Comparing two correlation coefficients

#1
I am trying to compare 2 independent correlation coefficients. Can I compare 2 independent rs using fisher z transformation test which one is a Pearson’s r and the other is Spearman’s r or do they have to be the same?
 

jocos

New Member
#3
Hi

I performed a Pearson's Correlation on two normally distributed variables and Spearman's on two other non-normal variables. All I want to know please, is right to compare the two r values e.g. 0.3 (using Pearson) and 0.25 (using Spearman's) using the fisher z transformation test?

Thank you
 

hlsmith

Omega Contributor
#4
I am guessing you could do that, but I FEEL you lose information in the transformation. I have no theory to support that though. Can you transform the latter variables to make them more normalized, then run them with Pearson?