The Distribution Map – One step beyond

The distribution map is used to visualise quantities of artefacts found within a site, patterns can be inferred from this. But several questions arise from this: how significant are these patterns? Are they real? At what scale is the interpretation? Also what happens when no patterns are obvious, do we give up?

Every site has some patterning of artefacts, be it clustered, dispersed or random. Visually we may interpret heterogeneity within a dataset which statistically is homogeneous. We can manipulate colour ramps to depict what we want to see, rather than what is there, even though this may be applied subconsciously.

Point patterns can be analysed in a number of ways as highlighted by Blankholm (1990), however collecting data at point precision can be problematic. Point collection of artefacts can be costly in terms of time and money as well as ironically imprecise. Grid collection methods are far more common practice upon settlements sites, yet this also yields problems. However, there are some techniques which can be applied to identify different types of clustering as well as statistically test the significance of the clustering. Some methods, such as moving average window techniques, may appear useful but when critically appraised problems can be identified. Other local methods such as the Getis and Ord’s Gistar (Gi*) and the Local Morans I (Ii) can yield significant insight and go beyond the distribution map.

The back drop of this presentation will be the Late Neolithic settlement site of Keinsmerbrug, located in the province of North-Holland, in The Netherlands, the spatial analysis forms part of a wider multi-disciplinary research group. Analysis of two further settlements is on-going and may be incorporated into the presentation if the data is ready in time.