Tuesday, 22 April 2014

PAS Database not a True Sample? Coin Finds Ignored?


A few weeks ago I was discussing a presentation by Helen Geake who was suggesting that the manner in which the information about archaeological finds made by the public differed from the range of artefacts discovered by professionals in excavations was significant in archaeological terms. She was interpreting this in terms of deposition processes. I pointed out that the PASD data was selective because collectors pick up finds differently from archaeologists (what is archaeological evidence is not always a "collectable"). A metal detectorist from central England highlights another problem. Gill Evans of Central Searchers organizes metal detecting rallies commercially, she says that the reason why very few of the finds from these events seem to be getting into the PAS database is that "our FLO has a standing invite to our digs but we never see her". In any case, according to Ms Evans' observations, the FLO "is only interested in the good stuff not broken artefacts and roman grots". If this is what is indeed happening, this is an amazingly un-archaeological manner of selecting data for the database, on completeness and missing out a whole category of archaeological evidence for site use - in what way can one study the Roman coinage of Britain using the PAS database if the basic record is missing unknown numbers of items negatively selected out for their "grottiness'? How is that criterion even defined in archaeological terms?

So what would be the effect of such practices on studies such as Helen Geake's mentioned above, or Philippa Walton's: Rethinking Roman Britain: Coinage and Archaeology (Moneta Monograph No. 137)? How can one "rethink" anything when part of the database one is relying on is skewed from the outset? In what manner can this 'non-grot-factor' be compensated for in the study of the material?  Can it?

UPDATE 27.04.14:
See now the further developments in the discussion of this topic, Monday, 28 April 2014 "PAS and the Pear-Shaped Database".

13 comments:

Cultural Property Observer said...

You may not want to hear this, but archaeological digs give an imperfect picture too. Blow through the upper layers to get to the good stuff. And then miss things where no metal detectors are used. Then miss more things when small stuff goes through screens when dirt is sieved. And hope that you are not digging in a place where the authorities suppress information they don't like i.e., in China where non-Han archaeological evidence is sometimes is not recorded or Arab countries which destroy evidence of Jews.

Paul Barford said...

Thus speaks Washington's great expert on archaeological methodology. I can see it now, the Peter Tompa Institute of Alternative Archaeology, renamed from the 'Foggy Bottom Senior Citizen Egg Sucking Research Institute'.

Unknown said...

Several thoughts on this:

1) FLOs do not select complete objects for recording. In the early days of the PAS, some metal detectorists would first offer up their more identifiable objects to FLOs. However, this was certainly no longer the case when I was an FLO (both in the North East and Cambridgeshire).
2) With regards to 'Roman grots', both Sam Moorhead and myself initiated a campaign in about 2005 to encourage metal detectorists to record all their Roman coins, regardless of denomination, rarity or preservation. We visited more than 100 metal detecting clubs and gave talks outlining the importance of recording assemblages rather than selected highlights. I think this campaign was pretty successful, reflected in the fact that I spent the period 2010-2013 recording huge numbers of poorly preserved Roman coins.
3) As you are aware, all archaeological data is subject to a variety of biases and constraints, regardless of whether it is collected by metal detectorists or archaeologists. If you had actually read any of my book, your question in this blog post would have been answered. I dedicated an entire chapter to a discussion of biases in the PAS dataset and how to approach interpretation of such data. Dr Katie Robbins has also expertly tackled these very same issues in PhD and post-doctoral research.

Unknown said...

Two very separate things spring to mind here:

Firstly, it is clear that you have been mis-informed regarding PAS recording policy and in this particular case, Roman grots. Reporting an off-the-cuff remark from a detectorist is not a substitute for actually looking at the database. Indeed, for the past 10 years, Sam Moorhead and myself have encouraged the recording of all coins regardless of denomination, rarity or preservation. Before moving on to Oxford in January, I spent 3 years recording some very poorly preserved assemblages of Roman 'grots'. These coins were certainly not the pick of the bunch, specially selected for FLO approval and recording!

