Dear DCM experts
When we run DCM post-hoc on a group of participants, we get the parameter estimates for the winning model in DCM_opt for each individual. However, when we specify a DCM model with the same structure of DCM_opt and estimate it in a conventional manner for the same individual we get entirely different parameter estimates.
What is the reason for that? Which one is more reliable?
Thanks a lot
Tali Bitan
|