Correct Positioning of Multiple Significance Labels on Dodged Groups in Ggplot

Correct positioning of multiple significance labels on dodged groups in ggplot

Although I really don't think this is a good idea for visualisation - here is a solution. If you use ggpubr, stay in the ggpubr syntax. And use faceting for subgrouping.

P.S. Try a table instead.

library(tidyverse)
library(ggpubr)

mydf <- filter(diamonds, color == "J" | color == "E")

comparisons <- list(c("Fair", "Good"), c("Fair", "Very Good"), c("Fair", "Premium"))

ggboxplot(mydf, x = "cut", y = "price", facet.by = "color") +
  stat_compare_means(
    method = "t.test", ref.group = "Fair", label = "p.format",
    comparisons = comparisons
  )

^{Created on 2020-03-20 by the reprex package (v0.3.0)}

Manually plotting significance relations between sub-groups on ggplot2 barplot

If you know on which barchart you want to add your significance labels, you can do:

library(ggsignif)
library(ggplot2)

ggplot(df, aes(x = activity, y = mean, fill = group)) +
  geom_bar(position = position_dodge(0.9), stat = "identity",
           width = 0.9, colour = "black", size = 0.1) +
  xlab("Activity") + ylab("Mean")+
  geom_signif(y_position = c(60,50), xmin = c(0.7,2), xmax = c(1,2.3),
              annotation=c("**", "***"), tip_length=0)

Does it answer your question ?

How to centre single bar position with multiple bars in position_dodge in ggplot2

OP, use position_dodge2(preserve="single") in place of position_dodge(preserve="single"). For some reason, centering bars/columns doesn't quite work correctly with position_dodge(), but it does with position_dodge2(). Note the slight difference in spacing you get when you switch the position function, but should overall be the fix to your problem.

Reproducible Example for OP's question

library(ggplot2)
set.seed(8675309)
df <- data.frame(
  x=c("A", "A", "A", "B", "C", "C"),
  grouping_var = c("Left", "Middle", "Right", "Middle", "Left", "Right"),
  values = sample(1:100, 6))

Basic plot with position_dodge():

ggplot(df, aes(x=x, y=values, fill=grouping_var)) +
  geom_col(position=position_dodge(preserve = "single")) +
  theme_classic()

When you use position_dodge2():

ggplot(df, aes(x=x, y=values, fill=grouping_var)) +
    geom_col(position=position_dodge2(preserve = "single")) +
    theme_classic()

stat_compare_means() for multiple groups in faceted ggplot data

Edit

The key thing here is that the comparisons are based on:

comparisons: A list of length-2 vectors. The entries in the vector are
either the names of 2 values on the x-axis or the 2 integers that
correspond to the index of the groups of interest, to be compared.

So you have to use interaction on the x-axis to be able to compare the boxplots per grouping. Here is a reproducible example:

library(ggplot2)
library(ggpubr)
library(dplyr)
df %>%
  ggplot(aes(x=interaction(subject, grouping),y=percent)) + 
  geom_boxplot(aes(group = interaction(subject, grouping)), outlier.shape = NA) + 
  geom_point(aes(color = grouping, group = grouping),size=1.75,position = position_dodge(width = 0.75))+
  scale_y_continuous(breaks = seq(0, 100, by = 25)) +
  scale_x_discrete("", limits = c("Four.A", "Four.B", "Four.C", "One.A", "One.B", "One.C", "Three.A", "Three.B", "Three.C", "Two.A", "Two.B", "Two.C"),
                   labels = c(rep(c("", "Four", "", "", "One", "", "", "Three", "", "", "Two", ""), 3))) +
  stat_compare_means(comparisons = list(c("Four.A","Four.B"),c("Four.A","Four.C"),c("Four.B","Four.C"), 
                                        c("One.A","One.B"),c("One.A","One.C"),c("One.B","One.C"),
                                        c("Three.A","Three.B"),c("Three.A","Three.C"),c("Three.B","Three.C"),
                                        c("Two.A","Two.B"),c("Two.A","Two.C"),c("Two.B","Two.C")),
                     method = "t.test",
                     paired=F,
                     tip.length=0,step.increase = 0.05) +
  facet_wrap(~ family) +
  theme_minimal()

^{Created on 2022-08-31 with reprex v2.0.2}

---

Old answer

Maybe you want something like this. Also make sure to remove the limits in your scale_y_continuous, otherwise the labels won't be shown. Here is a reproducible example:

set.seed(13)
df = data.frame(grouping = c(rep("A",24),rep("B",24),rep("C",24)),
                id = c(rep(1,12),rep(2,12),rep(3,12),rep(4,12),rep(5,12),rep(6,12)),
                subject = c(rep(c("One","Two","Three","Four"),72)),
                family = c(rep(c("Red","Red","Red","Red","Blue","Blue","Blue","Blue","Green","Green","Green","Green"),6)),
                percent = sample(1:100,288,replace=T),
                result = "Good")

library(ggplot2)
library(ggpubr)
library(dplyr)
df %>%
  ggplot(aes(x=subject,y=percent, group = interaction(subject,grouping))) + 
  geom_boxplot(outlier.shape = NA) + 
  geom_point(aes(color = grouping, group = grouping),size=1.75,position = position_dodge(width = 0.75))+
  scale_y_continuous(breaks = seq(0, 100, by = 25)) + 
  facet_wrap(~ family) +
  scale_x_discrete("") +
  theme_minimal() +
  stat_compare_means(comparisons = list(c("Four", "One"), c("Four", "Three"), c("Four", "Two"), c("One", "Three"), c("One", "Two"), c("Three", "Two")),
                     aes(label=..p.adj..),
                     method = "t.test",paired=F,
                     tip.length=0,step.increase = 0.05)
#> Warning: Using `as.character()` on a quosure is deprecated as of rlang 0.3.0.
#> Please use `as_label()` or `as_name()` instead.
#> This warning is displayed once per session.

^{Created on 2022-08-28 with reprex v2.0.2}

Grouping by a factor with dodged position in ggplot

I think you need to rethink your approach - dodging doesn't work well with lines. Why not use facetting and explicitly display trial on the x axis?

ggplot(d, aes(trial, y = value)) + 
  geom_line(aes(group = subject)) +
  geom_point(aes(colour = treatment)) +
  geom_text(aes(label = subject), hjust = 1.5, size = 3)  +
  facet_wrap(~ treatment)

.....

---

Related Topics