Over the last few months I have done tests with two groups under 4 different conditions at 5 levels of difficulty. Untill yesterday, I had never heard of Wilcoxon, correct me if I'm wrong; I'll have to use the unsigned variant of Wilcoxon with the mean values of all test persons at the 4 different conditions for one Wilcoxon. So I'll have to do it 5 times.
My question is: What can I do with the output I get? Is there another possibility to compare the Wilcoxon results of all 5 difficulty levels?
Does anyone have any experiences with this Wilcoxon test?
I will have to do it in MATLAB, so this if you ask me, this looks suitable for my work: http://www.mathworks.com/matlabcentral/fileexchange/25830-mann-whitney-wilcoxon-test
Some information of what kind of study I did: Tested 2 groups (normal hearing and cochlea implant). A test similar to OLSA (Oldenburger Satztest)
Thanks for your help
The only thing I found yet is: Check if p is higher or lower 0.05. If p got a higher value that 0.05, there is no significant different between these two groups.
Source: http://www.methodenberatung.uzh.ch/datenanalyse/unterschiede/zentral/mann.html#16 -> 2.3. Prüfung auf Signifikanz
edit: And a check for critical value of U: http://math.usask.ca/~laverty/S245/Tables/wmw.pdf
Hello Daniel,
As far as I know you are not supposed to repeat Wilcoxon tests between different conditions, as this repetion increases the probability of finding significant differences by chace. You might want to look into the Kruskal-Wallis test, a nonparametric version of ANOVA (or ANOVA itself, if your data is normal).
Good luck!
- Badges
- Science topic