How many replicates do you have for each level of your factor you are comparing?

When you get R = 1 and P values > 0.05 in pairwise comparison, it is often because there is not enough replicates to run a large number of permutations. The P value is extremely sensitive to the number of permutations you run. In your case I would guess you have probably 10-30 permutations possible. Am I right?

Have a look on the number of possible permutations and the number of permutations run for each pairwise test (given in the results sheet in Primer), and let us know how many you have and how many samples replicates you have.

If you have a limited number of permutations possible, there is nothing you can do with your pairwise test and I would not interpret these results but only keep the general ANOSIM result. A way to overcome this, would be to perform a cluster analysis coupled with SIMPROF to determine if each cluster of your dendrogram is unique or not from its surrounding cluster.

You will find more information on this in the following answers in RG:

https://www.researchgate.net/post/Can_anyone_help_me_in_understanding_and_clearly_interpreting_ANOSIM_Analysis_of_Similarityand_SIMPER_Similarity_percentage_analysisresults

Aimeric

Dear Aimeric,

Thanks for your reply.  It was very helpful. Unfortunately as you guessed, few of my samples did'n have many replicates.  |I have only two replicates for each sample.  The permutation numbers were low for pairwise test! I had only around 14 during ANOSIM analysis.

 i will try to do SIMPROF with cluster analysis as you advised.

Once again thanks a lot for your input.

regards

Venkat

A couple of additional notes: R is a scaled measure of the separation between groups of samples, so if the Global test is significant (there are differences between groups) you can interpret the relative values (close to 0, not different, close to 1, very different) even if you don't have sufficient power (samples) to run the pairwise tests.  Simprof tests a different hypothesis, so be careful that you understand what each test is actually testing.

The package is called Primer, not Primer-e (which is the company that sells it).

Hi Paul,

We met at Coffs Harbour. I'm writing up at the moment and have an ANOSIM result showing a global R value of 0.044 and a p value of 0.19. Can I still use this? and any tips on how to present/word the finding? 

Kind Regards

Adam

Hi Adam,

As the value is close to zero (and p is >>0.05) you can use it, but only to say there is no evidence of a difference

Thanks Paul,

Much appreciated 

Thanks for the information Paul, 

Regards

Venkat

Dr. Paul Somerfield , can i use "p" as "Rho" value"

No. They are different things. p is a probability, rho is a correlation. Ubnless I misunderstand your question. If you are asking if you can use the letter p to indicate rho, no you can't, rho is a greek symbol and looks a little different.

Hi there, as stressed before, the P value is strongly influenced by the number of replicates thus permutations in your desing. I believe that for ANOSIM and PERMANOVA for setting up and alpha of 0.05 you need at least five replicates in each of your treatments. If you have, lets say three replicates, if there is strong dissimilarity among treatments R will be close to 1, but P will indicate no significance.

The recomendation in the PRIMER manual is that the interpretation of your results must be based on R values rathern than P. R is a true measurement of the dissimilarity among samples.

Cheers

Also, an alternative would running PERMANOVA with Monte Carlo simulations when there is a low number of replicates. Cheers

Dear Siddarthan,

I hope that this manuscript is helpful, you can download it from the following URL, it is free.

Article Ecological importance and sociological characters of fourtee...

The Author