Emily Lena Jones and David A. Hurley 0000-00-00 00:00:00
Emily Lena Jones is Assistant Professor in the Anthropology Program at Utah State University. David Hurley is the Digital Services Manager for the Albuquerque/Bernalillo County Public Libraries. The developing "digital world" of archaeology holds great promise for expanded communication of and collaboration on archaeological projects. The number of projects that take advantage of digital technologies to share and integrate data is on the rise; the Alexandria Archive Institute (http://www.alexandriaarchive.org/) and Bone Commons (http://www.alexandriaarchive.org/bonecommons/), Open Context (http://opencontext.org/), and the Chaco Research Archive (http://www.chacoarchive.org/cra/) are just a few of the digital archaeology collaboration spaces currently available. Clearly, archaeologists are more focused on increasing data sharing and on integrating data from multiple sources. However, for these digital communication/collaboration projects to fully realize their potential, all archaeologists, whether in the field or in the lab, need the skills that allow them to be effective in the digital world. Aside from the obvious (though always shifting) technical proficiencies, students need to develop conceptual models for understanding archaeological data that facilitate such collaboration. The most fundamental, and useful, such model is the relational model of data (Keller 2009). This model is the basis for the relational databases that are the foundation of virtually all current and emerging datasharing projects. Despite the importance, usefulness, and ubiquity of the relational database, when we informally surveyed our colleagues, we learned that only a few American graduate (or undergraduate) anthropology programs explicitly cover relational databases anywhere in the curriculum. Even more surprisingly, many of the archaeologists we spoke with were using flat files (e.g., Excel spreadsheets) instead of relational databases, and quite a few didn't exactly know what a relational database was. Given the importance of relational databases in digital data- sharing, this is distressing. Our opportunity to engage with this issue came in autumn of 2009. One of us, an assistant professor at Utah State University, was assigned to develop and eventually teach a class in zooarchaeology for graduate and advanced undergraduate students, and decided to use this course to introduce zooarchaeologists- in- training to basic database thinking. The challenge in this situation would be to keep the relational databases as a structural, rather than a content, area of the class; and to assist with this, a librarian with expertise in teaching database skills joined the project. The choice to include a focus on relational databases in a zooarchaeology class may be surprising, but we feel we had a strong case for it. The students in this class were primarily drawn from Utah State's MS in Cultural Resource Management program, which seeks to prepare students for leadership careers in private and government archaeology; or advanced undergraduates in the Anthropology major, which focuses on applied anthropology. Few of these students are likely to pursue zooarchaeology beyond the class; they will, however, encounter faunal assemblages and zooarchaeological data in their archaeological careers. At the same time, this was the first exposure that many of these students had to raw archaeological data, and, therefore, there was an opportunity for them to conceptualize the data relationally from the beginning rather than having to unpack their understanding of the data later on. In other words, by linking the learning curve of thinking relationally to the learning curve of doing zooarchaeology, we hoped both would be made easier. We saw this as a "now or never" scenario for teaching relational database concepts. We thus developed a course with three learning objectives. The first two are typical of zooarchaeology courses: to provide students with basic identification skills (with a special focus on distinguishing human from nonhuman bone, a skill of great importance for any archaeologist) and to introduce them to how zooarchaeological research can inform larger archaeological questions. The third (and much more unusual) objective was that "students will understand relational databases as a means of organizing zooarchaeological data". Implementing the Objectives In order to best teach relational database thinking without shorting the other learning objectives, we decided to use relational databases as a framework for the class, rather than as a content area per se. We started with an orientation to relational data, and then used this as an organizational framework for the rest of the class content. This allowed us to talk about databases without significantly impacting the class time devoted to the more traditional class goals. The power of a relational database lies in the relational structure of the data–that is, in explicitly defining the relationships between discrete data elements. Compared to a flat file, where all the data is treated as properties of the individual specimen (a very simplified example can be seen in Figure 1), the relational model offers greater analytical flexibility by preserving multiple access points (in other words, data can be analyzed based on any property) while ensuring data integrity (Keller 2009 discusses this in more detail). The orientation to databases therefore focused on helping students understand the types of relationships that are made explicit in the relational model. We made the orientation heavily visual, and integrated it with a general discussion of different types of zooarchaeological data. For instance, Figure 2 shows an example of a one- to- many relationship. An individual specimen can, logically, only have one element identification. It is either a tibia, or some other element. One specimen can only have one element. An element identification, however, can logically be applied to more than one specimen. There can be more than one specimen identified as tibiae in a given collection. Hence, the relationship between specimen and element is one- to- many. Other times, the relationship is what is called many- to- many. For example, a specimen can have multiple modifications: it can be burned, weathered, and have multiple types of