I have data which is not normally distributed. Each participant ranked six actions in order of preference (so this is ordinal data right?) I want to see whether there is a significant difference in their preferences depending on the group they are in. As there are more than three groups, I believe that I could conduct a Kruskal Wallis (as this is a non-parametric test) however I also think that a Kendall tau may be appropriate (but as my data is groupings -categorical against rankings-ordinal, I'm not sure correlation is the right thing to do).

I'm still not entirely sure if I am clear with any to analyse this so any help would be much appreciated.

  • 1
    $\begingroup$ If you believe my answer is wrong (I could be wrong), please tell me what I went wrong. I might edit my answer or even delete it. $\endgroup$ – HelloWorld Apr 8 '17 at 13:06

I believe this is an application for Friedman Test, which is a non parametric test for repeated measurement.

Let's look at the Wikipedia:

n wine judges each rate k different wines. Are any of the k wines ranked consistently higher or lower than the others?

  • You have participants, the example has wine judges
  • Your participants rate 6 items, the wine judges rate k wines

Does that sound familiar?

  • $\begingroup$ Thank you for your reply. I don't believe my data to be repeated measures as I would like to compare the ranking between groups and each of the 100 participants (I had slightly more than this) are separated into 4 different groups (so this would make it non-repeated right?) So say if participants in Group 1 significantly preferred item 1 over the others, compared to the other groups. I don't think Friedman would fit with this? $\endgroup$ – Dragonfly Apr 4 '17 at 14:50
  • 1
    $\begingroup$ @Dragonfly Yes, it is. When a participant rates something for 6 times in a row, it's repeated. $\endgroup$ – HelloWorld Apr 4 '17 at 14:51
  • 1
    $\begingroup$ @Dragonfly Data for each participant can be broken into 6 rows. One for each item. This is how marketing professional code their data. $\endgroup$ – HelloWorld Apr 4 '17 at 14:52

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.