# Wilcoxon signed-rank test - Effect size with small sample size

#### ibarradas

##### New Member
Good afternoon.

I have two dependent variables and I want to demonstrate that one is higher than the other. My number of observations (N) is equal to 9 (it is a small sample). Since my distribution is not normal, I am using the Wilcoxon signed-rank test. I would also like to calculate the effect size. However, in the literature regarding nonparametric tests, I just find the Wilcoxon effect size r, which is given by Z/sqrt(N).
Since I have a small sample, I believe I should NOT use the approximate method of the Wilcoxon signed-rank test, and therefore I will not have Z to calculate r.
Is there a different way to calculate the effect size for a small sample size when I am using this test?
I hope I was clear. It would be really helpful to at least get some literature where I could check this!

Thank you.