Models of government formation processes in multi party democracies are usually highly sensitive to the rules that govern the selection of formateurs (i.e., the parties selected to propose a potential government). The theoretical literature has focused on two selection rules: selection proportional to seat share, and selection in order of seat share. In this paper we use a new data set on government formations in 11 parliamentary democracies to empirically assess which selection rule most closely approximates the data. We find that while there is little empirical support for selection in order of seat share, proportional selection fits the data well.