Is there an implementation of bayesglm in Stata? (That is, approximate maximum penalized likelihood estimation with specified normal or t prior distributions on the coefficients.)
Is there an implementation of bayesglm in Stata? (That is, approximate maximum penalized likelihood estimation with specified normal or t prior distributions on the coefficients.)
I have some code written with normal priors (the user specifies the prior mean and variance) for logistic regression and poisson regression. I haven't documented it as well as i should yet and haven't had time to finish up the code for other common regression models. its on my 'to do' list for when i'm done teaching in a few weeks. happy to send you the program if you want to play around though.
Unfortunately there's very little Bayesian anything in Stata. There's a library for interfacing Stata with WinBUGS that you could implement GLM in, but I've always opted to use BRugs instead so I can't vouch for it (but here's someone using it on your schools dataset: http://yusung.blogspot.com/2007/01/using-winbugs-… Out of curiosity, why do you ask?
Stata is not very good in this department. I have been trying to convince Stata folks to make the move to R in this regard. There is a neat function called MCMCglmm which is capable of doing binary, multinomial, poisson, etc. random effects models in a Bayesian framework.