Vignette on generating data with brms

Hi all,

I’m interested in exploring prior sensitivity by generating simulated data according to a model and then having a way to check the sensitivity by looking at parameter recovery, intervals, and (if possible) differences in loo values resulting from the simulation. I understand this can be done in brms but I am unable to find a clear vignette on the topic. If there is one out there please send me the link.

Thanks in advance,

David

This isn’t exactly what you’re asking for, but I’ve blogged on how to do simulation-based power analyses with brms models, which could get you started. To adjust for your needs, you’d change the evaluation criteria from power to CI width, parameter bias, and so on. You can find the first post in the blog series here.

1 Like

I think this will do the trick. Thanks.

Cheers!