# Brm\_multiple - divergent transitions

**URL:** <https://discourse.mc-stan.org/t/brm-multiple-divergent-transitions/25448>\
**Category:** Modeling\
**Tags:** fitting-issues\
**Created:** [November 26, 2021, 10:42am UTC](https://discourse.mc-stan.org/t/brm-multiple-divergent-transitions/25448 "2021-11-26T10:42:44Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![L.Sch](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.mc-stan.org/l.sch/32/12563_2.png) [@L.Sch](https://discourse.mc-stan.org/u/L.Sch)\
**Post date:** [November 26, 2021, 10:42am UTC](https://discourse.mc-stan.org/t/brm-multiple-divergent-transitions/25448/1 "2021-11-26T10:42:44Z")

</div>

I fit a multilevel model using imputed data (m = 50, using _mice_). I use `brm_multiple` from _brms_ for this because I can simply use the mids object there. Now I get, for example, 4 divergent transitions in one model after the warmup. If I select `combine = F` in the model and look at the 50 fitted models individually, only some of them have divergent transitions, but most of them do not. My question now is whether this is an artifact of _brm\_multiple_, such as the incorrect R-hat warning, or whether this is something I should be concerned about. There is little difference in the estimates of the models with and without divergent transitions. I have already made the following adjustments:

```
control = list(adapt_delta = 0.99, stepsize = 0.01, max_treedepth = 15)

```

Edith: running the model on one combined set only (`mice::complete()`) with `brm` does not give any divergent transitions.