Dear all,
I am facing negative estimates of components of variation (ECV) and high p-values for some terms in a PERMANOVA analysis (working with Primer-e v6 here). According to PERMANOVA+ manual, page 54 (attached), this is an issue in nested designs that might be solved by sequentially pooling these terms when re-doing the analysis. However, this "pooling" option that appears in the manual (Fig. 1.32 in page 58) is not available for me when I try to make the analysis and I can only see the "exclude terms" option (see attached picture). This is unfortunate because, according to the manual, the correct way to go would be pooling rather than excluding. Did any of you have the same issue? do you know if I can do this in Primer-e v.7 or using the Vegan package in R?
Thank you very much in advance for your help.
Best,
Juan
Dear Juan,
I have not used PRIMER but I've been this week precisely working with PERMANOVA in R with vegan (with adonis2). I exactly had the same problem at first (P values of 1), until I realized the terms should be nested!
For example, my model has 4 fixed categorical factors, and three of them are nested into the first one (the symbol "/" is used in R for that). I want to test the single effects + nested effects:
permanova F)
F1 ............................. 0.001 ***
F2 ............................ 0.001 ***
F3 ............................0.001 ***
F4 ............................ 0.001 ***
F1(F2) ............................0.104
F1(F2(F3)) ............................ 0.001 ***
F1(F2(F3(F4)))) ............................ 0.012 *
Residual ............................
Total ............................
I hope this can help you out, don't hesitate contacting me if you need some help or codes for it :)
Cheers,
Raquel
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts