Paulina F. Przystupa, Ph.D., Alexandria Archive Institute
L. Meghan Dennis, Ph.D., Alexandria Archive Institute
In this case study, we introduce archaeological data literacy by drawing from a data set titled Neolithic and Bronze Age Cattle Data from Switzerland collected by Elizabeth Wright and colleagues (Wright et al 2021). This data set aggregates previously published data from a variety of sources, and focuses on data about cattle during the Neolithic and Bronze ages. Wright et al (2021) define the Neolithic and the Bronze ages as between the years of about 4500 and 800 cal BC. They divide the data into specific sub-temporal units that exist within each epoch. The data mainly come from archaeological sites in present-day Switzerland but also from other parts of West-Central Europe, such as Germany and Liechtenstein.
As explored in Wright (2021), creating a data set like this is complicated but worthwhile to address trends such as regional adaptations and changes to human behavior. Specifically, this data set supports research focusing on how people exploited domesticated and wild cattle during the Neolithic, by leveraging bone measurements and morphological differences to separate wild and domesticated cattle. Another use for this data set can be to explore how people in that region changed how they used cattle over time. Changes in use demonstrated significant relationships with adjacent regions, which shows that during the Neolithic, groups of people didn’t necessarily live in isolation from one another. They had relationships with neighbors hundreds of miles away!
These are the sorts of things that you can learn and do if you have good archaeological data literacy. But what does it mean to have archaeological data literacy? Well, archaeological data literacy is when a person knows how to read, work with, analyze, argue, and communicate utilizing archaeological data. Whenever we’re learning something new in archaeology, either on our own or in a classroom setting, we’re engaging with that new information in a variety of ways. It could be that we’re reading an article on a new topic, or we’re looking at a graph from a publication, or we’re replying to a post on social media. Or, we might be tossed a spreadsheet full of numbers and words we’re not familiar with. When presented with new information we can react in a bunch of different ways.
Some people feel comfortable looking at the numbers as numbers, playing with them using a spreadsheet program, or using a programming language like R to examine trends. In contrast, others will want to read more articles to contextualize those numbers before doing anything. Those are just a few ways to react to the data but any response requires a certain level of established data literacy. Folks who have that kind of data literacy already understand different kinds of numbers, and feel comfortable exploring questions or finding keywords that they are familiar with. They might not need much guidance to get started with the data. If that’s you, great! If it’s not you, it’s okay! Not everyone is comfortable jumping right into spreadsheet or code-based analysis. Some folks might be afraid that they’ll break the data, or make the wrong conclusions because they don’t know enough about the topic. That’s okay, too! These fears are understandable and common, but they can be surmounted!
Either way, if you’re interested in playing with the cattle data on your own, we have a full exercise as a PDF to guide you through it. Those who enjoy coding can explore the same resource in R markdown as part of a larger data literacy project. The goal of that exercise is to acclimatize you to making mistakes, messing up, and the idea that there’s no way to break the data! Anxieties you have about numbers, spreadsheets, and data, are normal. They’re a part of advancing your data literacy. In this text though (the one you’re already reading), we’re going to use examples from this data set to explore what it’s like to learn to ‘read’ data.
If you’ve gotten this far you’re probably heaving a sigh of relief, knowing that data literacy isn’t only about looking at numbers. A lot of the data that we engage with in archaeology has a numerical component, but the numbers aren’t always numbers… In Kansa and Kansa (2013), Open Archaeology: We All Know That a 14 Is a Sheep: Data Publication and Professionalism in Archaeological Communication, they show how the numbers that we see in a table may mean more than what they appear. The numbers might be keys that link to data in another table (such as linking bones to measurements taken on them), or they might be codes used as shorthand that are translated into text in another document (i.e. “14” = “sheep”). This highlights how there are numbers or data of different kinds and they cannot always be treated as the same! And this is where the data literacy component of “reading” both draws from what you know and extends it a little further.
To dig into this a bit deeper, there are different kinds of numbers that may be categorized or understood by the acronym NOIR. This is short for nominal, ordinal, interval, and ratio. These different kinds of numbers limit how we can explore the data to examine trends, or apply statistical tests. Let’s quickly go through what each of these kinds of numbers are, and how you might use them. Keep in mind, each of these types of numbers is actually a type of data.
- Nominal numbers, or nominal data, are data which can be organized into categories. In an archaeological analysis of our cattle bones, this could be by sex if we want to separate bulls from cows, or by location if we want to put all of the bones from Liechtenstein together. It’s important to see that nominal data are things that can’t be divided numerically, only categorically. For example, a “1” for “male”, a “2” for “female”, and a “3” for “unknown sex” serve to separate these two categories, not rank them or allow them to be tallied.
- Ordinal numbers, or ordinal data, are data which can be put into a ranking, or a list based on an order. We could order cattle bones by weight from lightest to heaviest, or by their length from shortest to longest, or by their excavation date from most recently to longest ago. When working with ordinal data, we’re not concerned with how much difference there is between our data points, just that there is a difference that we can rank.
