which test for a 2-way repeated measures design with multivariate data in Matlab or R

[marohde]

Hello,

I have one of the annoying 'which test'? questions.

I have trivariate metric data (x1,x2,x3) from a repeated measures design with 2 factors (one with 2 levels, one with 8; both IVs are within participant). 16 participants were tested. I would like to know whether either of the two factors or their interaction has an effect on the data.

(1) I assume the correct test for this is a 2-way repeated measures MANOVA?

(2) If yes, would someone be so kind to explain to me in a "for dummies" way how to perform one of these in Matlab (R2015b with statistics toolbox)? Alternatively, I could also use R.

I have spent several hours googling and feel a bit lost (not much experience with multivariate data and all examples I found are for between participant designs).

Thanks!

Marieke

 

[mmercker]

Re: which test for a 2-way repeated measures design with multivariate data in Matlab

Hi, 2-way repeated measures MANOVA sounds right, here are some information for the possibility in R:

https://stat.ethz.ch/pipermail/r-help/2010-October/258020.html

alternatively, you could conduct three ANOVAs (which is porbably not the most elegant way but also not the worst). You could follow this with a Discrimimant analyses (DA) in order to check if certain factors or levels can be discriminated via combinations of the outcome

 

[marohde]

Re: which test for a 2-way repeated measures design with multivariate data in Matlab

Hello!

Thanks so much for your answer, I only see it now. Somehow, my notification settings are not right, I have to sort that out, I never realized there was a reply to my post. Thanks a lot for your help. I could have really used your first answer, but independently settled on the sub-optimal second answer.

Thanks!

Marieke