卡方检验结果的解释

如何解决卡方检验结果的解释

我正在探索我的实验结果（在 R 中），结果可以显示为 5x4 列联表。我已经运行了独立性的卡方检验，它给出的 p 值远低于 0.05，这意味着治疗和结果之间存在关系。

x = structure(c(17479L,256L,1332L,66L,1919L,4242L,87L,394L,26L,761L,6359L,40L,349L,22L,823L,5666L,75L,325L,39L,809L),.Dim = 5:4,.Dimnames = list(c("G1.G0","early.S","S","late.S.G2","G2.M"),row.names = c("Control","Treatment1","Treatment2","Treatment3")),class = "table")
ch.t = chisq.test(x)
ch.t
ch.t$residuals

输出

Pearson's Chi-squared test
data:  x
X-squared = 236.55,df = 12,p-value < 2.2e-16

               row.names
               Control Treatment1 Treatment2 Treatment3
  G1.G0      1.3744793 -4.2432348  1.5180210 -0.2012234
  early.S    1.3854929  3.2597551 -4.8551234 -0.2396874
  S          2.9011770  4.0127320 -4.4966957 -3.9322811
  late.S.G2 -1.4033400  1.2079478 -1.1821301  2.6092210
  G2.M      -6.1967331  7.5868936  0.9130591  3.0832144

我坚持“成对”比较。例如，与对照组相比，我如何判断处理 1 对结果“S”是否具有统计上的显着影响？我应该对每对治疗进行 chi-test 还是有办法从这个单一测试中“提取”答案？

解决方法

使用 chisq.posthoc.test 包 install.packages("chisq.posthoc.test")，您可以基于 T. Mark Beasley & Randall E. Schumacker (1995) .

解释：G1.G0 与对照组、治疗 1、治疗 2 之间存在显着相关性，但与治疗 3 之间没有显着相关性。

Dimension        Control    Treatment1 Treatment2 Treatment3
G1.G0   p-Value  0.000063   0.000000   0.001368   1.0000000

换句话说，控制、治疗 1、治疗 2 都对 G1.G0 有显着影响。但处理3对G1.G0没有显着影响。

x = structure(c(17479L,256L,1332L,66L,1919L,4242L,87L,394L,26L,761L,6359L,40L,349L,22L,823L,5666L,75L,325L,39L,809L),.Dim = 5:4,.Dimnames = list(c("G1.G0","early.S","S","late.S.G2","G2.M"),row.names = c("Control","Treatment1","Treatment2","Treatment3")),class = "table")
ch.t = chisq.test(x)
ch.t
ch.t$residuals

library(chisq.posthoc.test)

chisq.posthoc.test(x)

   Dimension     Value   Control Treatment1 Treatment2 Treatment3
1      G1.G0 Residuals  4.662387 -10.799208   3.981811 -0.5225417
2      G1.G0  p values  0.000063   0.000000   0.001368  1.0000000
3    early.S Residuals  1.995708   3.522918  -5.407864 -0.2643080
4    early.S  p values  0.919313   0.008536   0.000001  1.0000000
5          S Residuals  4.282599   4.444246  -5.132860 -4.4437546
6          S  p values  0.000369   0.000176   0.000006  0.0001770
7  late.S.G2 Residuals -2.013868   1.300592  -1.311795  2.8664955
8  late.S.G2  p values  0.880468   1.000000   1.000000  0.0830090
9       G2.M Residuals -9.382259   8.618532   1.068996  3.5737213
10      G2.M  p values  0.000000   0.000000   1.000000  0.0070390