Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [5] is duplicated

[matasV99]

Hi,

I am trying to run your Cummings estimation plot for multiple group comparison but i am running into the error in the name of this issue.

Here is a reproducible example:
df <- data.frame (first_column = c(0.5, 0.55, 0.65, 0.64, 0.6125, 0.6, 0.525, 0.6125, 0.4375, 0.475, 0.525, 0.425, 0.6, 0.54, 0.64, 0.58),
second_column = c("O pre", "O pre", "O pre", "O pre", "Sham pre", "Sham pre", "Sham pre", "Sham pre", "Sham after", "Sham after", "Sham after", "Sham after", "O after", "O after", "O after", "O after" )
)
colnames(df)[1] <- "light"
colnames(df)[2] <- "condition"

multi.two.group.unpaired <- df %>% dabest(condition, light,
idx = list(c("O pre", "O after"),
c("Sham pre", "Sham after"),
c("Sham after", "O after")),
paired = FALSE)
multi.two.group.unpaired.meandiff <- mean_diff(multi.two.group.unpaired)
plot(multi.two.group.unpaired.meandiff)

Thank you for your time,
Matas

[josesho]

Hi Matas,

Unfortunately the plotting function is unable to plot repeated groups directly .

You will want to duplicate the repeated columns, so something like

df <- data.frame(first_column = c(0.5, 0.55, 0.65, 0.64, 
                                    0.6125, 0.6, 0.525, 0.6125, 
                                    0.4375, 0.475, 0.525, 0.425, 
                                    0.6, 0.54, 0.64, 0.58,
                                    0.4375, 0.475, 0.525, 0.425, 
                                    0.6, 0.54, 0.64, 0.58 ),

                 second_column = c("O pre", "O pre", "O pre", "O pre", 
                                   "Sham pre", "Sham pre", "Sham pre", "Sham pre", 
                                   "Sham after", "Sham after", "Sham after", "Sham after", 
                                   "O after", "O after", "O after", "O after" ,
                                   "Sham after2",  "Sham after2", "Sham after2", "Sham after2", 
                                   "O after2", "O after2", "O after2", "O after2" )
)        

colnames(df)[1] <- "light"
colnames(df)[2] <- "condition"

followed by

multi.two.group.unpaired <- df %>% dabest(condition, light,
                                                  idx = list(c("O pre", "O after"),
                                                             c("Sham pre", "Sham after"),
                                                             c("Sham after2", "O after2")),
                                                  paired = FALSE)
multi.two.group.unpaired.meandiff <- mean_diff(multi.two.group.unpaired)
plot(multi.two.group.unpaired.meandiff)

You can use a vector graphics software to edit the x-axis labels.
Hope this helps!

[matasV99]

Dear Josesho,

Thank you so much! May I inquire what exactly you mean by vector graphics software? I understand dabest uses ggplot, however, I was unable to change the axis labels or labels by just adding ggplot objects to dabest objects...

Thank you again for your answer this is very helpful,
Matas

[josesho]

Vector graphics software are software that can directly edit SVG output, such as Adobe Illustrator, and Inkscape. So you'd produce the estimation graph to SVG, then open it in a vector graphics software, and directly edit the text. Not ideal, but solves the problem!