Dear SPM- Experts,
I run an rfx Bayesian Model Selection including 14 subjects with 64 model each. I noticed that the ranking of the models and also the values of model expected probability and model exceedance probability change every time I repeat the analysis although I did not change model space. Model A for example is two times the best model, however, when running the analysis a third time, it is only the fourth best model. Why are such differences? Do you have any idea how to proceed in such a case?
Thanks a lot for any explanations!
carmen
|