- Interval numbers, or interval data, are data that can be categorized and ranked in equal intervals, and where the differences between numbers mean something but there’s no true zero. For the cattle bones, a bone’s deposition date is an example of this, though…it’s a little complicated. A bone from Wright et al (2021) dates to 1600 BCE. What that actually means is that someone or thing left it there 1600 years Before the Common Era (BCE) with the common era event being our zero. But if we want to know how long ago 1600 BCE was from today, today being in the year 2023, we’d need to add the two, 1600+2023, to get 3623 years ago. So while the units are the same (equal intervals of one year) the amount (number of years in the range) changes depending on when or where you measure from.
- Ratio numbers, or ratio data, are data that have consistent units occurring at equal intervals, can be ranked, and have a ‘true zero’ (an absolute lower limit that you can’t go below zero). Continuing with our cattle data, a bone’s weight, or length has an absolute zero because you can’t have a negative weight or length for something that exists! This is where the measurement comes in: while the units might drive the change in number, the actual value measured will not change no matter what units you record in. An existing cow bone will have a fixed length and fixed weight in whatever units you pick. It’s very informative to use ratio data when working with bones, because since ratio data is never negative, you can use it to calculate frequencies (how often you find a kind of bone), and standard deviation (what is ‘normal’ for a given kind of bone).
Understanding the kind of numbers (or data) you’re working with is important to figure out how you can analyze them. In the data story we linked to near the beginning of this text, we have folks focus on terms related to time. “When?” is an important question in archaeology and matters regardless of geographical locations and kind of excavation. And to explore that question, based on what we learned above, will probably require us to consider either ordinal or interval data. You can explore both ways of understanding time, as ordinal or interval, in the Cow-culating Your Data exercise by using a spreadsheet program or by using the statistical program R.
However, as we’ve said before, there are more ways to engage with data than by using numbers. Another way to explore and cultivate data literacy skills is through considering your data as narrative, practicing the arguing and the communicating components of data literacy. What stories does your data tell? How can you think and share the past through your data in a way that other folks, especially non-archaeologists, can relate to? Data may be collected in numbers, but it can be expressed in words. An important part of growing your data literacy is the ability to share your results in clear, concise language that engages your reader.
You can do this by considering different kinds of voices you may use to tell the story of a data set. Take a look at Neolithic and Bronze Age Cattle Data from Switzerland by clicking on the Data Records hyperlink and selecting a data type. Play around with the settings of the various images on the page, moving sliders, filtering and clicking buttons, practicing your “reading” data skill. Then consider how you might tell someone else about the things you observe in the data set. Consider how the way you tell someone might change based on their age or cultural background? How would you change the way you spoke if you were telling a professor versus your elementary school teacher? These are all people who might be interested in knowing about these bones but their interest in the story may vary depending on what pieces you highlight and how you tell them about it.
Data literacy is a process of continual growth and learning. Throughout a career in archaeology, your archaeological data literacy will grow, and your ability to synthesize information will increase. Learning how to approach data is a lifelong process, and one that will, ultimately, make you a better archaeologist!
About the Authors
Paulina F. Przystupa, PhD is a Filipine-Polish-Canadian-American Postdoctoral Researcher in Archaeological Data Literacy at the Alexandria Archive Institute / Open Context. Trained as a historical landscape archaeologist exploring socialization and assimilation at children’s institutions, Paulina works in pedagogy-focused digital archaeology contexts to understand the dynamic ways people learn and teach archaeology and the ways that education shapes national and community identities.
L. Meghan Dennis, Ph.D., Alexandria Archive Institute
References
Kansa, E. C., & Kansa, S. W. (2013). Open Archaeology: We All Know That a 14 Is a Sheep: Data Publication and Professionalism in Archaeological Communication. Journal of Eastern Mediterranean Archaeology & Heritage Studies, 1(1), 88-97. DOI: https://doi.org/10.5325/jeasmedarcherstu.1.1.0088
Schibler, J. (2017). Zooarchaeological results from Neolithic and Bronze Age wetland and dryland sites in the Central Alpine Foreland: Economic, ecologic, and taphonomic relevance. In U. Albarella, M. Rizzetto, H. Russ, K. Vickers, & S. Viner-Daniels (Eds.), The Oxford Handbook of Zooarchaeology. Oxford University Press. DOI: https://doi.org/10.1093/oxfordhb/9780199686476.013.6
Wright, E. (2021). Investigating cattle husbandry in the Swiss Neolithic using different scales of temporal precision: Potential early evidence for deliberate livestock “improvement” in Europe. Archaeological and Anthropological Sciences 13(36), DOI: https://doi.org/10.1007/s12520-020-01252-6
Wright, E., Schäfer, M., Stopp, B., Marti-Grädel, E., Ginella, F., Kerdy, M., Bopp-Ito, M., Deschler-Erb, S., & Schibler, J. (2021). Neolithic and Bronze Age cattle data from Switzerland. Open Context. <http://opencontext.org/projects/c89e6a9e-105a-4368-9e90-26940d7bf37a> DOI: https://doi.org/10.6078/M7H13049
