We conducted a sensitivity analysis of the results of weighted voting experiments by varying two features of the experimental protocol by Montero et al. (2008): (1) the way in which subjects? roles are reassigned in each round (random role, RR, versus fixed role, FR) and (2) the number of proposals that subjects can simultaneously approve (multiple approval, MA, versus single approval, SA). It was observed that the differences in those protocols had impacts on the relative frequencies of minimum winning coalitions as well as how negotiations proceed. Our analysis favors a protocol with FR-SA for the future research, because this protocol induces subjects to commit fewer errors in their decision making than the protocol with RR-MA, and because proposal-objection dynamics are more frequently observed under FR-SA than under RR-MA.