Wilcox.test 为相同的数据返回两个不同的值

Question

我有一个包含两组 A 和 B 的 df。

group <- c(A, A, A, A, A, A, B, B, B, B, B, B, B, B, B, B, B)
measure <- c(3.4, 2.76, 3.37, 2.66, 2.73, 3.91, 9.06, 4.87, 3.4, 3.01, 1.95, 1.88, 1.85, 1.77, 1.77, 0.95, 0.25)
df <- data.frame(group, measure)

如果我说

wilcox.test(measure ~ group, data = df)

我得到 p = 0.06（带有关于关系和确切值的警告）。但是，如果我使用完全相同的数据创建两个向量 A 和 B

 A <- c(3.4, 2.76, 3.37, 2.66, 2.73, 3.91)
 B <- c(9.06, 4.87, 3.4, 3.01, 1.95, 1.88, 1.85, 1.77, 1.77, 0.95, 0.25)

然后说

wilcox.test(A,B, paired = FALSE, alternative = "two.sided")

我得到 p = 0.1908（关于领带的相同错误，我明白了）。这两个值都不是“重要的”，但差异令人不安。我究竟做错了什么？？ TIA

Answer 1

无法重现该问题。我认为在第一种情况下，您替换的是变量 A 和 B 而不是标签“A”和“B”

A <- c(3.4, 2.76, 3.37, 2.66, 2.73, 3.91)
B <- c(9.06, 4.87, 3.4, 3.01, 1.95, 1.88, 1.85, 1.77, 1.77, 0.95, 0.25)
wilcox.test(x=A, y=B)

    Wilcoxon rank sum test with continuity correction

data:  A and B
W = 46.5, p-value = 0.1908
alternative hypothesis: true location shift is not equal to 0

Warning message:
In wilcox.test.default(x = A, y = B) :
  cannot compute exact p-value with ties
 

group <- c("A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B")
measure <- c(3.4, 2.76, 3.37, 2.66, 2.73, 3.91, 9.06, 4.87, 3.4, 3.01, 1.95, 1.88, 1.85, 1.77, 1.77, 0.95, 0.25)
df <- data.frame(group, measure)
wilcox.test(measure ~ group, data = df)

    Wilcoxon rank sum test with continuity correction

data:  measure by group
W = 46.5, p-value = 0.1908
alternative hypothesis: true location shift is not equal to 0

Warning message:
In wilcox.test.default(x = c(3.4, 2.76, 3.37, 2.66, 2.73, 3.91),  :
  cannot compute exact p-value with ties