I am doing a fourth-corner (co-inertia on 3 matrices to evaluate relationships between two matrices) and the analysis is known to increase the occurrence of false positives (multiple comparison). It is advised to adjust the p-values with the False Discovery Rate method.

However, I wonder how to evaluate whether to use non-adjusted vs. adjusted p-values.

A colleague and the literature mentioned that there is no strong basis whether or not to adjust p-values but that it mainly depends on the results.

What are the results supposed to look like so that you know if you should adjust the p-values or not?

I would like more opinions on this. :)

I plan to use the p-values further in my analysis as an indicator for relationships between my variables and will change the range so that it represents strong and weak relationships (X - pvalue).

As non-adjusted pvalues range between 0-1, I would do 1- non-adjusted pvalues.

However, my colleague mentioned (if I understood) that for adjusted pvalues, the range should go from 0.5-1 from the significant p-values. I am not sure why though and not sure I grasp how my non-adjusted pvalues relate to the adjusted counterparts.

Any lights on this?

Thanks in advance,