Testing interactions in nonparametric data.

#1
I am using SPSS to analyze data from agricultural trials. Some of the data is border-line nonparametric so I have been using both conventional ANOVA and nonparametric methods to analyze main effects. Now I have encountered a problem in that I am interested in assessing the significance of main effect interactions. For the parametric data I can use a GLM Univariate analysis but what can I use for the nonparametric data?


Cheers

Nic
 
#2
Hi Nic,

This is exactly the limitation of nonparametrics--you can't test interactions (at least not to my knowledge).

What kind of data do you have? You can run parametric models on non-normal data.

Karen