Spatial lag model with brms and predictions/marginal_effects

Hi All,

I am estimating a spatial SAR model in brms using the cor_sar() function. I am now trying to get marginal effects and predictions out of the model, but I am unsure whether the functions in brms or those included in strengejacke take the spatial structure and covariance into account, i.e., whether the total impact and feedback is calculated in those functions.
I have been looking around for a while, but I can’t find any information on this. Sorry if I missed something, thank you very much for any help.


Please also provide the following information in addition to your question:

  • Operating System: Mac Osx Mojave 10.14.5
  • brms Version: brms_2.9.0

marginal_effects will always exclude autocorrelation structures, but when you compute predictions via posterior_predict, they will be included.

1 Like

Awesome, thank you so much for the quick reply.

And thanks for all your work on this awesome package!