Mi() to propagate uncertainty?

Interfaces brms
1

Quick question, just want to make sure I’m not doing something incorrect:

I have a regression model I’m planning on running in brms. The x and y variables were determined by a separate model, which means each is defined by a posterior distribution. I could run brm_multiple where each iteration is a different combination of the posterior distribution, but that takes forever to compile. I can use mi() instead with the se of the posterior distribution for each variable, correct?

2

I am not an expert on that topic, but yes I believe you can. That seems analogous to what McElreath covered on chapter 15 of his 2nd edition. See here for a brm() version: 15 Missing Data and Other Opportunities | Statistical rethinking with brms, ggplot2, and the tidyverse: Second edition. I’m very curious on others’ thoughts, though.

2 Likes