Problems using alternative tests (i.e. anova) about ggsignif HOT 2 CLOSED

Hi, I am happy to hear that you like the package.

the argument that is given to test must satisfy certain requirements, which anova does not so that is the reason, why the significance layer is not appearing.

But you can easily wrap the anova call in your own function, as I do in this example:

a <- rnorm(n = 10, mean = 5, sd = 1)
b <- rnorm(n = 10, mean = 5.8, sd = 1)


library(tidyverse)
#> Loading tidyverse: ggplot2
#> Loading tidyverse: tibble
#> Loading tidyverse: tidyr
#> Loading tidyverse: readr
#> Loading tidyverse: purrr
#> Loading tidyverse: dplyr
#> Conflicts with tidy packages ----------------------------------------------
#> filter(): dplyr, stats
#> lag():    dplyr, stats
library(ggsignif)

data <- data.frame(label = c(rep("A", 10), rep("B", 10)), id = c(1:10, 1:10), 
  value = c(a, b))

ggplot(data, aes(x = label, y = value)) +
  geom_boxplot() + 
  geom_signif(comparison = list(c("A",  "B")), y_position = 7.2, test = "wilcox.test") + 
  geom_signif(comparison = list(c("A",  "B")), y_position = 7.6, test = "t.test") +
  geom_signif(comparison = list(c("A",  "B")), y_position = 8, test = function(a, b) {
    list(p.value = summary(aov(a ~ b))[[1]][["Pr(>F)"]][[1]])
  })

I actually call aov instead of anova, but this should not be that important. Please note that the package is only designed to support 2 way comparisons (unlike the original intend of an anova), so you might want to compute the individual p-values separately and then use the annotations parameter to draw them.

from ggsignif.

Thanks a lot const-ae for your immediate reply ! Highly appreciable !
I have few doubts :

1. " [[1]][["Pr(>F)"]][[1]] " - could you please explain what this indicates ? Because the p values are different when I run anova separately for each and every values in the data. With geom_significance it is showing as "NS" whereas if i run anova separately for the same set of values, it shows "***" Significant difference.
2. In the place of anova, can I use tukey or duncan test for multiple comparisons?
3. I would like to have "", "", "", "NS" in different colors. since I didn't specify annotations separately, I couldn't give different colors. could you please help me ?

geom_signif(comparisons = list(c("High", "Low")), y_position=c(4.5, 7.7),test = function(a, b) { list(p.value = summary(aov(a ~ b))[[1]][["Pr(>F)"]][[1]])}, map_signif_level=T, tip_length=0)
Once again thanks for your help and for your valuable time to help me out !

from ggsignif.