I have a test design in which I have 1 base condition and around 14 test conditions. Each of those 15 conditions has a total of 15 observation points (so equal sample size). The base condition represents the data points in the original sample, whereas the test conditions all represent the data points after a different intervention. The data is continuous.

Now I would like to compare all conditions with each other, so that I can end up knowing which intervention was the most successful, essentially a ranking.

Most likely, the distributions of the conditions are not normal and variances will differ across the conditions, because of this a simple ANOVA cannot suffice.

I am now thinking that the best solution might be a non-parametric test like the Wilcoxon or Kruskall-Wallis, however, I would prefer it if there is a parametric option viable.

Which test is appropriate in this situation?

  • $\begingroup$ Could you please provide us with some more information about your data? How are they distributed? Some plots would be really helpful to see if a parametric option is viable. Also for the non parametric case I suggest looking at this answer here: stats.stackexchange.com/questions/107945/… $\endgroup$ – Patrick Oct 18 '19 at 17:11

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.