The Heteroskedastic Probit Model in brms (w/ Bernoulli or Binomial)

Hey @paul.buerkner or @matti!

I may be out of the loop, so forgive me if this has been clearly answered before (and I hope you do not mind my tagging you), but is there an analog to the disc parameter in the implementation of binary probit regression (e.g., using the bernoulli or binomial families) within brms. As @matti might guess, I am hoping to fit an unequal variance SDT model to some data for which we have binary (old/new) responses. From @matti’s earlier (but somewhat old at this stage) blog posts, I gather this may not be possible (without confidence ratings or reversion to a cumulative link function).


Perhaps this following paper could help: [1905.09501] Bayesian Item Response Modeling in R with brms and Stan