Félix -

I do not fully understand your question, though I think it a good idea in general to note that a statistical 'test' is not so conclusive as one might be lead to believe, and in particular, an isolated p-value is rather meaningless.  

Perhaps, however, the second comment found through the link below may be of some interest to you, though as I said, I'm not really certain that this is the kind of thing which you are studying. 

 https://www.researchgate.net/publication/271072864_Analyzing_n_Samples_of_2_Observations_Each

The link found in that second comment may not apply to your kind of data.  I'm not sure.  And computer programming would certainly have to be updated and adjusted to the sample sizes that you actually have.

Cheers - Jim

Thanks James,

I will look into it once I manage to download the full pdf.

I fully agree that the p-value grail is something reviewers too often abuse of.

best

felix