我正在分析一个数据集,其中随时间测量了来自不同治疗方法的重复种群。我想用人口随机效应进行重复测量anova
,并用Mauchly球形度检验和p值惩罚(当违反球形度时需要)。
我正在尝试使用Anova
软件包中的car
命令,但是我很难添加随机效果。我可以使用car
软件包对球性和G-G校正进行Mauchly检验,并进行重复测量ANOVA。
以下是使用sleepstudy
数据集的示例代码
#change to wide format
sleepwide<-dcast(sleepstudy, Subject ~Days, mean, value.var="Reaction")
#rename columns
names(sleepwide) = c("Subject","x0", "x1", "x2", "x3", "x4", "x5",
"x6", "x7", "x8", "x9")
#create treatment column to used as fixed effect
sleepwide$treatment<-c('a','b','a','b','a','b','a','b','a','b','a','b',
'a','b','a','b','a','b')
#run a linear model for all time points with treatment as a fixed effect
z1<-lm(cbind(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9) ~ treatment, data=sleepwide)
#dictate the sample design
sample<-with(sleepwide, ordered(c("x0", "x1", "x2", "x3", "x4", "x5", "x6",
"x7", "x8", "x9")), levels=c("x0", "x1", "x2", "x3", "x4", "x5",
"x6", "x7", "x8", "x9"))
idata=data.frame(sample)
#run Anova in car package
ranova<-Anova(z1, idata=idata, idesign=~sample, type="III")
当我尝试以随机效果运行主题时,运行与上面相同的代码(例如,z2
代替上面的z1
),它将运行测试。
z2<-lme(cbind(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9) ~ treatment,
random=(~1|Subject), data=sleepwide)
有人能用随机效应进行莫希利检验+校正来进行重复测量方差分析的方法吗?