bivaritate data which is hierarchically structured.Methods Multilevel bivariate models,Ordinary least square estimation.Results Multilevel models can present the variance covariance metrics of two dependent variables in every levels,and make out the functional expresses of correlation coefficient with covariates.Conclusion Multilevel bivariate models is effective in analyzing the correlativty of bivariate data,and ordinary least square method could not compared with it in this sense.

关 键 词：双变量 多水平模型 相关性 体重 身高 

分 类 号：R311[医药卫生—基础医学]