In an experiment involving plant clones there are: 20+ treatments, 5 technical replicates for each treatment (plant clones in which the treatments were tested). Variable of interest is Biomass.
Due to the small number of technical replicates, I believe a non-parametric test (like Kruskal-Wallis) should be used ... am I right in this assumption? Also ... I believe that as I am working with plant clones I should use a test for dependent samples (then Kruskal-Wallis does not seem to be that suited for this). Which test is more suitable to this situation?
Extra question: If I conduct the exact same analysis but in a different medium of growth than the previous one, I believe I would be adding another variable to the experiment, therefore should I use some sort of analogous 2-way ANOVA but suited for these conditions?