与R的比例两两比较

Question

我是 R 新手。所以这是我的问题，我想进行比例的成对比较。在我的数据中，我有三个会话和每个会话的时间段。我添加了每个时隙的计数，并知道 M1 会话的样本大小 = 181、M2 = 67 和 M3 = 106。 我想比较每个会话、每个时段的比例。就像时间段 18h-19h 一样，我有来自 M1vsM2 、M2 vs M3 和 M1 vs M3 的 p 值。

data <- tribble(
  ~Horaire, ~SessionM1, ~NbM1, ~SessionM2, ~NbM2, ~SessionM3, ~NbM3,
"18h-19h","M1", 6, "M2", 7, "M3", 6,
"19h-20h","M1", 6, "M2", 10, "M3", 17,
"20h-21h","M1", 14, "M2", 10, "M3", 16,
"21h-22h","M1", 23, "M2", 5, "M3", 9,
"22h-23h","M1", 15, "M2", 9, "M3", 4,
"23h-00h","M1", 19, "M2", 4, "M3", 6,
"00h-01h","M1", 19, "M2", 5, "M3", 10,
"01h-02h","M1", 21, "M2", 3, "M3", 7,
"02h-03h","M1", 8, "M2", 5, "M3", 7,
"03h-04h","M1", 15, "M2", 4, "M3", 10,
"04h-05h","M1", 15, "M2", 4, "M3", 9,
"05h-06h","M1", 20, "M2", 1, "M3", 5)

PS：我编辑了数据。

有人可以帮忙吗？谢谢

我希望获得每个时间段内每种会话组合的 p 值。谢谢

Answer 1

您可以对每一行使用

pairwise.prop.test()

。

t(apply(data[grep("NbM", names(data))], 1, \(m) {
  p <- pairwise.prop.test(x = m, n = c(181, 67, 106))$p.value
  p[!is.na(p)]
})) |>
  as.data.frame() |>
  setNames(combn(1:3, 2, \(x) paste0('M', x, collapse = 'vs')))

#    M1vsM2 M1vsM3 M2vsM3
# 1   0.166  0.771  0.771
# 2   0.005  0.001  1.000
# 3   0.289  0.232  1.000
# 4   1.000  1.000  1.000
# 5   0.431  0.431  0.121
# 6   0.796  0.707  1.000
# 7   1.000  1.000  1.000
# 8   0.447  0.483  0.803
# 9   1.000  1.000  1.000
# 10  1.000  1.000  1.000
# 11  1.000  1.000  1.000
# 12  0.096  0.211  0.482