Skip to contents

Plot a group comparison

Source: R/plot_compare.R
plot_compare.Rd

Create a publication-ready comparison plot using a sensible visual selected from the outcome type and study design.

Usage

plot_compare(
  data,
  variable,
  group,
  paired = FALSE,
  id = NULL,
  type = c("auto", "box", "bar"),
  display = c("proportion", "count"),
  show_points = TRUE,
  show_p = FALSE,
  test = c("auto", "t_test", "welch_t", "wilcox", "anova", "welch_anova", "kruskal",
    "chisq", "fisher", "mcnemar", "rm_anova", "friedman", "cochran_q"),
  var_equal = FALSE,
  palette = NULL,
  base_size = 14,
  title = NULL,
  caption = NULL,
  xlab = NULL,
  ylab = NULL,
  legend_title = NULL
)

Arguments

data

A data frame.

variable

Outcome or response variable, supplied as a bare name or character string.

group

Categorical grouping variable, supplied as a bare name or character string.

paired

Logical; whether observations are paired or repeatedly measured.

id

Participant identifier required when paired = TRUE.

type

Plot type: "auto", "box", or "bar".

display

Categorical display: within-group "proportion" or "count".

show_points

Logical; show individual observations for continuous data.

show_p

Logical; add the selected test and p-value as a plot caption.

test

Test passed to compare_groups() when show_p = TRUE.

var_equal

Logical; passed to compare_groups() when show_p = TRUE. With test = "auto", TRUE selects the equal-variance parametric route for independent, non-skewed continuous outcomes. Default FALSE retains Welch methods; no variance test is performed.

palette

Optional character vector of colours. It must contain at least one colour per displayed group or outcome level.

base_size

Base font size.

title, caption

Optional plot title and caption.

xlab, ylab

Optional axis labels.

legend_title

Optional legend title.

Value

A ggplot object.

Details

The minimal call is plot_compare(data, variable, group). Continuous outcomes are shown as boxplots with individual observations, categorical outcomes as stacked within-group proportions, and paired continuous outcomes as participant-level connected observations. The returned object is a standard ggplot and can be customized with ordinary ggplot2 layers.

When show_p = TRUE, the annotation is obtained from compare_groups() with the same test, var_equal, paired, and id settings. The test name is always shown with the p-value. The annotation and plotted denominators use the same complete observations; non-finite continuous values are excluded.

Examples

plot_compare(mtcars, variable = mpg, group = am)


plot_compare(
  mtcars,
  variable = vs,
  group = am,
  display = "proportion",
  show_p = TRUE
)