Plot_model() How to change legend title when color is set to black and white

Question

In plot_model() function (sjPlot package), setting colors to black and white, colors="bw", does not allow the user to modify the character string in the title of the legend in legend.title. Is there a workaround for this without modifying the data itself?

EDIT

Example with colors="bw"

plot_model(mod4, type="int",
           colors="bw",
           mdrt.values="minmax", 
           title = "",
           axis.title = c("Variable 1", "Response"),
           legend.title="Distance (Meters)"
           ) 

With colors="Set1"

plot_model(mod4, type="int",
           colors="Set1",
           mdrt.values="minmax", 
           title = "",
           axis.title = c("Variable 1", "Response"),
           legend.title="Distance (Meters)"
           ) 

Answer 1

The 'internals' of sjPlot are based on ggplot2, so you can use ggplot2 commands to customise your plot.

Example plot (from https://strengejacke.github.io/sjPlot/articles/plot_marginal_effects.html):

library(sjPlot)
library(ggplot2)
data(efc)
theme_set(theme_sjplot())

fit <- lm(barthtot ~ c12hour + neg_c_7 + c161sex + c172code, data = efc)

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"))

With colors = "bw":

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"), colors = "bw")

Changing the legend title:

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"), colors = "bw") +
  scale_linetype_manual(name = "New Title",
                        values = c(1,2,3))

Changing the color and size of the title:

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"), colors = "bw") +
  scale_linetype_manual(name = "New Title",
                        values = c(1,2,3)) +
  theme(legend.title = element_text(colour = "red"))

^{Created on 2023-10-19 with reprex v2.0.2}

---

Edit

Also, it seems that changing legend.title does work in my example:

library(sjPlot)
library(ggplot2)
data(efc)
theme_set(theme_sjplot())

fit <- lm(barthtot ~ c12hour + neg_c_7 + c161sex + c172code, data = efc)

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"))

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"), colors = "bw")

plot_model(fit, type = "pred", terms = c("c12hour", "c172code"),
           colors = "bw", legend.title = "new title")

^{Created on 2023-10-19 with reprex v2.0.2}

Are you able to provide a minimal reproducible example to illustrate your problem? Are you running the most up-to-date version of R/sjPlot?