Prior on intercept(s) for ordinal regression in brms

Interfaces brms
1

Please also provide the following information in addition to your question:

  • Operating System:
  • brms Version:

I’m modeling likert scale data. The likert scale has 7 points. I’m plotting priors to make sure the model I’m fitting makes sense, but I can’t seem to get the priors on the intercepts right. I’m trying to tell brms that the most likely response is the 5th and 6th likert scale point but I only seem to be able to increase the probability of the first or last likert scale point with a prior like the one below.

          prior = c(set_prior("normal(-1,.5)", class = "Intercept")

Any advice?

2

The problem is that brms currently does not allow to set different priors on the different intercepts of an ordinal model, but I agree this should be possible.

Would you mind opening an issue on https://github.com/paul-buerkner/brms/issues?