Please help me select the appropriate test for my analysis

[davidbeattie]

Hi,

I have a within-subjects experimental design that features 2 independent factors: (driving condition - 2 levels, sound type - 4 levels), and 1 dependant factor: (emotional response - 4 levels). The data is ordinal. I would like to assess the emotional responses for both independent factors.

I initially used a Kruskal-Wallis H test to assess the difference between emotional responses for each of the individual sound types (of which there are 4). I'm now unsure if this is in fact appropriate.

Please advise on how best to analyse my data.

Thanks!

 

[GretaGarbo]

Maybe a Polytomous (Multinomial) Logistic Regression with two explanatory variables (driving condition - 2 levels, sound type - 4 levels).

I am a little bit confused by:

I have a within-subjects experimental design

Does this mean that each individual has tested each of the 2*4=8 experimental condition? So, does that mean that the individuals should be considered as a random effect and the model should be estimated as a mixed model?