Monday 16 April 2012

Detecting Under the Microscope: The PAS as a Source of the Bias?

The PAS is going to spend 150K to try and tell archaeologists how to use their database and ignore the bias which it seems they feel they can no longer ignore. I wonder how detailed their account will be of where we should take their own activities into account? I mentioned the other day the way that it has been observed that in attempts to use "data" (I use the term loosely) from metal detecting to look at intra-site distributions, that the patterns are skewed by the fact that the finder gets to take the goodies home, so they want to search in the areas where the best and most "stuff" is found, which of course messes up any attempt to treat this as objective sampling.  It gets worse, as this account from a rally indicates:
I went to the Flos tent about 2pm and queued to get my early finds recorded; and they had a map of the fields there with all the find spots marked! Having seen where the "Hot spots" were I proceeded tothose areas and had more finds
One might suggest that this could even be a deliberate tactic. Having identified which rally participants are more likely to bring their finds in for recording with the PAS, it certainly does the database statistics the world of good to show these individuals where most of the finds are being made, so (as we see above) those who bring finds in for recording head straight out there, to bring back even more finds to boost the statistics which the PAS seems to regard as their whole raison d'etre. This would be regardless of the fact that concentrating artefact hunters in those areas, rather than having them scattered, concentrates their destructive effects on certain areas of the archaeological record. If this is deliberate policy, it is an archaeologically destructive one.

We must also remember that if you or I try to find out where precisely a particular artefact on the PAS database was found (for example somewhere near our own home), we cannot. PAS hides this information from the people who pay their salaries on the grounds that some of us cannot be trusted not to go out with a metal detector and visit the same spot to see what WE can find, even though the vast majority of people with access to the database are never likely to every have a metal detector in their hands in their life. Yet here, metal detector owners are being shown where other metal detector owners have had a lot of success finding things to collect and sell.  The policies of the PAS on access to information on findspots seem totally lacking in consistency.

Telling artefact hunters on a commercial rally how to maximize hits and depletion really does not square too well with then spending £150K telling researchers how to make allowance for bias in the resultant database! One way the PAS could facilitate this is writing an individual report about each commercial rally they collect their "data" from, including information not present in the individual records on the database. Such a report would set out the various factors which will affect the information acquired (precise boundaries of area searched, intensity of activity across different parts of the areas), what is likely to have been missed (or simply walked-off with without recording), and most importantly reflecting on the state of what is left of the archaeological record at that spot when they leave. That's how to allow researchers using their records as a source of information can learn what biases may have been introduced by the way the information was collected in each individual case, and only that will allow them to use that information in any form of analysis more sophisticated than a simple and crudely macro-scale presence-absence-here dot-distribution map. 

No comments:

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