I am unsure as to which test to use in R. Here is the oversimplified sampling procedure, with easy-looking numbers :
- We have 20 patches of the same size in a field.
- Inside these patches, we look for 50 different species (10 species of grass, and 40 species of insect).
- Every time we find a species of grass, we record its coverage on a rough logarithmic scale from 1 to 4.
- Every time we find a species of insect, we count them and record their abundance on a rough logarithmic scale from 1 to 4.
So my data sort of looks like this:
My problem is, how do I test which species are significantly associated? How do I detect clusters? Multivariate analysis? Half weight index? Bootstrap?
I'm not exactly gifted when it comes to statistics, so any help would be greatly appreciated!