Secondly, it seems abundantly clear that you have not actually read my book. If you had, you would be able to come some way to answering your own question regarding the validity of my analysis. I devote a chapter to investigating the biases inherent in PAS data and how we can mitigate for these and throughout test PAS data against assemblages from more than 400 excavations throughout the UK. As you are surely aware, all archaeological data, regardless of whether it comes from an excavation, a field survey or metal detecting day out, is subject to a range of biases and constraints. This does not mean that we should disregard it out of hand.

Paul Barford said...

Readers will know I do not generally publish anonymous comments. It is totally unclear who this "Unknown" is. But since he or she would probably complain I was "suppressing information" if I deleted it, I'm putting it up. The next one from this account will be rejected if not signed.

Certainly there is a lot of information going around metal detectorist discussion lists (we don't know if "Unknown" ever reads them) that FLOs are hard pushed to get through all the finds offered and some detectorists are persistently saying that FLOs are selecting objects for recording.

Since they are the (only) ones to know what they submitted to the FLO and see what of it ended up on the PAS database, I am inclined to believe that this is happening. I suggest if "Unknown" wishes to argue this with the detectorists reporting it, then he or she should get over there and have it out with the tekkies.

2) There may have been "a campaign in about 2005 to encourage metal detectorists to record all their Roman coins", but many people currently in the hobby started after that. Probably more than 6000 of them. What specific outreach has been done on this since then?

3) Everyone (even Peter Tompa, bless him) is aware that "all archaeological data is subject to a variety of biases and constraints", especially if you study it in central Europe where such questions are central to a certain methodological school.

The point is that one simply CANNOT treat "data" (inverted commas because they are not) collected by metal detectorists and that collected by archaeologists as in any way comparable.

Then the anonymous critic fulminates: "If you had actually read any of my book, your question in this blog post would have been answered" Wonderful, lovely, but whose book are we discussing here? Have I got it in the PACHI book box? I of all people would be very interested in a "discussion of biases in the PAS dataset and how to approach interpretation of such data", even if it is just one measly chapter. This is obviously of key importance to what the PAS have been claiming for a decade and a half, let's see what underlies it. So what is this mysterious book and how much does it cost?

Unknown anonymous then suggests
"Dr Katie Robbins has also expertly tackled these very same issues in PhD [...] research",
Well, I do not know what she's doing now, but the PhD is available online and I've been through it with great care.

No, no she does not in fact discuss this issue in sufficient detail. Her main focus is on broad distribution, rather than site assemblages, she treats the entire PASD as compiled from metal detectorists (of course it is not). For her study of the patterns of activity, her main external source of information was a questionnaire with loaded questions addressed to artefact hunters.
http://paul-barford.blogspot.com/2010/10/this-could-be-dynamite-if-she-can-get.html

She missed a great deal by not going onto the forums and using the candid comments there, if she'd done that. I feel some of her conclusions would be modified.

But it is nice to see her independent estimate of the numbers of unrecorded objects - something the PAS and everyone else has maintained a discrete silence over, haven't they? Wonder why? Do they think she's wrong?

Paul Barford said...

As for the second rather redundant anonymous comment... Certainly as somebody who has spent a considerable amount of time researching policies on artefact hunting in the UK and elsewhere, it is not exactly appreciated that some anonymous person comes on here being patronising about what I allegedly do not know.

This person will not know how difficult it always has been getting ANY sensible and detailed information from the PAS concerning their policy on anything at all. This was even in the days when they had a forum for interacting with members of the public. They are a very hermetic and opaque organization who feel no obligation to keep anyone informed about what's going on when the questions start penetrating beyond the 'everything's-going-spiffingly-well' facade.

Since Anonymous-Unknown cannot be bothered to sign their comments, I have no idea which book it is I am supposed to have 'not-actually-read' but let me assure this person that if I'd not read it, I'd not be here questioning "the validity of [their specific] analysis" - analysis of what precisely? Bonkers.

I look forward to learning more about "how we can mitigate for these [biases]" in using artefact hunting as an ersatz archaeology. My own investigations indicates that unless you lower the enquiry to the simplest of Kossinnist primitivism, this is most likely not possible.

