How to move significant level on plot to the top?

Question

I have this barplot using ggpubr with the code

visualiseplot <- visualisepwc %>% add_xy_position(x = "Condition")
visualiseplot3 <- ggbarplot(data_sum, x = "Condition", y = "Q70.7", fill = "Condition") +
  stat_pvalue_manual(visualiseplot, hide.ns = TRUE)
visualiseplot3

My graph I got from it looks like this:

I got a message saying I can only have a link to the picture so I hope you can see the image, but if you can't basically the significance levels are at the bottom of the graph and overlap, and I would like it to be above the graph at the top.

this is an example of how it should look but using a boxplot - it seems to work with boxplot but not barplot :

Any recommendations for what I should do? -

Answer 1

Without your data and the visualisepwc output, it is difficult to replicate the issue. However, you could avoid the issue altogether by using the ggsignif package. I've tried to replicate the aesthetics of the ggpubr plot as much as possible.

library(ggplot2)
library(ggsignif)

## Fictional Data (your data_sum)
data_sum <- data.frame("Condition"=c("Fictional","Socratic","Textbook"),"Q70.7"=c(52,45,20))

ggplot(data_sum,aes(x=Condition,y=Q70.7,fill=Condition))+ # map data
  geom_bar(stat="identity",color="black")+ # plot bars
  geom_signif(comparisons = 
    list(c("Fictional","Textbook"),c("Socratic","Textbook")), # indicate which pairs to connect with significance bars
          map_signif_level = TRUE,
          y_position = c(60,50), # indicate where to put the bar for each pairing
          tip_length = 0.1, # set vertical size of bars
          color="black",
          annotations = c("**","**"))+ # indicate signif level for each pairing
  theme_classic()+ 
  theme(legend.position = "top")