Hello!
I am using the Fisher's z-Test to compare two Pearson-Correlation-Coefficients. I am currently researching the statistical assumptions Fisher's z-Test makes. Most importantly I'd like to ask, wether both of the compared coefficients need to be significant in order to use the test. I have found conflicting positions on that. I'm hoping somebody can provide a citable source.
Thanks in advance and
best of health
David
As far as I can think, I don't think it matters at all if they are both "significant" or not. One approach is testing the hypothesis that r is different from zero, and the other is testing whether two r's are different. I don't see how one hypothesis test has any bearing on the other. I absolutely have no references for this idea, tho. I'm just consulting my brain, which is, eh, a semi-reliable machine.
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation