如何向R语言的htest对象添加额外的自定义组件？

问题原因 #

• 字段添加不符合要求且未显式命名：你需要添加的是显著性水平alpha，但你修改代码时直接向列表传入了level，且没有指定元素的键名，正确的添加写法应为alpha = alpha。
• htest类的默认打印逻辑不会输出自定义字段：R内置的print.htest()方法只会输出预设的标准字段（统计量、p值、自由度、置信区间、数据名等），你新增的自定义字段不会出现在控制台打印结果里，但实际上已经存储在返回的列表对象中，你可以通过$运算符提取验证。

验证方法 #

你可以运行如下代码验证字段已成功存储：

# 调用自定义函数
res <- testvar(c(1,2,3,4,5), c(2,3,4,5,6))
# 提取自定义的alpha字段
res$alpha

完整修改代码 #

如果需要同时实现存储alpha字段+打印时显示alpha，可以参考如下修改：

testvar <- function(x1, x2, level = 0.95) {
  df1 <- length(x1) - 1
  df2 <- length(x2) - 1
  alpha = 1 - level
  s1 <- var(x1)
  s2 <- var(x2)
  dataname <- paste(deparse(substitute(x1)), "and", deparse(substitute(x2)))
  if(s1 > s2) {
    ts = s1/s2
    pval = 2 * (1 - pf(ts, df1, df2))
    conf = c(s1/(s2*qf(1-alpha/2, df1, df2)), s1/(s2*qf(alpha/2, df1,df2)))
    # 记录自由度匹配原生htest打印规范
    attr(ts, "parameter") <- c(df1 = df1, df2 = df2)
  }
  else if(s2 >= s1) {
    ts = s2/s1
    pval = 2 * (1 - pf(ts, df2, df1))
    conf = c(s2/(s1*qf(1-alpha/2, df2, df1)), s2/(s1*qf(alpha/2, df2,df1)))
    attr(ts, "parameter") <- c(df1 = df2, df2 = df1)
  }
  
  names(ts) = "F"
  attr(conf, "conf.level") <- level
  # 显式添加alpha字段，补全原生htest需要的method字段
  output <- list(statistic = ts, p.value = pval, conf.int = conf, 
                 data.name = dataname, alpha = alpha,
                 method = "F test to compare two variances")
  attr(output, "class") <- "htest"
  
  # 自定义打印逻辑，新增alpha输出
  cat("\n", output$method, "\n\n")
  cat("data: ", output$data.name, "\n")
  cat("F = ", round(output$statistic,4), ", df1 = ", attr(output$statistic,"parameter")[1], 
      ", df2 = ", attr(output$statistic,"parameter")[2], ", p-value = ", round(output$p.value,4), "\n")
  cat("alternative hypothesis: true ratio of variances is not equal to 1\n")
  cat(attr(output$conf.int,"conf.level")*100," percent confidence interval:\n", sep="")
  cat(" ", round(output$conf.int[1],4), " ", round(output$conf.int[2],4), "\n")
  # 新增显著性水平打印
  cat("significance level alpha: ", output$alpha, "\n\n")
  
  return(invisible(output))
}

内容的提问来源于stack exchange，提问作者feonyte