reverse dodge order works differently in position_jitterdodge() vs., for example position_dodge2()

[davebraze]

Reversing the order of dodged elements using position_dodge2(), one passes in reverse=TRUE. To reverse dodged element positions in position_jitterdodge(), it is necessary to use a negative value to dodge.width=-.75; there is no 'reverse' argument. This is undocumented in position_jitterdodge and inconsistent with other dodge functions.

[clauswilke]

Please provide a reprex.

[smouksassi]

davebraze is asking why there is no reverse argument for jitterdodge or at least document negative width behavior

library(ggplot2)
ggplot(mtcars,aes(as.factor(cyl),mpg,fill=as.factor(vs),color=as.factor(vs)  ))+
  geom_boxplot(position = position_dodge2(reverse = TRUE),color="black")+
  geom_point(position =position_jitterdodge(dodge.width = -0.75))
#> Warning: position_jitterdodge requires non-overlapping x intervals

^{Created on 2019-11-07 by the reprex package (v0.3.0)}

[joelgautschi]

the same is true for position_dodge() with geom_point(): position_dodge(width = -0.5) gives a warning. However, the order is reversed as I had hoped.

library(tidyverse)

df <- tribble(~dimension, ~group, ~value,
              "dim 1", "A", 0.6,
              "dim 1", "B", 0.75,
              "dim 1", "C", 0.25,
              "dim 2", "A", 0.4,
              "dim 2", "B", 0.4,
              "dim 2", "C", 0.5)

plot <- ggplot(data = df,
               mapping = aes(x = dimension,
                             y = value,
                             color = group)) 

# warning
plot +  
  geom_point(position = position_dodge(width = -0.5))
#> Warning: position_dodge requires non-overlapping x intervals

^{Created on 2021-01-26 by the reprex package (v0.3.0)}

The question remains whether this is 'good practice' or if it is better to use an approach like color = fct_rev(group) within aes() to reverse the order.