Illustrating distributional hierarchical model

I’ve adapted the vignette on distributional models ( for my data:

  • Participants split into two groups (~40 per group)
  • Each participant provides measurements over ~ 50 days
fit= brm(bf(symptom ~ day*group+(day|ID), sigma~group+(1|ID)),data=df)```

I find in my results that sigma_group is different between the groups. I was wondering what is the best to illustrate this? I think the brms call conditional_effects does not show the right thing as it just shows effects on the mean (day, group, day*group), but not on the sigma.

I guess I could manually do a bar plot where first I compute for each person the standard deviation of their measurements and then compute the mean (and standard error) for each group to make the two bars?

Many thanks

Take a look at argument dpar of conditional_effects which you can use to illustrate the predictions on sigma.

thank you! I didn’t realise this setting existed…