Draft vignette with package comparison

inst/WORDLIST

@@ -1,3 +1,4 @@
+’
 ’s
 ANCOVAs
 Arel

pkgdown/_pkgdown.yml

@@ -96,6 +96,10 @@ navbar:
           href: articles/introduction_comparisons_4.html
         - text: "Contrasts and comparisons for zero-inflation models"
           href: articles/introduction_comparisons_5.html
+        - text: -------
+        - text: "Comparison to other R packages"
+        - text: "Packages related to predictions, marginal means and effects"
+          href: articles/practical_comparison.html
     - icon: fa fa-newspaper
       text: News
       href: news/index.html

vignettes/overview_of_vignettes.Rmd

@@ -69,3 +69,7 @@ All package vignettes are available at [https://easystats.github.io/modelbased/]
 * [Slopes, floodlight and spotlight analysis (Johnson-Neyman intervals)](https://easystats.github.io/modelbased/articles/introduction_comparisons_3.html)
 * [Contrasts and comparisons for generalized linear models](https://easystats.github.io/modelbased/articles/introduction_comparisons_4.html)
 * [Contrasts and comparisons for zero-inflation models](https://easystats.github.io/modelbased/articles/introduction_comparisons_5.html)
+
+### Comparison to other R packages
+
+* [Comparison of R packages related to predictions, marginal means and effects](https://easystats.github.io/modelbased/articles/practical_comparison.html)

vignettes/practical_comparison.Rmd

@@ -0,0 +1,143 @@
+---
+title: "Case Study: Comparison of R packages related to predictions, marginal means and effects"
+output: rmarkdown::html_vignette
+vignette: >
+  %\VignetteIndexEntry{Case Study: Comparison of R packages related to predictions, marginal means and effects}
+  %\VignetteEngine{knitr::rmarkdown}
+  %\VignetteEncoding{UTF-8}
+bibliography: bibliography.bib
+---
+
+```{r set-options, echo = FALSE}
+knitr::opts_chunk$set(
+  collapse = TRUE,
+  comment = "#>",
+  dev = "png",
+  out.width = "100%",
+  dpi = 300,
+  message = FALSE,
+  warning = FALSE,
+  package.startup.message = FALSE
+)
+
+pkgs <- c("emmeans", "marginaleffects", "ggeffects")
+if (!all(insight::check_if_installed(pkgs, quietly = TRUE))) {
+  knitr::opts_chunk$set(eval = FALSE)
+}
+if (getRversion() < "4.1.0") {
+  knitr::opts_chunk$set(eval = FALSE)
+}
+```
+
+This vignette compares the **modelbased** package to other common packages that can be used to compute adjusted predictions, marginal means, marginal effects, or contrasts and pairwise comparisons.
+
+**modelbased** is built on top of the two probably most popular R packages for extracting marginal means and effects, namely **emmeans** [@russell2024emmeans] and **marginaleffects** [@arel2024interpret]. Thus, you obtain the same results either from **modelbased** or one of the other two packages. This vignette shows how to replicate results using the different packages.
+
+```{r}
+# to create data grids, we use `insight::get_datagrid()`
+library(insight)
+
+# the four packages, which we compare
+library(modelbased)
+library(emmeans)
+library(marginaleffects)
+library(ggeffects)
+```
+
+## Categorical predictors
+
+```{r}
+# a very simple model
+data(iris)
+model <- lm(Petal.Length ~ Species, data = iris)
+
+# modelbased
+estimate_means(model, by = "Species")
+
+# emmeans
+emmeans(model, "Species")
+
+# marginaleffects
+avg_predictions(model, by = "Species")
+
+# ggeffects
+predict_response(model, "Species")
+```
+
+## Continuous predictors
+
+```{r}
+# a very simple model
+data(iris)
+model <- lm(Petal.Length ~ Sepal.Length, data = iris)
+
+# create a range of representative values
+grid <- get_datagrid(model, by = "Sepal.Length")
+grid
+
+# modelbased - defaults to create a range of 10 values from
+# minimum to maximum for numeric focal predictors
+estimate_means(model, by = "Sepal.Length")
+
+# emmeans
+emmeans(
+  model,
+  "Sepal.Length",
+  at = list(Sepal.Length = grid$Sepal.Length)
+)
+
+# marginaleffects
+avg_predictions(
+  model,
+  by = "Sepal.Length",
+  newdata = data.frame(Sepal.Length = grid$Sepal.Length)
+)
+
+# ggeffects
+predict_response(model, "Sepal.Length [4.3:7.9 by=0.4]")
+```
+
+## Interaction between continuous and categorical
+
+```{r}
+# a very simple model
+data(iris)
+model <- lm(Petal.Length ~ Sepal.Length * Species, data = iris)
+
+# create a range of representative values
+grid <- get_datagrid(
+  model,
+  by = c("Species", "Sepal.Length"),
+  range = "grid",
+  preserve_range = FALSE
+)
+grid
+
+# modelbased
+estimate_means(model, by = c("Species", "Sepal.Length"), range = "grid")
+
+# alternative notation - for "Sepal.Length", we want mean and +/- SD
+estimate_means(model, by = c("Species", "Sepal.Length = [meansd]"))
+
+# we could also pass a data grid to the `newdata` argument...
+estimate_means(model, by = c("Species", "Sepal.Length"), newdata = grid)
+
+# emmeans
+emmeans(
+  model,
+  c("Species", "Sepal.Length"),
+  at = lapply(grid, unique)
+)
+
+# marginaleffects
+avg_predictions(
+  model,
+  by = c("Species", "Sepal.Length"),
+  newdata = grid
+)
+
+# ggeffects
+predict_response(model, c("Species", "Sepal.Length"))
+```
+
+# References