Wednesday, 10 September 2014

Focus on PAS: Trying to get to the Figures

Due to the way the PAS database is organized and the search engine functions, stripping out the spin from the way the PAS data are presented is a thankless task (see also 'Bumping Up the Figures' PACHI 17 May 2011). As we have seen, the staff in PAS central office in Bloomsbury have clearly been attempting to bump up the figures of the Portable Antiquities Scheme by putting on their database all sorts of material besides that generated by members of the public voluntarily reporting things they have found. One of the principal means by which this has been achieved is putting Treasure items (compulsorily reported and not only those found by members of the general public) on their database. The table below (utilising the search facility of the database, ticking the 'treasure' tag) illustrates what has been happening:

 Treasure Items incorporated on the PAS database (complied PMB 9th September 2014)

Treasure results (records)
Treasure numbers
498 (max 11)
221 (max 43)
2,785 (max 1026)
236 (11 hoards)
642  (max 149)
630 (30 hoards)
8,323 (max “2000”)
973 (44 coin hoards, 33 hoards )
14,694 (max 7,065)
981 (49 coin hoards, 40 hoards)
17,054 (max 9,315)
1,600 (82 coin hoards, 33 hoards)
78,724 (max 52,504)
1,194 (74 coin hoards, 37 hoards)
18,162 (max 4,653)
1,292 (115 coin hoards, 51 hoards)
12,588  (max 4,957)
1,104 ( 52 Coin hoards 33 hoards)
5,787 (max 1,470)
(< 9th  Sept)
456 (18 coin hoards, 9 hoards) + the 22000 added on 8th September
3,172 + 22000
158504 + 22000 = 180504

 The table shows a number of interesting things. First of all some 9000 records on the PAS database are of Treasure origin, and an astonishing 180000 individual finds (that's a fifth of the claimed total of a "million"). Oddly enough, though the mysterious "22000" coins recorded by Dr Drost appear in the grand total displayed proudly on the front page of the database, and in the overall statistics count, they for some reason fail to turn up in other searches. 

But look what else has been happening. It seems that there was no finds category "hoards' and coin hoards in the database before 2006. Also from 2010 we see a tendency for hoards not to treated as one record, but split into a number of records. Look also at the increasing trend in the number and size of hoards being found. This cannot be sustainable.

So if we remove those Treasure items (including what -for the lack of more information at present we must call the "Drost 22000") from the figures, we find that today instead of reading "1,021,374 objects within 631,423 records", the Treasureless figure would be:

 840,870 objects within 622, 356 records.
But that is not the end of the obfuscation. As I reported earlier, 'Bumping Up the Figures' (PACHI 17 May 2011). The PAS database includes some totally extraneous information added in March 2010. The Celtic Coin Index (CCI not a PAS initiative) was added, accounting for 37,930 objects and the same number of records and the Iron Age and Roman Coins in Wales (IARCW) accounts for another 52,812 objects in as many records. Some of these records refer to finds made (for example by ploughboys) well before the invention of metal detectors). That's an additional 90,742 records in the database which are nothing to do with PAS outreach to members of the public, they come from other sources entirely.

If we adjust the PAS figures for this insertion too, we get the real figure which is:
  750,128 objects within 531,614 records.
...which is still a long way from "a million" objects voluntarily handed in to the Portable Antiquities Scheme for recording. In other words, the next time you hear a metal detectorist or pro-collecting archaeologist tell you that the PAS has recorded "one million finds handed in by responsible detectorists which should make us all very happy", subtract at least 271,246 objects (and 998 809 records) from that for a more realistic idea. 

These figures can also yield one other piece of information. Taking the overall statistics, according to PAS figures, artefact hunting with metal detectors accounts for 764,706+42,728 of the total 1,021,322 finds claimed. That means 213,888 finds currently in the PAS database did not come from metal detecting. Removing the Treasure items (reporting of which is obligatory) will give an indication of the degree to which artefact hunters voluntarily bringing material to the Scheme for recording is actually occurring. Subtracting the total of 180,504 from the number of metal detecting finds gives 626,930 objects. 

It can readily be seen that 630,000 objects compared with 214,000 objects is far from the sort of proportions supporters of artefact hunting claim "detectorists are contributing". Divided by 8000 artefact hunters, that is 79 objects reported each in eighteen years. Divided by today's more realistic figure of 16000 active detectorists, it is a very poor showing indeed. Obscuring such features and concerns is one result of the way the PAS has been boosting the figures for its own benefit. This does the UK heritage debate no credit and a great disservice.Can we now have some presentation of the statistics in a manner that facilitates their being used to show where these "data" come from and to what they really refer, and whether they really are comparable within the database, or is that asking for too much transparency for seventeen million pounds?

No comments:

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