3

I have different variables (A, B, C, D) which are all ordinal scaled. I have another variable (Z) which is ordinal scaled, too. My sample size is approx. 1.500.

Is it in the statistical sense correct, to calculate the Spearman's $\rho$ for each relationship - (1) A & Z, (2) B & Z, (3) C & Z and so on - and compare these values subsequently?

Maybe using the Fisher Z-Test?

Possible result: None of the relationships differs significantly from another one but (1) is slightly stronger than (2) but not as strong as (3)?

ali_m
  • 295
  • 3
  • 17
jan
  • 89
  • 1
  • Related: http://stats.stackexchange.com/questions/99741/how-to-compare-the-strength-of-two-pearson-correlations/99747#99747 – jona Jul 17 '14 at 14:21
  • Thanks. Is it possible, to say "(1) is slightly stronger than (2) but not as strong as (3)" without touching upon the significance-issue? – jan Jul 17 '14 at 18:49
  • Thanks. Is it **(at least)** possible, to say "(1) is slightly stronger than (2) but not as strong as (3)" without touching upon the significance-issue? – jan Jul 17 '14 at 20:21
  • Calculate the CIs for all 3 and look at them. Though, is it possible you're interested more in something like a regression coefficient than the correlation coefficient? – jona Jul 18 '14 at 01:00
  • OK. And calculating the confidence intervall tells me, whether there are differences between the correlations or not? If they "overlap" the reader has to choice, whether the differences are relevant or not. Right? – jan Jul 18 '14 at 08:03

0 Answers0