Multivariate model with truncated distribution

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

  • Operating System: Windows 10
  • brms Version: 2.7.3

I have been trying to fit a multivariate model with a the response variables bounded above 0. I imagined that something like this would work:

fit <-brm(mvbind(re_var1, re_var2, re_var3, ...)|trunc(lb=0) ~ ex_var1+ (1|Factor)

but it gives me the following error:

Error: Only 'se', 'weights', 'mi' are supported addition arguments when 'rescor' is estimated.

Is there a way to perform such an analysis with brms?

You could use non-gaussian families with a natural boundary at 0.

Thank you for your response. Indeed, that was my first port of call.

To be crystal clear, interpreting your answer, it is not possible to fit a multivariate model with a truncated Gaussian distribution using brms?

Indeed this it not possible when you model the residual correlations due to the cdf of the MV normal distribution not being implemented in stan yet. It becomes possible as soon as you dont model the residual correlation.

Ok, thank you for the answer and the explanation.