Fig. 4
From: The multivariate analysis of variance as a powerful approach for circular data
Continuation of our hypothetical example, analyzing the usefulness of the MANOVA approach for biological data. In addition to the yearly increase of clustering, we now manipulated the orientation direction for each group, using either a 90° (a, b) or a 180° (c, d) difference. This was then either combined with a group difference in yearly clustering increase (b, d), or not (a, c). The left panel shows the power for the intercept of MANOVA approaches as well as Rayleigh and HR test, the center panel shows the power of detecting the linear effect, the right panel shows power for detecting group differences