credits Rubin (1983) and Little (1983) in their comments on another paper. But basically the above paper shows how to model the weights in a Bayesian framework as opposed to just considering them fixed like rstanarm does.
1 Like