Skip to contents

Test whether groups share the same mean direction (Watson-Williams)

Source: R/circular_statistics.R
test_mean_directions.Rd

Wraps watson.williams.test – the circular analogue of the parametric F-test for equal means. Assumes von Mises- distributed data with equal concentrations across groups; if concentrations differ substantially or the distribution is non-von-Mises, consider a non-parametric alternative.

Usage

test_mean_directions(
  hd,
  group_col,
  angle_col = "heading",
  pairwise = FALSE,
  p_adjust = "none"
)

Arguments

hd

Data frame with heading and group columns.

group_col

Column identifying conditions or groups.

angle_col

Heading column in radians. Default "heading".

pairwise

Logical. FALSE (default) returns a single omnibus test across all groups. TRUE returns all pairwise comparisons.

p_adjust

Multiple-comparison correction method passed to p.adjust. Default "none". Applies only to the pairwise output; a p_value_adj column is added. Strongly recommended when pairwise = TRUE: use "BH" (Benjamini-Hochberg) or "holm" (family-wise control). Ignored for the omnibus test (single p-value, no adjustment needed).

Value

Tidy data frame. Omnibus result has columns n_groups, statistic, df1, df2, p_value, test. Pairwise result additionally has group1, group2, and p_value_adj (when p_adjust != "none").