The Portable Antiquities Scheme announces ?150K research project to study its databas

[BAJR]

[ATTACH=CONFIG]1070[/ATTACH]

The Leverhulme Trust has given the British Museum a 3-year Research Project Grant of ?149,805 for the project, 'The Portable Antiquities Scheme database as a tool for archaeological research' which starts today. Katherine Robbins, currently completing a Collaborative PhD at the University of Southampton in which she is analysing the data gathered by PAS in three pilot areas (Hampshire, the Isle of Wight and Northamptonshire) will be employed as Research Assistant, with Roger Bland, Keeper, Department of Portable Antiquities & Treasure.

[ATTACH=CONFIG]1071[/ATTACH]The project will analyse the factors that underlie the Portable Antiquities Scheme database, which currently consists of over 750,000 archaeological objects found by the public across England and Wales. It will analyse the spatial distribution of the data, comparing it with other datasets; it will also survey finders and will produce a report and web resource which will enable the many researchers who use the data to understand the biases in the dataset.

The PAS was founded in 1997 and has operated across England and Wales since 2003: it has a network of 39 locally-based Finds Liaison Officers (FLOs), managed by 4 staff at the British Museum and supported by 5 National Finds Advisers. Over 14,000 metal detectorists and other members of the public have offered finds for recording. The PAS database is providing a rich and detailed source of information with national scope that is available online at http://finds.org.uk/database. It is increasingly used by academic and professional archaeologists to study the past and to inform planning decisions.

However, although 62 PhDs, 12 major research-council funded projects and 109 MA or BA dissertations are known that currently use the PAS database, the data is not being used to its full potential because there has been little detailed research on the nature of the data and some archaeologists do not use it for this reason.

There are wide variations in the frequency of finds recorded on the PAS database from around England and Wales: there are 28.6 objects per km2 from the Isle of Wight, compared with 0.7 per km2 from Devon. The counties of Norfolk and Suffolk account for 20.4% of all finds of Treasure since 1997, but for only 7.0% of the land mass of England and Wales and 4.1% of all Scheduled Monuments. As yet, we do not fully understand how to interpret these variations. Do the concentrations of finds from the Isle of Wight and East Anglia mean that these areas were much richer than other parts of England or do they simply reflect the amount of metal detecting there? There is an urgent need to understand in greater detail the factors that influence the geographical distribution of the data and the relationships between collection practice, artefact type and space. The project will therefore answer the question: what underlying factors govern the spatial distribution of finds recorded by the PAS?

By providing a clear analysis of the factors underlying the dataset, this project will enable the rapidly growing PAS database to be exploited to the full in future research on the archaeology of the UK. The PAS database also provides an unparalleled scale of data that can be used in the study of sample bias in archaeology.

The survey of collection bias among amateur finders will have applications outside archaeology (e.g., in natural history where a great deal of data is gathered by amateurs) and it will also have an impact on studies of similar finds outside England and Wales. The use of spatial statistical analysis on this type of data is also cutting edge and the combination of spatial analysis with a study of finder behaviour has not, to our knowledge, been undertaken in any other field. This study will help to transform the use of the PAS database in research.

Under the supervision of the Principal Investigator (Roger Bland), the Research Assistant (Katherine Robbins) will use techniques already developed in her PhD, which looked at 3 pilot areas only, to extend the study across the whole of England and Wales. Katherine will continue to work with Graeme Earl as a Visiting Fellow in the Archaeological Computing Research Group at the University of Southampton, and continue to be advised by Chris Lloyd from Queen's Belfast.

She will map the PAS database against key archaeological datasets, especially Historic Environment Records, and others that she omitted from her PhD for reasons of time. She will use spatial statistical techniques, within a Geographic Information System, to generate intensity maps of find locations and selected classes of finds. She will explore correlations between the intensity maps and will build a model to incorporate information relating to finder activity and other perceived biases. Robbins will also analyse finder and Finds Liaison Officer practice, using a combination of statistical techniques with qualitative data from questionnaires and surveys.

An academic panel will advise on the project and a conference in spring 2014 will include papers from experts who will analyse and discuss data supplied by the PAS covering a range of periods, artefact types and geographical areas. These will be published online and as a book. There will also be a Report written by Robbins and Bland which will bring together the results of the project to produce a definitive study of the data recorded in the PAS database. This will identify and analyse key features of the data and will define the best ways to present the data, with their inherent biases, in a transparent fashion. This will be a British Museum Research Paper available in print and online. In addition, guidance for researchers on how to interpret the spatial distribution of PAS data will be developed on the PAS website, besides articles in popular magazines and two peer-reviewed journals.

