Dear Guillaume,
Thank you for the quick reply and for your feedback.
I tried modelling only the interaction group x condition, and I was able to define contrasts in order to extract betas and SE for every single column of the design matrix (this time, with 4 columns).
My issue is that this model does not only include the interaction term, but rather both main effects (of group/condition) and interactions. The betas I get this way are the same as the betas I got with the previous model (8 columns) when testing for a single regressor in the interaction group x condition (e.g., G1C1). However, they are not the same as the betas I extracted previously from my last 4 columns, which reflected only the interaction term. I am specifically interested in the SE of the betas of the interaction effect, since that is what I am representing in my graphs.
I hope you can give me an idea if this is possible.
Best regards,
Cecilia
|