The estimated standard deviation of the posterior distribution. It is calculated as
.
where N is the number of MCMC observations, B is the number of burn-in observations, is the value of the selected estimand on the i-th observation, and
is the estimated posterior mean.