I am still new to STAN and trying to familiarize myself as much as possible. My coding skills in STAN aren’t at a point for me to build my own models quite yet, so I am relying heavily on rstanarm and brms for now. I am familiar with rstanarm’s stan_clogit() function, but is there is a straight forward way to fit a clogit model using brms?
I don’t think so, but if you only have one success per group by design then the log-likelihood contribution for each group is not that difficult to code yourself. It is just the linear predictor of the success minus the log_sum_exp of the linear predictors in that group.