Thanks a lot peter. I will try doing this. Thanks for your help. I really appreciate it
Kushal J Kapse
Aphasia Research Laboratory
Boston University
________________________________________
From: Zeidman, Peter [[log in to unmask]]
Sent: Wednesday, July 31, 2013 4:39 AM
To: Kapse, Kushal Janardan; [log in to unmask]
Subject: RE: DCM.U.dt = SPM.Sess.U(1).dt;
Hi Kushal,
I'm glad to hear you are making progress.
> Should the models Pp<0.9 CI should be dropped as they suggest not the
> strong probability of direct effect they depict?
You could try Bayesian Model Averaging (which is an option when you specify Bayesian Model Selection). This will average the parameters across all your models, but weight each parameter by its parent model's posterior probability. That way the average will many reflect models 3 and 4, and only to a lesser extent models 1 and 2.
Best,
Peter.
|