I have conducted two Maxdiff studies and estimated using MNL and HB. The results are very similar for all except one item in both maxdiff studies (they are studies with completely different items).
The MNL model has a rather low RLH fit but my HB fit is very good and, as they both give similar results I assume this means the low MNL fit is due to the heterogeneity in preferences in the sample. But I am at a loss to explain the one difference in both studies.
Do you have any suggestions for further investigation?