The 2 functions provided perform a nonparametric permutation test based on Watson's U2 statistic for 2-sample testing of circular data (angles).
Watsons_U2 computes Watson's U2 statistic for 2 samples of circular data. The code is derived from the equations of Zar JH, Biostatistical Analysis, 4th ed. (1999) and has been verified against the numerical examples provided therein.
Watsons_U2_perm_test computes the probability that, according to the null hypothesis, 2 samples of circular data come from the same population, or from 2 populations that have the same direction. The p values thus obtained have also been verified against the numerical examples provided in Zar (1999).
As an alternative to the permutation test, significance tables for critical values of Watson's U2 have been published (e.g. in Zar, 1999).
Please do not hesitate to contact me for any suggestion or to report a problem about this function.
Moved source code to GitHub.
The functions that compute Watson's U2 statistic and the nonparametric permutation test using that statistic have been merged.
Inspired by: Circular Statistics Toolbox (Directional Statistics)