One population effect missing in brmsfit object

Hi everyone,
I’m a beginner in brms, so I might be missing something obvious here.
So I’m fitting a number of ordinal probit models on an ordinal variable. One of the independent variables is a categorical variable “condition” with 5 levels: active_arrhTMS, active_rhTMS, baseline, sham_arrhTMS and sham_rhTMS. Logically, I would expect all of these levels to appear in population-level effects. However, active_arrhTMS is missing. Please see the model and the resulting brmsfit object below. Might there be a particular reason for an effect to be discarded when fitting?

model_task_5 <- formula(probe.response ~ zbv * zlog.apen + probeix + block_num + condition + visit + (1|subj/condition))
model_task_5_fit <- brm(model_task_5, data = dataset, family =cumulative("probit"), init=init, iter = 5000, control = list(adapt_delta = 0.99, max_treedepth = 12)) %>% 

Operating System: macOS Monterey Version 12.3.1
brms Version: 2.16.3

I believe that, if factor levels aren’t specified, brms takes categorical variables like “condition” in alphabetical order and uses the first as the reference level by default (i.e., not shown in population effects, since the factor coefficients represent differences from reference).

1 Like