This document describes how to plot marginal effects of interaction
terms from various regression models, using the
plot_model()
function. plot_model()
is a
generic plot-function, which accepts many model-objects, like
lm
, glm
, lme
,
lmerMod
etc.
plot_model()
allows to create various plot tyes, which
can be defined via the type
-argument. The default is
type = "fe"
, which means that fixed effects (model
coefficients) are plotted. To plot marginal effects of interaction
terms, call plot_model()
with:
type = "pred"
to plot predicted values (marginal
effects) for specific model terms, including interaction terms.type = "eff"
, which is similar to
type = "pred"
, however, discrete predictors are held
constant at their proportions (not reference level). It internally calls
via .type = "emm"
, which is similar to
type = "eff"
. It internally calls via .type = "int"
to plot marginal effects of interaction
terms in a more convenient way.plot_model()
supports labelled data
and automatically uses variable and value labels to annotate the plot.
This works with most regression modelling functions.
Note: For marginal effects plots, sjPlot calls functions from the ggeffects-package. If you need more flexibility when creating marginal effects plots, consider directly using the ggeffects-package.
Note: To better understand the principle of plotting interaction terms, it might be helpful to read the vignette on marginal effects first.
To plot marginal effects of interaction terms, at least two model
terms need to be specified (the terms that define the interaction) in
the terms
-argument, for which the effects are computed. To
plot marginal effects for three-way-interactions, all three terms need
to be specified in terms
.
A convenient way to automatically plot interactions is
type = "int"
, which scans the model formula for interaction
terms and then uses these as terms
-argument.
library(sjPlot)
library(sjmisc)
library(ggplot2)
data(efc)
theme_set(theme_sjplot())
# make categorical
efc$c161sex <- to_factor(efc$c161sex)
# fit model with interaction
fit <- lm(neg_c_7 ~ c12hour + barthtot * c161sex, data = efc)
plot_model(fit, type = "pred", terms = c("barthtot", "c161sex"))
For type = "int"
, no terms need to be specified. Note
that this plot type automatically uses the first interaction term in the
formula for the x-axis, while the second term is used as grouping
factor. Furthermore, if continuous variables are used as second term,
you can specify preset-values for this term with the
mdrt.values
-argument, which are then used as grouping
levels.
In this example, the second term is a factor with two levels (male/female), so there is no need for choosing specific values for the moderator.
To switch the terms, in this example barthtot and
c161sex, simply switch the order of these terms on the
terms
-argument and use type = "pred"
.
To switch the terms for plot-type type = "int"
, you need
to re-fit the model and change the formula accordingly, i.e. using
c161sex as first term in the interaction.
# fit model with interaction, switching terms in formula
fit <- lm(neg_c_7 ~ c12hour + c161sex * barthtot, data = efc)
plot_model(fit, type = "int")
By default, for continuous variables, the minimum and maximum values
are chosen as grouping levels, which are 0 and 100 - that’s why the
previous two plots are identical. You have other options as well,
e.g. the mean-value and +/- 1 standard deviation (as suggested by Cohen
and Cohen for continuous variables and popularized by Aiken and West
1991), which can be specified using mdrt.values
.
Since the terms
-argument accepts up to three model
terms, you can also compute marginal effects for a
3-way-interaction.
# fit model with 3-way-interaction
fit <- lm(neg_c_7 ~ c12hour * barthtot * c161sex, data = efc)
# select only levels 30, 50 and 70 from continuous variable Barthel-Index
plot_model(fit, type = "pred", terms = c("c12hour", "barthtot [30,50,70]", "c161sex"))
Again, type = "int"
will automatically plot the
interaction terms, however, using mdrt.values = "minmax"
as
default - in this case, the “levels” 0 and 100 from continuous variable
barthtot are chosen by default.
#>
#> [[2]]
#>
#> [[3]]
#>
#> [[4]]
Aiken and West (1991). Multiple Regression: Testing and Interpreting Interactions.