Thanks you to both @ricardoV94 and @OriolAbril for the in-depth answers! Sorry for taking some time to respond! I’ve been coming back to the following post several times, trying to formulate my thoughts.
I’ve actually never used Stan, but Oriol’s comment is a very interesting read!
I’ve tried responding several times here, but never quite managed to formulate my thoughts correctly (they’re to do with understanding the shape of model with and without observed data). As I work more with pymc, I think this topic will become clearer to me. 