I am trying to test for vocal differences between 5 individuals with sample sizes 102, 86, 115, 45, 12. My initial plan was to use PERMANOVA, but I am concerned about the sample size limitations and differences between sample sizes of the groups.
The variables in the dataset are all numeric measurements of attributes of the audio signal and there is 24 of them. Does the high number of variables matter for PERMANOVA and if so, would it be correct to use PCA to reduce the number of variables?