假设我有这个数据集：
require(rms)
newdata <- data.frame(eduattain = rep(c(1,2,3), times=2), dadedu=rep(c(1,2,3),each=2),
random=rnorm(6, mean(1000),sd=50))
我将因变量和自变量都转换为因子
newdata$eduattain <- factor(newdata$eduattain, levels = 1:3, labels = c("L1","L2","L3"),
ordered = T)
newdata$dadedu <- factor(newdata$dadedu, levels = 1:3, labels = c("L1","L2","L3"))
并使用权重进行简单的有序逻辑回归：
model1 <- lrm(eduattain ~ dadedu, data=newdata, weights = random, normwt = T)
警告信息：
In lrm(eduattain ~ dadedu, data = newdata, weights = random, normwt = T) :
currently weights are ignored in model validation and bootstrapping lrm fits
我有理由相信，如果使用权重，结果会完全不同。