0

当我在使用 mantelhaen.test 时设置替代 =“更大”时，返回的置信区间上限为 Inf。为什么会发生这种情况，我可以解决它吗？我不确定这个问题是否源于编程（可能是我的数据格式？）或者它是否更多是一个统计问题。

这是我的输出：

       Mantel-Haenszel chi-squared test with continuity correction

    Mantel-Haenszel X-squared = 3.7413, df = 1, p-value = 0.02654
    alternative hypothesis: true common odds ratio is greater than 1
    95 percent confidence interval:
     1.192457      Inf
    sample estimates:
    common odds ratio 
      2.59944

下面是一些示例代码来演示：

    library(vcd)
    library(DescTools)

    example =("
                      Year Treatment Dead      Count
                      2015  Control   no          50
                      2015  Treat     no          60
                      2015  Control   yes          1
                      2015  Treat     yes          4
                      2016  Control   no          40                                  
                      2016  Treat     no         115                
                      2016  Control   yes          2                  
                      2016  Treat     yes          9 
                      2017  Control   no         142                 
                      2017  Treat     no          33      
                      2017  Control   yes          6                 
                      2017  Treat     yes          5                 
                      ")

    example = read.table(textConnection(example),header=TRUE)

    example =
      mutate(example,
             Year = factor(Year, levels=unique(Year)),
             Treatment = factor(Treatment, levels=unique(Treatment)),
             Dead = factor(Dead, levels=unique( Dead))
      )

    example.xtabs = xtabs(Count ~ Dead + Treatment + Year, 
                                  data=example)   


    ftable(example.xtabs)                      


    mantelhaen.test(example.xtabs) 
    mantelhaen.test(example.xtabs, alternative = "greater") #one sided - why 
                                                             #does CI = Inf
4

0 回答 0