Unknown said...

I think your declarations of ignorance regarding my identity are pure rhetoric, Paul. As my comment was a response to your post about Rethinking Roman Britain, I thought my identity was abundantly clear. Nevertheless, there was no intention to post anonymously. I was under the mistaken impression that if one signed in through google,identity was made clear. As for posting twice, an error on my part, rather than a deliberate repetition. The first post appeared not to have been accepted.

Dr Philippa Walton

Andy Baines said...

To be honest I have heard many tales through speaking to other metal detectorists and reading on the forums plus my own personal encounter, that there is a growing tendency with the current backlog of finds that need recording to just ignore some of the less glamorous finds. I obviously cannot say this is the case for every flo in every county but it does sound like it is a growing problem and a worry. I dont want to undermine the great and probably very tiresome job flo's do, however there are problems that need to be addressed for the scheme to live up to its full potential. That goes for detectorists and flo's.

Paul Barford said...

Andy, I think we have here exactly the same phenomenon which this blog tries to highlight, there is what the PAS "say" and what the PAS "do". Here the litmus test of that is what we see at the grassroots end, detectorists' experience of what the PAS actually does or does not do.

Andy Baines said...

I think it is probably unfair to label the full PAS scheme and all of its corresponding FLO as the same. I think its more a case of a few rogue's who probably feel over worked and under payed who are letting the side down. Who can blame them though? I cannot begin to imagine how many objects there are to be recorded. This current system has a shelf life and I think we are getting to its expiry date now. The current influx of new metal detectorists as the hobby becomes more popular does not help the current situation and I believe the way the PAS is run now needs to be modified before its crumbles. I really dont think glossing over the cracks is going to help.

Cultural Property Observer said...

Mr. Barford, this sounds like "the perfect" is being made the enemy of the good. No system is perfect but the PAS is still far better than the deplorable situations in places like Cyprus, Greece, Italy, Bulgaria where the State's bad laws discourage reporting of finds. FLOs are human after all and likely have to make calls as to what's important enough to record.

Paul Barford said...

To "unknown", the post is not about "rethinking Roman Britain" I merely used it as an example of the many studies which draw on these "data" and no, I did not remember I'd mentioned it (in passing) in the post written several hours previously so I really did not have any firm idea who was writing to me anonymously. It is simple courtesy either to use a real-name account (what's the problem?) or at least sign the post.

Anyhow, glad we got that sorted out.

I have not yet got your book, it is a bit pricey (90 euros) and I'm a bit allergic to numismo-arguments these days. Also to be honest, I really do not see that in the research on which it is based (http://discovery.ucl.ac.uk/1318144/1/1318144.pdf pp 58-61) you really DO give anything much in the way of a proper breakdown of the "biases in the PAS dataset and how to approach interpretation of such data". Again, from my point of view, your research would have benefited from looking more at what detectorists DO, rather than approaching it (as your thesis does) from "what we have".

Certainly Ms Walton, your thesis chapter 3 does NOT answer "the questions I have".

Paul Barford said...

To Peter Tompa (real name)... is "the perfect being made the enemy of the good" the argument used by US cops gathering evidence at a crime scene? We have all watched with incredulity the effects in court of such lax data-gathering by your police.

I do not see that we can use "second best" evidence as a basis for anything here either. Second best and biased evidence is no evidence at all, except of poor data collection.

I hope there are no criminologists in the US suggesting that money could be saved by getting random members of the public to collect forensic information at crime scenes and then try to use that information to secure a conviction. Are there?

This blog is not about "reporting of finds", this blog is about artefact hunting and collecting and their effects on the preservation of the archaeological record. Please go to an archaeology data management blog to moan about Cyprus, Greece, Italy, Bulgaria and the United States of America not having a PAS-like organization. Try your trite justifications on another audience.

 
Creative Commons License
Ten utwór jest dostępny na licencji Creative Commons Uznanie autorstwa-Bez utworów zależnych 3.0 Unported.