[Dinosaur]

Interesting that they're going to look at the behaviour/practices of the metal-detectorists/finders and the FLOs etc. On the Catterick Metal Detecting Project a few years back (published in YAJ, can't remember which year, ?2009ish?), apart from recording the objects and their locations, we also recorded who found what, what type of machine they'd been using at the time, ground conditions and also detectorist behaviour: whatever you did they were unable to stick to 'organised' searches, resulting in completely false 'hotspots' where areas were deemed to be 'rich' and hence stripped of every last microscopic scrap of metal using more and more sophisticated equipment at the expense of adjacent areas that probably contained just as many finds, and this was all just in one field. One would suspect that the PAS is to some extent recording similar behaviour at a rather more macro scale.

[BAJR]

Something I also noted as well when I did the Rallies. the Brownian motion of detectorists - This was back in 2007-9 when I did a similar numbered finds, GPS etc... names... and took machine type. but was never sure why -

Could be interesting though.

[Dinosaur]

... and took machine type. but was never sure why -

We'd naively thought that the comparative results would be worth loads to the winning manufacturer? - turned out it wasn't allowed though :face-crying:

- although of course we've now got a good machine with some interesting customised software loaded

[geo]

Seems that you used the wrong people on the survey. Did they have a random search brief or were transects marked out and so on ? The type of machine used and their success rate is a bit of a myth as much depends upon ground conditions on a particular day, individuals search methods and often downright luck

[Dinosaur]

Actually you'd be surprised the difference between the different equipment and software (and yes, this varies with ground conditions, particularly how much air's in the soil, which seems to matter for some reason), but this was nearly a decade ago so maybe the technology has converged? As BAJR seems to agree, you can organise as much as you like, lay them out lanes, whatever, but by the end of the afternoon they'll all still be crammed in one small corner - except for the one bloke 200m away who thinks he knows better....

Screening the web sales sites would add unreported data, it's amazing how often the vendors are quite happy to blatently gove the provenance of the plundered booty. We (local archaeologists, EH, the police etc) know there's an Anglian cemetery being dug out somewhere around Catterick cos thats what's tagged on all the finds for sale on the web (which include non-metal beads etc so we know they're taking a spade to burials), so its possible to add a vague spot to distribution maps and even get some idea of date etc, just would be nice to know exactly where it is and get it either protected in some way or dug properly

[geo]

Yes there are so many factors which affect detection capability in the field. The amount of air in the soil is one as you say and you will see detectorists avoid freshly tilled soil for this reason especially if it dry. The use of power harrows to make a seed bed often produces this affect when drilling late spring crops such as peas and despite a very inviting field few will bother to search. Frost is another factor with a better response after a good hard freeze. I think these factors still remain as technology can give you more programmes to input plus more value add ons ,but at the end of the day it is often a case of going over an item and even the general run of the mill detector will outscore the die hard professional with his latest ?1000 plus gismo.

I do take your point on organisations which is why such surveys need to have strictly defined conditions as to how they are carried out otherwise the end results can be far from acceptable. Transects have their limitations as broken pots with many sherds will go along way in the ploughsoil whereas a few metallic casual losses will not and so the results from fieldwalking and detecting along the same transects are just not comparable. I find total coverage of selected areas a better method using detectorists who will stick to the requirements.

As for the putative plunderers all i can say is a certain archaeoblogger is making hay with your comment and predictably using them in his usual selective way as if he is trying to prove the world is indeed flat so i will not elaborate and give him more to blog with.

[Unitof1]

An academic panel will advise on the project and a conference in spring 2014 will include papers from experts who will analyse and discuss data supplied by the PAS covering a range of periods, artefact types and geographical areas. These will be published online and as a book. There will also be a Report written by Robbins and Bland which will bring together the results of project to produce a definitive study of the data recorded in the PAS database. This will identify and analyse key features of the data and will define the best ways to present the data, with their inherent biases, in a transparent fashion.
This will be a British Museum Research Paper available in print and online. In addition, guidance for researchers on how to interpret the spatial distribution of PAS data will be developed on the PAS website, besides articles in popular magazines and two peer-reviewed journals.

Just imagine coming out of university trowel in hand sorry phd (sorry uncompleated collabarative pdh)and landing this baby. What I dont undersatnd is why is this being done "in house". Isnt all the data freely available to anybody to interpretate in any way they want. Why does it need some official hand and ?50K a year for three years, but then this is the transparant world of secondments between the department of media sport and culture and The Britsish muusuleum, presumably who ever funds the PAS (is it a lottery) and now some factors called Unilever

There is an urgent need to understand in greater detail the factors that influence the geographical distribution of the data and the relationships between collection practice, artefact type and space. The project will therefore
answer the question: what underlying factors govern the spatial distribution of finds recorded by the PAS?

Just what is so urgent about what underlying factors.

Just thinking out allowed what if they, whoever that might be, had set up a compertition: ?150 k to the best urgent answer to what underlying factors govern the spatial distribution of finds recorded by the PAS. Would it have generated anything better?

Not really the done thing but this is. oh theres an academic panal, advice and a conference

[Unitof1]

we used to have a pet metal detectorist who might have asked if they should have got the job- well maybe not, bit over their head