R数据处理与卡方检验

Question

我有一个数据框架：

gender   group   count    total
female    A       8        10
female    B       23       30
female    C       22       25
male      A       18       28
male      B       23       30
male      C       40       70

我应该如何处理数据并应用于性别之间的每一组的卡方检验？

Answer 1

首先，使用dput功能可以方便地复制数据。

dput(dat)
structure(list(gender = c("female", "female", "female", "male",
"male", "male"), group = c("A", "B", "C", "A", "B", "C"), count = c(8L,
23L, 22L, 18L, 23L, 40L), total = c(10L, 30L, 25L, 28L, 30L,
70L)), class = "data.frame", row.names = c(NA, -6L)).   

然后可以按组(或性别)使用by函数，如下所示

by(dat, dat$group, function(x) chisq.test(data.frame(x$count, x$total)))
dat$group: A

Pearson's Chi-squared test with Yates' continuity correction

data:  data.frame(x$count, x$total)
X-squared = 0.011266, df = 1, p-value = 0.9155

dat$group: B

Pearson's Chi-squared test

data:  data.frame(x$count, x$total)
X-squared = 0, df = 1, p-value = 1

dat$group: C

Pearson's Chi-squared test with Yates' continuity correction

data:  data.frame(x$count, x$total)
X-squared = 1.0981, df = 1, p-value = 0.2947

但是，通过查看该方法创建的2*2矩阵，确保其测试完全符合您的要求是值得的：

例如

> by(dat, dat$group, function(x)data.frame(x$count, x$total))[1]
$A
  x.count x.total
1       8      10
2      18      28