(Part IV) Empirical studies of networks based on archaeological data are on a rapid rise. So far, the adoption of network methods from other fields has outpaced the development of new techniques and heuristics for dealing with the sometimes peculiar qualities of archaeological network data. Key among the issues faced by archaeologists interested in using networks are the impact of uncertainty and missing data on the properties of the networks we generate. We often must build networks based on an incomplete universe of nodes (because our units of analysis lack current archaeological information or have been destroyed) as well as incomplete information about the nodes we do have (due to sampling issues, different recording conventions, etc.). Further, we often have no consistent way to estimate how much information we are missing. The prevalence of such known unknowns and unknown unknowns suggest that we must carefully temper inferences drawn from networks defined using archaeological data. Importantly, all hope is not lost and these challenges are not unique to archaeology or network data alone. In this session, we ask contributors to explore the potential impact of missing data on empirical archaeological networks and/or test tools and approaches for identifying robust patterns in archaeological networks despite such challenges. Approaches may include, for example, the use of probabilistic estimates and sensitivity analysis already popular in many other areas of archaeological statistical analysis such as seriation or methods specific to network data drawing on the large body of research focused on estimating the shape and properties of so called “dark” networks (common in studies of covert organizations, epidemiology, and infectious disease). In addition, this session welcomes archaeological applications of network methods in general.
