Abstract

AbstractAbstractSince 1997, the Portable Antiquities Scheme (PAS) has been recording over 860,000 objects in an online database (www.finds.org.uk), that is unique the world over. These objects have been recovered by the general public from across the rural and urban landscape of England and Wales, and range in date from the Palaeolithic period to the early modern era. These data have been used by academics in over 350 projects (www.finds/org.uk/research), but there are as yet no established methodologies for dealing with the ‘bias’ inherent in the data. This article looks at the history of sampling theory and presents a new seven-stage categorisation of collection bias (deposition, preservation, survival, exposure, recovery, reporting and recording). Through discussions of the impact of such ‘bias’ on the spatial distribution of PAS finds, this article explores the scales at which their effects are most prevalent. Examples are drawn from Northamptonshire, Hampshire and the Isle of Wight, and cover a number of different approaches that combine descriptive and inferential statistics, and ethnographic and textual analyses.