A reviewer of our recent paper wants us to have exact p values throughout the paper as opposed to p<001. When I run anosim in mothur, that’s the result I get: p<0.001*. Does anybody know how to get the exact p-value? Many thanks!

# How do I calculate exact p-value when p<0.001

Hi!

If that’s the worst of the critiques, I’d say you’re doing pretty well! I’d gently point out that’s a silly expectation. This is a permutation test and if you want more precision to the p-value you will have to increase the iters option to something more than 1000. Regardless, the actual value of the p-value is not meaningful. All that matters is that it’s less than 0.05.

I’ve actually gotten the opposite critique. I had a reviewer ask us to remove the p-values and instead simply state whether the p-value was significant or not. Can’t win!

Pat

You might also indicate how many times, within all your iterations, ANOSIM gave a result higher than the true R value (if I recall properly, that it how the p-value is calculated in ANOSIM - (number of permutations > R) / (total number of permutations)).

But, if he/she is asking for the exact p-value, I don’t think that reviewer understands p-values. Unless you have done many similar tests, and is trying to do a Bonferroni correction or something like that.

Leo

exactly, pardon my french but what a moron! On one hand (s)he does not understand how pvalues work on the other hand (s)he asks for sth meaningless, so tired of those people, hang in there @Ischaef!

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.