I’m fairly new to BRMS and Bayesian statistics in general. I was wondering if there was a prior to decrease inner group variance. For example, I’m pulling the 5th and 95 percentiles from the posterior distribution and the values seems a little too extreme based on my domain knowledge. I’m dealing with groups, so it would likely have to be adjusting the standard deviation of the random effect. However, when setting a prior for class = “sd” on the group parameter, the results get even more extreme. I’m not sure if this is the correct prior I should be setting. Any help would be appreciated!
Hi and welcome. A couple of things that will help folks answer your question:
Can you tell us what version you are running? I am guessing brms in R?
Can you post your model and a snippet of data (or simulated data if you can’t share the raw data)?
Also posting what priors you are using or the defaults if that is the case is great too.