On establishing ceramic chemical groups: exploring the influence of data analysis methods and the role of the elements chosen in analysis

Abstract

Since the 1970s, archaeologists have increasingly depended on archaeometric rather than strictly stylistic data to explore questions of ceramic provenance and technology, and, by extension, trade, exchange, social networks and even identity. It is accepted as obvious by some archaeometrists and statisticians that the results of the analyses of compositional data may be dependent on the format of the data used, on the data exploration method employed and, in the case of multivariate analyses, even on the number of elements considered. However, this is rarely articulated clearly in publications, making it less obvious to archaeologists. In this short paper, we reexamine compositional data from a collection of bricks, tiles and ceramics from Hill Hall, near Epping in Essex, England, as a case study to show how the method of data exploration used and the number of elements considered in multivariate analyses of compositional data can affect the sorting of ceramic samples into chemical groups. We compare bivariate data splitting (BDS) with principal component analysis (PCA) and centered log ratio-principal component analysis (CLR-PCA) of different unstandardized data formats [original concentration data and logarithmically transformed (i.e. log 10 data)], using different numbers of elements. We confirm that PCA, in its various forms, is quite sensitive to the numbers and types of elements used in data analysis.

A Comprehensive Workflow for Compositional Data Analysis in Archaeometry, with Code in R

“…The field of archaeology spends a lot of time collecting data to find groups (for example, Michelaki et al (2013); Rosenow and Rehren (2014); Waksman (2016); Jackson and Paynter (2016)). The reason to group may arise from a hunch that there is interesting patterning in the data.…”

The production and circulation of indigenous lead-glazed ceramics in northern Peru during Spanish colonial times

“…PCA using all of the remaining elements determined by INAA failed to provide much discrimination of the Zaña and Chicama valley groups, possibly due to the fact that some elements were acting as noise in the dataset, effectively obscuring patterns pertaining to region. Such situations in statistical analysis have been discussed by Glascock (2011), Michelaki, Hughes andHancock (2013), and Baxter and Jackson (2001), who discuss eliminating variables/elements from PCA. In our study, this led to the analysis of datasets comprised of 21, 12, 11, eight, six and four elements, with elements for use in PCA determined by the study of biplots and means and standard deviations for each element in each valley group.…”

Iconographic and special links of Uherské Hradiště tiles from the end of the 15th and the mid-16th century

“…The character and scope of the analogies to the Uherské Hradiště material and the need for a better understanding of the creation and production of stove tiles, trading with them and supplying regional centers gave an impetus to the analyses, which have been in progress for the second year with material from Uherské Hradiště, Uherský Brod, Strážnice and Buchlov and Cimburk castles. Their goal is to better identify mutual links and contacts within the production process of stove tiles production (compare Hložek-Loskotová 2014;Holl 1995;Holl-Balla 1994;Loskotová-Hložek 2016;Menoušková 2011;2018;Menoušková-Vitanovský 2007;Michelaki-Hughes-Hancock 2013;Römer-Strehl et al 2005;Rzeznik-Stoksik 2011;Smetánka 1968), and thus contribute to the identification of local production centers. Our analyses are based on the collections of finds that we know from individual sites from archaeological research (for Uherské Hradiště: Geisler-Vitula 1993;Procházka 1979Procházka -1982for Uherský Brod Šimík 2020, 2020a2020b).…”