0

I would like to run a multivariate mixed regression MCMC model with two response (independent) variables, namely Boldness scores (continuous variable) and Aggression ranks (ordinal ranks). Trial numbers (integers) are the fixed effect while individual ID is the random effect. I'm using a mixed model approach to partition between-individual co-variance from within-individual co-variance. I would much appreciate if someone lets me know how to do this, and which package to use, preferably in R and what priors to specify. Thank you very much in advance!

BP86
  • 1
  • 1
  • Sounds like an interesting problem, but StackOverflow is more about practical coding problems. Higher-level modeling questions might be better received on [CrossValidated](https://stats.stackexchange.com). As for MCMC in R, RStan is state-of-the-art for Bayesian analyses; `rstanarm` is a good starting place for more standard models. – merv Apr 21 '20 at 17:33
  • @merv: I posted this question in CrossValidated, but it was subsequently removed because this is a practical question. Thanks for the suggestion with rstanarm. I have used this approach before, but this cannot be used for a combination of continuous and ordinal dependent variables. For example, family=c("guassian", "ordinal") usage is not possible in rstanarm...... – BP86 Apr 22 '20 at 09:06

0 Answers0