Tukey HSD Test和Dunnett是常用的多组数据间的差异显著性检验方法。
Tukey's HSD（忠实显著性差异），也称作 Tukey HSD、WSD 或 Tukey(a) 检验，控制假阳性率判断族。这意味着，如果在 0.05 级别上执行检验，那么在执行所有成对比较时，获得一个或多个假阳性的几率是 0.05。
It reduces Type I error at the expense of Power. It is appropriate to use this test when one desires all the possible comparisons between a large set of means (6 or more means).
Post Hoc tests that assume equal variance.
It reduces Type I error at the expense of Power. It is appropriate to use this test when one desires all the possible comparisons between a large set of means (6 or more means).
Post Hoc tests that assume equal variance.
将一组处理与单个控制平均值进行比较的成对多重比较 T 检验。您可以选择第一个类别或最后一个类别作为默认控制类别。双向检验任何级别（除控制类别之外）的因素平均值是否不等于控制类别的因素平均值。
控制检验任何级别的因素平均值是否小于控制类别的因素平均值。
控制检验任何级别的因素平均值是否大于控制类别的因素平均值。
Post Hoc tests that do not assume equal variances
控制检验任何级别的因素平均值是否小于控制类别的因素平均值。
控制检验任何级别的因素平均值是否大于控制类别的因素平均值。
Post Hoc tests that do not assume equal variances
0. 做几组数据数据
set.seed(1080)
data <- data.frame(group = rep(c("P1", "P2", "P3"), each = 40),
values = c(rnorm(40, 0, 3),rnorm (40, 0, 6),rnorm (40, 1, 5)))
data$group<-as.factor(data$group)
head(data)
1. 先做ANOVA
model <- aov(values~group, data=data)
summary(model)
输出：
Df Sum Sq Mean Sq F value Pr(>F)
group 2 161.2 80.60 3.588 0.0307 *
Residuals 117 2628.5 22.47
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
2-1. Tukey HSD Test
TukeyHSD(model, conf.level=.95)
输出
Tukey multiple comparisons of means
95% family-wise confidence level
Fit: aov(formula = values ~ group, data = data)
$group
diff lwr upr p adj
P2-P1 1.066363 -1.4496236 3.582350 0.5744712
P3-P1 2.811853 0.2958665 5.327840 0.0244720
P3-P2 1.745490 -0.7704965 4.261477 0.2302467
2-2. Tukey投图
plot(TukeyHSD(model, conf.level=.95), las = 2)
3-1. Dunnett Test
install.packages("DescTools")
library(DescTools)
DunnettTest(x=data$values, g=data$group, control ="P2")
输出：
Dunnett's test for comparing several treatments with a control :
95% family-wise confidence level
$P2
diff lwr.ci upr.ci pval
P1-P2 -1.066363 -3.4395508 1.306825 0.5001
P3-P2 1.745490 -0.6276977 4.118678 0.1788
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
参考文章：
https://www.r-bloggers.com/2021/08/how-to-perform-dunnetts-test-in-r/
https://www.r-bloggers.com/2021/08/how-to-perform-tukey-hsd-test-in-r/