混合效应线性模型与单因素方差分析在重复测量数据中的应用比较

The Application and Comparion of Mixed Effects Linear Model with Single Effdcts of Variance in Repeated Measures Data

目的:通过混合效应线性模型与单因素方差分析在重复测量资料中的应用比较,旨在说明两方法在处理重复测量资料时的应用特点。方法:用混合效应线性模型和单因素方差分析处理重复测量资料并比较。结果:混合效应线性模型和单因素方差分析都是处理重复测量资料的重要统计方法,前者在选择协方差结构下可对重复测量资料的固定效应和随机效应参数及协方差矩阵进行参数估计和统计检验,后者可对重复测量资料的固定效应做出统计推断。结论:混合效应线性模型是处理重复测量资料的有力方法,它对资料的协方差结构要求宽松,且结论可靠;单因素方差分析对资料的协方差结构有严格的限定。

Objective: In this article,a repeated measures data was analyzed in order to explain the application and characteristic of mixed effects linear model which was compared with single effects analysis of variance in order to account for it's use.Method: Make use of mixed effects linear model and single effect analysis of variance to deal with the repeated measure data.Results: Both mixed effects linear model and single analysis of variance were important statistics methods when dealing with repeated measure data,the former can carry the parametric estimate and hypothesis testing of fixed effects random effects through the choosed covariance;the latter can make a statistical inference.Conclusion: Mixed effects linear model was a good method dealing with repeated measures data,and it demand a loosed covariance structure and it's conclusion was credible;single effect analysis of variance have a strict limit of the covariance structure.