Collection Bias and the Portable Antiquities Scheme

Katie Robbins

The Portable Antiquities Scheme (PAS) was set up in 1997 to record archaeological finds made by members of the public. Since then, over 540,000 artefacts have been recorded on an online database, creating a unique dataset that is the envy of many countries worldwide. This dataset is increasingly being used by academic researchers to study artefact classes and to identify previously unknown archaeological sites, but so far there has been little detailed research into the quality of the data itself. In particular, there is a need to understand those factors that influence the geographical distribution of the data, so for example, to what extent are the data shaped by the number of active metal detectorists in the area, the extent of cultivation, or the relationship between detectorists and members of the PAS. Until we understand the biases behind the collection and recording of these data, we cannot expect to be able to understand the data themselves.

This paper will outline the many ways in which bias can enter the archaeological record. It will draw examples from a wide range of disciplines, including geography, ecology and palaeontology as well as archaeology, to discuss collection bias in terms of four categories – geographical bias, methodological bias, visual bias and individual bias. Using examples from Hampshire, Northamptonshire and the Isle of Wight, this paper will illustrate the biases inherent within the PAS dataset. Although focussed on the PAS, this research will highlight the need to understand the biases inherent in any dataset, for without a true understanding of the data we are using, we cannot expect to improve the quality and accuracy of our research.