Assume this code (written in MatLab). I'm about to calculate average of 50 random angles using Circular Statistics (Circular Mean), with nested averaging method:
rm = (rand(5,10) - 0.5) * 12; % generate random numbers in -6 .. 6
av1 = circ_mean(circ_mean(rm)')
av2 = circ_mean(circ_mean(rm')')
av1 and av2 expected to be the same, but! they are not even close to each other. Is that about implementation?