如何在箱线图中展示双因素ANOVA的紧凑字母显示(CLD)

问题背景 #

使用multcompLetters4()函数获取4组数据双因素ANOVA的紧凑字母显示(CLD)后，绘制带字母的箱线图时，geom_text()无法区分不同的Tx分组，字母全部显示在同一Group位置，无法对应到各自的填充组。

原始数据与代码 #

数据 #

data1 <- structure(list(Tx = c("T", "T", "T", "T", "T", "T", "T", "T", 
"T", "T", "E", "E", "E", "E", "E", 
"E", "E", "E", "E", "E", "T", "T", 
"T", "T", "T", "T", "T", "T", "T", "E", "E", "E", 
"E", "E", "E", "E", "E", "E", 
"E"), Group = c("I", "I", "I", "I", "I", "I", 
"I", "I", "I", "I", "I", "I", "I", "I", "I", "I", "I", 
"I", "I", "I", "G", "G", "G", "G", "G", "G", "G", "G", 
"G", "G", "G", "G", "G", "G", "G", "G", "G", "G", "G"
), Nb = c(24, 
21, 13, 18, 11, 12, 10, 17, 22, 13, 3, 6, 12, 11, 1, 5, 10, 10, 
10, 2, 18, 15, 15, 16, 19, 16, 22, 23, 19, 11, 10, 11, 16, 5, 
15, 14, 14, 17, 16)),  class = c("tbl_df", "tbl", 
"data.frame"), row.names = c(NA, -39L))

原始绘图代码 #

aov1 <- aov(Nb~Tx*Group, data=data1)

tukey <- TukeyHSD(aov1)

tukey.cld <- multcompLetters4(aov1, tukey)
cld <- as.data.frame.list(tukey.cld$`Tx:Group`)
cld

summary <- data1 %>%
  group_by(Tx,Group) %>%
  summarise(
    w=mean(Nb),
    sd=sd(Nb),
    max=max(Nb)) %>% 
  arrange(desc(w)) 

summary$Tukey <- cld$Letters

ggplot(data1, aes(x = Group, y = Nb, fill = Tx)) +
  geom_boxplot(show.legend = TRUE, outlier.shape = NA) +
  geom_text(data = summary, aes(x = Group, label = Tukey, y = max), inherit.aes = FALSE)

解决方法 #

核心是让文本位置与箱线图的分组偏移匹配，使用position_dodge()对齐不同Tx分组的字母，同时确保分组逻辑一致：

修改后的完整代码 #

aov1 <- aov(Nb~Tx*Group, data=data1)

tukey <- TukeyHSD(aov1)

tukey.cld <- multcompLetters4(aov1, tukey)
cld <- as.data.frame.list(tukey.cld$`Tx:Group`)
cld

summary <- data1 %>%
  group_by(Tx,Group) %>%
  summarise(
    w=mean(Nb),
    sd=sd(Nb),
    max=max(Nb)) %>% 
  arrange(desc(w)) 

summary$Tukey <- cld$Letters

# 调整文本位置匹配箱线图分组
ggplot(data1, aes(x = Group, y = Nb, fill = Tx)) +
  geom_boxplot(show.legend = TRUE, outlier.shape = NA, position = position_dodge(width = 0.75)) +
  geom_text(data = summary, 
            aes(x = Group, label = Tukey, y = max + 1, fill = Tx),
            inherit.aes = FALSE,
            position = position_dodge(width = 0.75),
            size = 4)

关键调整说明 #

• position_dodge(width = 0.75)：箱线图默认分组偏移宽度为0.75，让文本使用相同参数，即可让每个Group下不同Tx的字母对应到各自箱线正上方。
• y = max + 1：给字母位置添加小偏移，避免与箱线图最大值重叠，可根据数据范围调整数值。
• 保留fill = Tx：确保ggplot识别分组逻辑，正确应用位置偏移。

内容的提问来源于stack exchange，提问作者arnaudm