I am using the function rda from vegan to assess the effect of food availability - FOOD - (3 levels: low, medium, high) and tillage - TILL - (2 levels: yes, no) on earthworm species composition. Due to practical reasons, I had to design the experiment in a split plot: There are in total 4 blocks, and within each block, the whole-plot factor is TILL, while the split-plot factor is FOOD, which is randomized within each level of the whole-plot factor(TILL).
To summarize, FOOD is nested within TILL, and the combination FOODxTILL is replicated 4 times, resulting in 24 observations (3FOOD x 2TILL x 4BLOCKS)
I am carrying out a partial RDA (with Block as Condition) for this. The next logical step would be to test the significance of my treatments using the anova function and restricting the permutations with permute (function how, I think).
However, I am struggling in how to design the restricted permutation scheme to test the significance levels using the package permute.
In particular, I am not managing to understand how to code the "within", the "plot", and the "block", as well as the "strata" and the "type" for each of the hierarchical levels of the restricted permutations in the function "how" of the permute package.
Does anyone know how to code this? I'd appreciate some insights on this!