Operating System: Windows 10
Interface Version: R versions 3.4.4/ rstanarm 2.17.3 (also tried brms)
I’m trying to fit a basic multilevel model where I have data from three groups (about 35 per group) and the outcome variable consists of values falling between 0 and 1.0:
stan_glmer(outcome ~ (1|group), data=data, control=list(adapt_delta=.999), family = mgcv::betar)
While I expected this simple model to work, I get a warning message that there were divergent transitions. I then used the pairs () plot to try and diagnose things:
I can see that most of the divergences are occurring within a certain area, but I’m still unsure how to address these divergences. Thus far I’ve tried (a) increasing adapt_delta to .99 (and even higher (b) fitting the model in brms, which I’ve read does non-centered parameterization and © using a bunch of different priors [e.g., ranging from N(0,.1) to N(0,10)] for the group variance parameter - including very informative ones as I only have 3 groups and (d) adjusting the max tree-depth.
Are there any other avenues I should pursue?