Grouping techniques employ similarities within data to create new entities,which lend themselves to the interpretation process. This article presents three different grouping approaches, each... Show moreGrouping techniques employ similarities within data to create new entities,which lend themselves to the interpretation process. This article presents three different grouping approaches, each originally developed independently, and applied to a common dataset of archaeological finds. The aim is not to search for the right approach or results, in a competing way, but rather to present the methods as complementary. It is also our intention to stress that a tight connection between theory and statistical modelling is indispensable. Indeed, the use of a particular methodology must be supported by an adapted theory; similarly, a theory without a proper methodological realisation will often not have any actual utility. The integration of theory and method is exemplified in the three case methods. The first method uses metal objects as cultural indicators. The study area is divided into a set of identical geographical units, characterized according to the type and proportions of indicators and grouped using hierarchical clustering. The second approach deals with cultures as standardisations between individuals, using ‘Typenspektrum’ as significant data for identifying different cultures. Groups are defined through kernel density estimation and a cluster analysis, followed by internal and external validation techniques. A third method characterizes the funerary ritual and grave-goods, using a similarity algorithm coupled with clustering procedures to compare the graves with one another. The outcome is validated with exploratory methods and compared to patterns from different contexts. The complementarity of the results shows that each approach sheds light on a certain facet of the same whole. Show less
