Significance annotation layer

Adds statistical significance brackets and labels between groups. This is a syntax-sugar composite mark that combines mark_rule and mark_text internally.

Usage

mark_significance(
  plot,
  comparisons,
  y_position = NULL,
  y_offset = NULL,
  line_color = ._MARK_STYLE$ink,
  line_width = ._MARK_STYLE$lw_thin,
  text_size = ._MARK_STYLE$txt_note,
  tip_length = 0.02,
  ...
)

Arguments

plot

A plotit object

comparisons

A data frame with columns: group1, group2, label, and optionally y_position. Character columns are matched against the x-axis variable.

y_position

Numeric vector of y-positions for the brackets. If omitted, auto-computed from data range.

y_offset

Text offset above the bracket line (default 0.5). In data units.

line_color

Colour for the bracket lines (default ._MARK_STYLE$ink = "grey30").

line_width

Width of bracket lines (default 0.5).

text_size

Size of significance label text (default 3.2).

tip_length

Length of bracket end-tick lines (default 0.02). Units follow the axis type: a fraction of the level count on discrete x axes, a fraction of the bracket's numeric span on continuous axes.

...

Additional arguments passed to the label annotation (ggplot2::annotate("text", ...)).

Value

Modified plotit object

Details

Equivalent expansion:

  p |> mark_rule(x = comp$group1, xend = comp$group2,
                 y = comp$y_position, yend = comp$y_position) |>
       mark_text(x = midpoint, y = comp$y_position + y_offset,
                 label = comp$label)

Examples

df <- data.frame(group = c("A", "B", "C"), value = c(5, 8, 4))
comp <- data.frame(
  group1 = c("A", "A"), group2 = c("B", "C"),
  label = c("**", "ns")
)
plotit(df, encode(x = group, y = value)) |>
  mark_bar() |>
  mark_significance(comp, y_position = c(9, 6))