# Sigma interpretation in Bayesian Linear Model?

I have two question concerning my output of my bayesian linear regression.

1) I have all beta posterior and obviously, having used a prior for Sigma, i have a posterior for Sigma too, but what can i say about this parameter?

2) i have done two linear regression, one with a normal y[i] dependent variable and one with a t- student value dependant variable. the t-student regression have a lower posterior values for Sigma, can i say it performs better cause there is a smaller variance?

• You are making some kind of distinction between (a) "having a posterior" for a parameter and (b) "saying" something about. What exactly is that distinction? In your second question, you really need to explain your models more clearly so we can understand what "Sigma" actually refers to. – whuber Dec 7 '18 at 17:48