On Bayesian estimation of marginal structural models

Formalizing the notion of using a pseudo‐population as a data generating mechanism with particular characteristics