We describe a combination of methods applied to obtain reliable estimations of population density using archaeological data. The combination is based on a hierarchical model of scale levels. The necessary data and methods used to obtain the results are chosen so as to define transfer functions from one scale level to another. We apply our method to data sets from western Germany that cover early Neolithic, Iron Age, Roman, and Merovingian times as well as historical data from AD 1800. Error margins and natural and historical variability are discussed. Our results for nonstate societies are always lower than conventional estimations compiled from the literature, and we discuss the reasons for this finding. At the end, we compare the calculated local and global population densities with other estimations from different parts of the world.