Using emmeans on brms multivariate model


I ran a multivariate brms model (e.g. DV1, DV2, DV3, DV… ~ pred1 * pred2 * pred3+(1+pred1 * pred2 * pred3| Subject)). Now I want to use emmeans to further investigate any significant interactions between my predictors. However, if I use emmeans the way that I normally do, I get the ouptut averaged over all my predictors (or at least it seems like that is the case). What I want is to investigate the interactions per dependent variable. Is there any way that I can get the emmeans ouptut per dependent variable?

Kind regards,



Have you had any success using emmeans for a multivariate brms model, or found a good work-around?

If you have difficulties with emmeans, you might consider the marginaleffects package instead: Marginal Effects Zoo

1 Like

Yes, I was able to use emtrends for each individual DV by using resp=“nameDV”. I think it works for emmeans too. Good luck!

1 Like