After performing a T-test, we can use to Cohen d test to calculate the effect size. What would be the equivalent for calculating the effect size after performing a Wilcoxon test?
The probability of observing a higher value in group A (or in B, as you wish) when from each group a random value is selected. You may express it as log(odds).
The tested hypothesis is Pr(xA>xB) = 0.5, equivalent to test log(odds) = 0.