Semiparametric Quasi-Bayesian Bootstrap Procedures For Dichotomous Outcomes