In my experiment I have 3 treatment groups. Behaviours are scored for each subject (while they are isolated) before and after the treatment is applied.

I thought that the Kruskal-Wallis test could only be used for independent data sets (i.e. not on data from the same subject before and after a treatment). However, I have been advised to use the Kruskall-Wallis test and analyse the data separately for "before" and "after" treatment. I understood previously that the Friedman test was the appropriate test as it is used for repeated measures taken from the same subject (but apparently this will be more complicated).

I have been told to use the Friedman test for a dataset containing behaviours scored each week following treatment (during this observation period the subjects are not isolated but can interact with each other). I think the confusion is arising from whether "independent" means i) non-repeated measures on the same subject or ii) independent of influence from the actions of other subjects. However, I am easily confused with statistics and perhaps I am missing a logical reason why using the K-W test for before and after treatment is ok/better than using the Friedman test...?

I understand that the situation is different due to the treatment being applied but then the behaviours being scored are the same.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.