repeated measure non parametric anova between & within subjects

Discuss statistics related things
Post Reply
mata_hari
Posts: 2
Joined: Thu Mar 02, 2023 4:05 pm

repeated measure non parametric anova between & within subjects

Post by mata_hari »

Hi,

I have two groups, one experimental (n=27) and the other one a control group (n=19).

I want to run a repeated measure non-parametric anova both between & within subjects. I have the time as a repeated measure factor (before and after therapy) and a between-subject factor (the group).

How can I run this analysis in jamovi? In the non-parametric anova with the Friedman test seems to me I can't insert the between-subject factor.

Thanks
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: repeated measure non parametric anova between & within subjects

Post by jonathon »

i'm not sure such a procedure exists ... but i'm not sure.

jonathon
User avatar
reason180
Posts: 274
Joined: Mon Jul 24, 2017 4:56 pm

Re: repeated measure non parametric anova between & within subjects

Post by reason180 »

I don't think there's an generally-accepted way to compute a truly non-parametric ANOVA interaction, whether repeated-measures or not. This is because of how "interaction" is defined with respect to variance-partitioning. People have tried various ranking approaches that have turned out to be invalid for correctly inferring interactions. A more fruitful approach might be to try to transform the data and/or eliminate outliers so that your data better fit the underlying assumptions of ANOVA.
Last edited by reason180 on Fri Mar 31, 2023 6:12 pm, edited 1 time in total.
User avatar
MAgojam
Posts: 421
Joined: Thu Jun 08, 2017 2:33 pm
Location: Parma (Italy)

Re: repeated measure non parametric anova between & within subjects

Post by MAgojam »

Hey @mata_hari,
the Friedman test is a nonparametric alternative to a one-way repeated measures ANOVA, but unfortunately, there are no nonparametric alternatives to a two/three-way repeated measures ANOVA. If you are in the situation where the assumptions are not satisfied, you may consider performing two/three-way repeated measures ANOVA on the transformed and untransformed data to see if there are any significant differences.
If both tests lead to the same conclusions, you may not choose to transform the outcome variable and continue with the two/three-way repeated measures ANOVA on the original data.
It is also possible to perform robust ANOVA tests using the WRS2 R package, but staying in jamovi you find Robust ANOVA in the Walrus module, available in the library.
There will be an update soon (I'm finishing it) with the output also for the interaction of two/three-way factors.

Cheers,
Maurizio
Post Reply