cutmarks. Likewise, all these modifications can logically (and likely will be) applied to multiple specimens (Figure 3). Hence, the relationship between specimen and modification is many- to- many. We started the class off with this visual recap, focusing on how individual relationships (as seen in Figures 2 and 3) contribute to a complex whole. This orientation, when we talked about the fundamental relationships, was the only class time devoted solely to database information. After this first class, all discussion of databases was embedded in a discussion of zooarchaeological data. As we moved through the traditional class goals, every time we discussed a new kind of zooarchaeological data, we included in our discussion an exploration of how this data related to other zooarchaeological data. In this way, during the course of the class, we built a hypothetical zooarchaeological database. The examples we provide in Figures 2 and 3 are rather clear cut, but often understanding relationships between data elements (and even what constitutes a discrete data element) requires analytical effort. Engaging in this effort provided students a deeper understanding of the data, while also resulting in a dataset that can be analyzed with a broader set of tools, notably Structured Query Language (SQL) and GIS. Approaching zooarchaeological data this way would, we hoped, teach students to embed relationships in their data collection from the outset, rather than the more limited, but still more common, flat file approach, which treats all data as the property of the individual zooarchaeological specimen. In addition to quizzes (which involved both specimen identification and questions about database relationships), seminar readings, and the hypothetical database, students also completed a small analysis project, including the creation of a small relational database. The goal here was to have the students put all three learning objectives, in a limited way, into practice. Results From a student satisfaction perspective, the database experiment was a success. Student evaluations provided no negative feedback about the database portion of the class, and 13 of 15 students had something positive to say about it. This was a surprise, given that this was the first incarnation of a brand- new class, and given that the project portion of the class was far from perfect. Assessing what students actually learned about databases was a more difficult task. The database that the students produced was only moderately successful as a class project, perhaps not surprisingly. Students were overwhelmed by this task, and eventually one technologically savvy student took over database creation; the final product is thus not an accurate way to assess students' learning on databases. In retrospect, we think having students create their own database in the course of the class was too much to ask, and in future incarnations of the class we will provide them with a premade database to work with. We don't feel that this will take away from the database learning objective, given that our goal was not to teach students to create databases, but rather to introduce them to the fundamentals of database thinking. When student learning is assessed in these terms–that is, did students learn relational database thinking–our results are much better, although more data would be useful. Quiz performances suggest that students, at least during class, learned the basic concepts of relational databases; even students who struggled with identification were able to master database relationships. In the final quiz, students averaged 89 percent on the database relationship questions. We are curious, however, about the persistence of their mastery. Will students who took this zooarchaeology class, for example, perform better in the GIS class (which relies heavily on databases, although data relationships aren't explicitly taught)? We look forward to pursuing this question further. Conclusions We are convinced that relational database thinking both needs to be, and can effectively be taught, somewhere in the graduate curriculum–and probably in the undergraduate curriculum as well. However, the particular method for integrating databases into an analysis class presented here was, of course, designed for our situation at Utah State University. While we are pleased with our initial results, and plan on continuing to integrate databases into our zooarchaeology class, this would not necessarily be a desirable curriculum in other programs. While our experience suggests that zooarchaeology classes may be a surprisingly good fit for this–we were able to teach a basic introduction to relational database thinking without impinging on the straight zooarchaeological content–there is no reason why such teaching couldn't easily be incorporated into other archaeology classes instead of or in addition to zooarchaeology. There is, however, a big caveat. There is a difference between teaching database creation and teaching database thinking. The students in this class did not learn how to program a database, and trying to create a database was, from the instructor's standpoint, detrimental to the class. Next time this class is taught, we plan to provide them with the basic database for their project. What the students did learn was relational database thinking–the big picture concepts. Database relations, we found, provided an easy and effective way to organize the straight zooarchaeological material in the class, and the students now have the groundwork set to become relational database users, rather than relying on flat files. We thus hope that these students are more prepared to be archaeologists in a digital age. Acknowledgments. Thanks to everyone who provided syllabi and/or had conversations about goals for zooarchaeology classes and how to teach databases (including but not limited to Madonna Moss, Virginia Butler, Don Grayson, Barnet Pavao- Zuckerman, Jenny Waters, Bruce Huckell, Jean Hudson, Sarah Neusius, Natalie Munro, Levant Atici, and Christyann Darwent). And most importantly, we thank the students of Anth 5320/6320 in Spring 2010, who went along with this experiment and shared their opinions on it! References Cited Keller, A. H. 2009 In Defense of the Database. The SAA Archaeological Record 9(5)26–32.
Published by Society for American Archaeology. View All Articles.