Sunday, January 7, 2018

The Joy of Data Management, or Where’s the fourth miscellaneous censer sherd from lot 307-2-2

By Angela Huster

One of my least favorite parts of the lab work at Calixtlahuaca was doing ceramic type changes. Not because I disliked redoing my own work – there’s actually something satisfying in knowing that it’s right now – but because every single change had to be recorded in at least two, and usually four places (the label on the sherd, the type collection datafile, the paper form for the lot, and the computer datafile for the lot). It just seemed like an awful bother for a single change. I regularly tried to pawn off making the changes on anyone else I could sucker into it. We even had a form to keep track of where a change had been entered!

However, as I move on to working on other projects, I have come appreciate exactly how comprehensively the Calixtlahuaca Project tracked their artifact data. As an example of this, a couple years ago, after we were finished with the ceramic classification, we decided that we needed to subdivide a type. (From Misc censers, to Misc censers and Cut-out censers, for anyone who cares.) At that point, a fair number of sherds had been removed from their original lots, either as examples for the type reference collection, or as samples for particular technical analyses. However, because of our data tracking, I could pull not only a list of lots with the type, but also lists of cases where the sherds in question were stored someone else, which made relocating the pieces for reanalysis far easier. For one of the legacy datasets that I’m currently working with from another project, this simply isn’t possible – we know pieces were removed from lots for various analyses, but what was pulled and when it was taken are unknown. (Kintigh 1981 is an interesting look at the same issue in museum collections, with a focus on the possible bias introduced by archaeologists disproportionately pulling high-interest decorated types out of a collection.)

Miscellaneous censer sherds in the the type collection, prior to splitting the type. Knowing that these particular pieces were in the type collection spared me having to fruitlessly look for them in their lot bags.
While I doubt I’m ever going to love double or triple-redundant data tracking systems, I can see the value, and on my current project, nothing comes out of a lot bag unless it’s recorded that it’s now stored somewhere else. I suppose I shall now proceed to torture another generation of students by making them record type changes in multiple places…

Worked Cited:
Kintigh, Keith W.
                1981       An Outline for a Chronology of Zuni Runs, Revisited: Sixty-five Years of Repeated Analysis and Collection. Annals of the New York Academy of Science 376:467-489.