家禽试验中重复测量数据的常用分析方法

Common Statistical Methods of Repeated Measurement Data in Poultry Experiments

在家禽试验中经常会需要对试验动物进行连续性观察而获得重复测量数据。对这类数据资料如果采用常规方差分析或者对重复测量数据的各时间点分别分析,其结果可能因为来自相同个体的多次观察数据间存在相关而加大误差,或无法充分反应数据连续观测的特性。针对动物试验中重复测量数据的统计方法,文章分析了重复测量数据的重复测量方差分析(RM-ANOVA)、多变量方差分析(MANOVA)法以及混合线性模型拟合法的适用条件,结合案例详述了如何选择合适的方法在R语言中实现对这类数据的分析,为家禽科学及相关专业涉及重复测量数据的统计分析提供参考。

Repeated measurement data are often obtained in poultry experiments,for the continuous observation of animals frequently occurs over time.For analyzing repeated measures,if routine ANOVA is used or the data of each time point are tested separately,the results may be more error-prone due to correlation between multiple observations from the same animal or may not adequately reflect the characteristics of continuous observations.Aiming at the problem of misuse of analysis methods for animal repeated measures data,this paper introduced RM-ANOVA,MANOVA and the linear mixed model analysis.The operate of this analysis in R language were demonstrated using an example.It could provide reference for poultry science and related majors involving repeated measures data analysis.