Parametric or non-parametric test?

[Maxilo]

Dear all,
I’m not fluent in statistical data analysis and I can’t afford the services of a professional. Please I really need your suggestions with the following:
The first column of my table represents the different sampling sites from which I obtained different parts of a plant species (roots, stems, leaves and reproductive structures). Columns two to six contain the Reproductive Allocation (RA) of Nitrogen, Calcium, Magnesium, Potassium, Sodium and Phosphorus respectively (calculated as percentage of these nutrients in reproductive structures).
I’m having a tough time choosing the right statistical tests that will
1) Detect if there are differences in RA across all sites and among nutrients
2) detect if there are any patterns (relationships/equivalence) in RA.
Is this all the information the table could provide?
Many thanks View attachment 6038

 

[rogojel]

hi,
I would first do a quick cluster analysis, though that is not a atatistical test per se, it could give you some relationships.

regards

 

[gianmarco]

As already suggested, Cluster Analysis could be a way to go. Maybe Principal Component Analysis could be put to work ad well.
If you had frequencies rather than percentages, you could use Correspondence Analysis.

Hope this helps
Gm