eBay and the PAS: the data

Hope you enjoyed the brief tour of artefact homes. Thank you to those of you who offered feedback and encouragement, it certainly seemed to interest people!

Tomorrow I’m going to be doing some fieldwork with the Bristol Wood Recycling Project, so there won’t be a post and I’ll be tweeting instead. Either on Sunday or Monday and into next week, I’ll complete the eBay half of my PA2015 month by addressing my failure to engage any actual eBay sellers in my research and making some suggestions as to how eBay still has some potential to be considered public archaeology, despite my lack of qualitative data. As I’ve said before, I hope the ‘failure’ bit can be more of a discussion, it’s not a subject we hear enough about.

Firstly though, as we have had our leisurely tour around Britain (and overseas…?) seeing where eBay-listed items reside, I want to briefly present some thoughts on the data itself. It’s not necessarily a prime concern of mine, I’m more into talking to people, but if eBay is ever to be used in a meaningful way, it will entail some level of data analysis, so I might as well present a bit.

Objects on eBay

Sample objects on eBay, November 2014

Sample objects on eBay, November 2014

As we see in this lovely pie chart, Roman coins make up nearly half of the available data. Spindle whorls have a good go in second place, followed by lampposts and musket balls. Jettons and clay pipes bring up the rear. Is this to be expected? It seems to make sense that there is a more established trade in valuable Roman coins, while spindle whorls can be attractive in their own way. Clay pipe fragments are probably the most common accidental find, but I suppose people just don’t think of selling them.

Objects on the PAS

Sample objects recorded by the PAS, as of end November 2014

Sample objects recorded by the PAS, as of end November 2014

Contrast that with the PAS data. This pie chart, looking uncannily like the old Happy Eater logo, shows Roman coins as by far the dominant artefact type with everything else paling in comparison. This makes complete sense and in fact highlights one of the ways in which eBay and the PAS are not quite equivalent datasets. Roman coins are far more likely to be recognised as being of interest and added to the PAS than, for instance, clay pipe fragments and I suspect the PAS could not survive if it was constantly recording the latter. However, the proportion of Roman coins to jettons is not too far off in both sets of figures, suggesting that although the PAS favours certain artefact types, there is correlation between those types in that and other public-driven contexts.

What does this mean for public archaeology?

Leaving aside for a moment the problems of the sale of potentially important artefacts, it would seem from this data that Roman coins are a key artefact to focus on if we want to contact sellers and understand the movement of objects in and around eBay. However, I tried another bit of analysis that tells a slightly different story. Please note that this analysis was the result of literally minutes work and may well not stand up to close scrutiny!

You will have seen from my Homes of eBay posts that I calculated the percentage of PAS material on sale on eBay in one month and also how long it would take eBay to sell an equivalent of the PAS database for each object type. These are two different versions of the same data. In analysis, I turned each of these into relative values by ascribing a maximum of 100 to the highest values and rating others accordingly. So, in terms of amount of material on eBay, clay pipes score 100 as the amount recorded has the highest percentage in relation to the PAS. In the other column, Roman coins score 100 because it would take longest for eBay to sell an equivalent amount of material. Subtract the second value from the first and you have a series of relative values that show the potential for different types of eBay object to help us understand the relationship between people and historic-archaeological objects.

Representative Network Potential

Representative Network Potential

This graph tells us a different story to the pie charts. Roman coins, by far the most frequently recorded item on the PAS and listed item on eBay, get a Public Representative Value* of -97. In short, there are a lot of this kind of object, but eBay won’t tell you much about the people-object networks of Roman coins because it only holds a tiny sample. On the other hand, clay pipes have the opposite value. There are a relatively small number, but it’s a higher percentage of the PAS and replicates the PAS every 18 months. So, these two quite different datasets become comparable, not in pure numbers or why people use them, but in terms of their potential to tell us about people and their objects.

If you want to know about members of the public who are part of the networks of Roman coin movement, don’t bother with eBay, consult the Portable Antiquities Scheme. If clay pipes or musket balls are your thing and you want to understand how people interact with them, forget the PAS and try to get eBay sellers to talk. I hope you have better luck than I did.

