Thank you very much for your answer! My approach was to check whether the best model I found in the first analysis also explains the data of another independent sample by comparing this best model with a model lacking any drug influence. Is this a valid approach at all? Or should I estimate all the models of the first analysis for the second data set, too, and perform BMS on the same model space as in the first analysis? Is there any method described how to validate a “best” model of one analysis in another independent data set?
Thanks again for your help!!
|