Relabelling predictors with jtools effect

I'm trying to edit the names of the factor levels of a predictor variable in R for an effect plot with the jtools package. I should be able to do this with the "pred.labels" argument, but the labels don't change. For example, using the mtcars data set I try to write out the predictor names ("four", "six", "eight") but there is no change in the figure.

data(mtcars)
mtcars$cyl <- factor(mtcars$cyl)
mtest <- lm(mpg ~ cyl   hp   wt, data = mtcars)
effect_plot(mtest, pred = cyl, pred.labels = c("four", "six", "eight"))

Figure produced with above script, where labels have NOT been changed

Now, if this were really my problem I would just overwrite the factor levels in the data. However, in my data set I have many factors that should have a space in their name for easy interpretation, so I don't want to recode my variables this way. Any idea how to fix this?

Thanks

CodePudding user response：

Since jtools seems to use ggplot2 you can just change the labels like you would if you were using ggplot2

label <- c("four", "six", "eight")
effect_plot(mtest, pred = cyl)   scale_x_discrete(labels= label)

effect_plot(mtest, pred = cyl)   scale_x_discrete(labels= c("four", "six", "eight"))

CodePudding user response：

Not sure how to rename the group labels in jtools but you can do it with scale_x_discrete in ggplot2 since the resulting object is from ggplot.

effect_plot(mtest, pred = cyl)  
   scale_x_discrete(labels = c("four", "six", "eight"))