Confidence interval for a circular mean

Computes an approximate confidence interval for a circular mean. The large sample method uses a normal approximation on the mean direction, while the bootstrap method resamples angles and forms an interval from angular deviations around the sample mean. These intervals are exploratory; they are not reliable when the mean resultant length is close to zero and the mean direction is weakly identified.

Usage

circular_mean_ci(
  x,
  level = 0.95,
  method = c("large_sample", "bootstrap"),
  R = 999,
  axial = FALSE,
  na.rm = TRUE,
  seed = NULL
)

Arguments

x

Numeric vector of angles in radians.

level

Confidence level.

method

Interval method.

R

Number of bootstrap resamples.

axial

Should data be treated as axial, modulo pi?

na.rm

Should missing values be removed?

seed

Optional random seed for the bootstrap.

Value

A tibble with mean, lower, upper, level, method, n and Rbar.

See also

Other circular summaries: circular_sd(), circular_summary(), circular_variance(), estimate_kappa(), mean_direction(), mean_resultant_length(), resultant_length()