I promise it makes sense. Hopefully I can replicate the experiment at some point with a larger dataset and some people to talk to.

*Sorry, couldn’t resist developing complex terminology.


PA2015: February preview – What happens when you ask people for their consent to be studied?

What happens when you ask people for their consent to be studied?

I’ve always been interested in whether deposition implies consent to be studied, and now it seems more relevant than ever.  I’ve recruited a handful of people, some of whom I know, and others I’ve only connected with online, to look at how our changing notions of privacy are altering what we see as valid material for archaeological study. Well, if they’re altering things at all.

Information and consent

Toward the end of my archaeology career, I had something of a crisis of faith in our ability to piece together anything remotely accurate from a bunch of disparate data points. Having more data didn’t seem to help because data can only tell you what happened, not what someone’s intention was. I also felt sort of…intrusive. Maybe it’s because I worked with relatively late (post-medieval) material that was often attached to an individual who had a name; it wasn’t so anonymous, and I wondered what would happen if it were more recent. If consent were actually possible, would people grant it?

At what point does information become ‘fair game’? Is some stuff just not useful? Is there a difference between “in the interest of posterity” and “of interest to posterity”? That, I thought — and still think — is pretty exciting. Now that everything in the commercial sector is about data, and I’m in a job where I work with it, I want to return to this question.

I want to know what we reveal and conceal about ourselves online and off. I want to know how being watched affects how we describe our behaviour. I want to know what people do and don’t want others to know.

The point of the project isn’t to come up with any kind of conclusions, but to start to identify some of the issues that exist now, and those that will exist in the future, when it comes to storage, guardianship, curatorship, access to, and narratives made with our personal and collective digital data, as a historical archaeological resource. I’ve talked about some of my initial questions here,but now I want to see what else comes up.

In short: do you care if posterity knows that you pooped?

The rough outline

The main task for participants is to choose five things that they do at least three times per week, and record them in some way. It can be anything: brushing their teeth, sending an email, performing a google search, taking their kids to school. I’m asking them to record them, in any way they choose, in any level of depth or detail, or using any (reasonable) medium.

They’ve also filled out a survey that includes questions about how they feel about the privacy of their quotidian data, and they’ll answer the same questions, plus some new ones toward the end of the month, to see if anything has changed.

Finally, happens when you put the privacy controls in the hands of your subjects? Because I want to look at where people’s boundaries are without intruding on them, participants are perfectly welcome to omit anything they’re not comfortable sharing, although I’ve asked if they’ll at least make a note where something has been redacted.


A screen-grab of my browsing history

A screen-grab of my browsing history


And then what?

Up to now, archaeologists have pretty much assumed, with some (ok, many) exceptions, that deposition of material implies consent to be studied. We also know that people who self-historicise behave differently than those who don’t (for example, by creating conscious archives).There are projects like this Cold War one at the Greenham Common site, where including some of the people who lived there helped shed light on otherwise impossible things, but also brought up issues of what is and isn’t an acceptable intrusion.

I’ve always wondered what would happen if we assumed the same level of concern for privacy, even where we couldn’t actually obtain consent. And now, we know that our data is being collected and stored by companies, governments, and all kinds of other institutions. We’ve always seen data collection as a form of surveillance — even when we haven’t called it that — but I don’t think we’ve ever been so collectively concerned about who knows what about us, even if it’s not a secret.

How is the digital world changing how we view what is ‘private’, or even ‘secret’? What new opportunities will arise, and how can we best make the most of them without being unwelcome intruders into people’s lives? Do we need to change our ways?

I’ve landed firmly in the business world now, and part of my job is to be in charge of people’s data, but I consider it part of my duty as a human being to make sure that we are really aware of the kind of resource we have — and what is and isn’t okay to use. Because what is legal is not necessarily ethical.

For me, the implications are twofold: doing business in a way that is more ethical and (maybe) understanding the potential secondary social-historical-archaeological uses of data collected for a business purpose, and also a plain-old issue of archaeological ethics. Because for me, that was always the most interesting part of archaeology: who is really entitled to join the dots?