PREVIOUS   NEXT   CONTENTS   HOME 

3.1.1 Data cleaning and enhancement

A detailed discussion of the process of cleaning and enhancement can be found elsewhere (Naylor and Richards 2005), but it is important to outline the main points here. The organisation of the two recording schemes is radically different, with the PAS database built by data entered from its individual Finds Liaison Officers (FLOs), based around the country. The EMC, in contrast, is based at the Fitzwilliam Museum, Cambridge, where a small team identify the coins reported to them, and enter records onto the database held there. The high number of FLOs results in a database where the data stored have not been entered in an entirely standardised manner and there can be variation, especially in the terminology used for 'object type' and 'class'. This is less of a problem for the EMC, where numismatic terms are applied throughout to rigorous standards. The level of specialist knowledge also varies between the organisations. The EMC is staffed by a number of numismatists specialising in the coinage of early medieval Europe, whereas FLOs deal with material from the Palaeolithic to modern periods, and cannot be expected to possess the same levels of specialist knowledge in all of these fields. As a result, the EMC data provided a ready-to-use dataset, whereas the PAS database, as it was in 2005, required more work before interrogation could begin. It is worth noting, however, that since 2003 the PAS has made efforts to standardise data recording, through its database and the appointment of period Finds Advisors, including ones for medieval objects and coins. This should make the work of any future projects much easier.

The enhancement of the PAS dataset was based around two main factors. First, the standardisation of the descriptive language used in order to aid confident interrogation of the data, and second, the application of a number of classification systems to the highest level possible (Table 1).

The choice of published classification schemes was based around several factors. Firstly, Geake's (2001) PAS Recording Guide made suggestions for some schemes that could be used. These were often utilised more readily by the PAS's FLOs than others and so retaining these as the basic classification was the most sensible option. Secondly, where Geake (2001) had not suggested any particular typology, a range of schemes may have been applied haphazardly across the database. Hinton's (1996) pin classification is a good example of a scheme that had been inconsistently applied. In these cases, the project tried to increase the consistency in the application of a single typology. Thirdly, there were artefact types where no classification had been applied to any great extent. These often included categories of finds that were relatively rare, or where no useful typology had been produced, including Late Saxon buckles, ansate brooches and hooked tags. For these, a simple classification system was produced (see archive).

Table 1: Artefact classification schemes
Artefact typeClassification system used
Strap-endsThomas 2003; 2004
Pins Hinton 1996
Stirrup-strap mounts Williams 1998
Stirrup terminals Williams 1997
Bridle fittings Williams 2007, enhanced by VASLE
Buckles (earlier varieties) Geake 1997
Buckles (Later Anglo-Saxon varieties) VASLE-produced classification
Ansate brooches VASLE-produced classification
Hooked tags VASLE-produced classification

 PREVIOUS   NEXT   CONTENTS   HOME 

© Internet Archaeology/Author(s) URL: http://intarch.ac.uk/journal/issue25/2/3.1.1.html
Last updated: Tues Apr 21 2009