Dear Peter,
I have a very similar question: I'd like to compare parameters between two experimental groups that both prefer different families of models explaining the underlying data (DCM for ERPs).
How do I proceed? If I'd like to calculate BMA, I always have to chose a family for running the analysis. Since the groups have different winning families, am I allowed to compare the parameters in the end? Do I have to average over all families?
I am a bit lost here...
Thanks in advance!
Best
Fabienne
|