Pedagogy First

A Lab-Led Model for Supporting DH Instruction on Campus

Founded in 2015, the DHTF program was created with the two goals of supporting the integration of DH into the undergraduate curriculum and providing semester-long mentorship for graduate students looking to learn and teach DH. Originally intended to support one graduate student per semester, the program quickly adopted a cohort model (typically between three and four students a semester) beginning in its second year. In addition to increasing the number of students the DHLab could mentor, this approach allowed us to create a peer community where students might learn from one another as well as from lab staff.5 The DHTF program follows a train-the-trainer model, where lab staff teach graduate student instructors, providing both technical and pedagogical training, so that the latter, in turn, feel confident teaching their students. The program has trained over thirty-five graduate students who collectively represent seventeen departments from across the humanities. The reach of the program has extended beyond the humanities as well, with a religious studies PhD student serving as a DHTF for a computer science class on DH applications, and a physics PhD student serving as a DHTF for an English class on science fiction and technology. The multidisciplinary dimension of the program gives students glimpses into how disciplines outside their own are engaging with DH. It also reaffirms the benefit to placing DH pedagogical instruction within a lab—it would have been a tall order for the seventeen departments to each stand up its own program with a DH mentor and peer community.

All Yale graduate students in the humanities or with a teaching appointment in the humanities are eligible to apply. In order to maximize the number of students who can participate, no prior experience in DH is required, and the class itself does not have to be a “DH class,” meaning DH topics and tools do not have to be a primary (or even secondary) focus. As Cordell argues, “You do not need an entire DH curriculum, or even a designated DH course, to introduce substantial digital pedagogy into your classes.”6 Strongly believing this to be true and recognizing that students’ control over their teaching varies by their department and appointment, we work with students on a class-by-class basis to determine what would be effective and realistic. Whereas some DHTFs serve as the instructor of record for a seminar class they designed, others are teaching assistants assigned to one or more sections for a lecture class led by a faculty instructor who might be open to DH but not a practitioner of it. In these cases, the program additionally functions as an introduction to the possibilities of DH for faculty, some of whom have since reached out to the lab to think about strategies for incorporating DH into their other classes.

In order to remain responsive to changing DHTF interests, class needs, technologies, and other time-intensive obligations the fellows and lab staff have, the program has few set requirements. At a minimum, fellows must participate in periodic cohort meetings where we plan class activities and discuss readings, projects, and methods that might be relevant to their specific courses. Learning objectives and course-specific constraints come first, specific methods and software second. We think through scaffolding. Would their goals be better served by introducing DH at a high level once or twice or through sustained engagement with a particular tool, technique, or theory? Would class time be better spent preparing students to create their own datasets or to work with prepared texts? The latter allows more time for fleshing out a method and its built-in assumptions, but the former provides the opportunity for students to build a dataset that is meaningful to them. What are the potential tradeoffs (privacy, cost, disk space, ease of access) to consider when deciding whether to teach with software that lives locally versus in the cloud? These cohort meetings also provide opportunities for fellows to share how their classes are going, receive help with any issues they might be facing, and get inspiration from their peers’ approaches.

Along with cohort meetings, fellows must also participate in at least one technical workshop that is geared toward learning how to teach a method or tool within their subject areas. Fellows are encouraged to (and almost always do) attend one another’s training sessions. Here is where the cohort model again offers advantages. If one fellow is interested in incorporating geospatial methods into their teaching, the other fellows can also develop expertise in this area, even if it does not apply to their immediate classes. While these sessions usually involve teaching a digital tool, I take care to avoid what John E. Russell and Merinda Kaye Hensley describe as “buttonology”—“software training that surveys different features of an interface in an introductory manner” without calling attention to larger research contexts.7 As Paul Fyfe writes, “The tools are easy. What is hard is imagining how to use them and, harder still, imagining the social conditions they might enable and, hardest of all, creating the institutional structures in which they will flourish.”8 When I teach a method or tool, I include metacommentary along the way—I mention how I selected the example use cases, tools, and datasets, and how institutional licenses factored into that selection. I point out why I ordered the steps in the tutorial the way that I did, and the kinds of questions I tried to anticipate or encourage. I ask the fellows to think through what their own approaches to teaching the method or tool might be given the context of their class or discipline.

For an example of a typical training session, here is an activity I use with fellows who are interested in creating maps from literary texts with the goal of prompting their students to think differently about how place functions in novels or poems. In preparation for the session, I ask the fellows to read a few “geographically rich” passages from a text in their class and add the locations they come across to the first column of a spreadsheet template I share in Google Drive. This template has only three columns to begin with: “Name,” “Latitude,” and “Longitude.” In the session, I show them that many locations can be found in Google Maps by looking at the URL after you have searched for a place by name. It is not long before problem cases arise—what do we do with references to the moon or locations that cannot be searched by name in Google Maps? How can we account for historical locations that have changed names, or imprecise references that we can estimate but not pinpoint? We talk through strategies for capturing this information, adding new columns to the spreadsheet as we go. We discuss why starting this work by hand is generally a good choice even if they intend to automate the process later, because the experience points out peculiarities in their datasets that might thwart geocoders—automation does not always save time. Once we have our populated spreadsheet, we turn to mapping software. More often than not, I will introduce ArcGIS Online because it is web based, it is relatively user friendly, it does not require custom programming, and Yale has an Enterprise account, prompting discussions about access and sustainability. After loading the dataset, we consider which basemap makes sense given the questions we want to investigate—satellite imagery is more helpful than a generic gray background for showing greenspaces and roads. If we are working with historical data, we might have to add a georeferenced layer. If so, we make plans to cover that in a follow-up workshop, underlining the need to build in ample class time for scaffolding the different components of a mapping project. What aspects are most important for students to learn and why? Is it how to create a spatial dataset or how to visualize it? Do students need to know how to georeference a map, or could the fellow provide one that is ready to go?

Not all fellows teach software during their fellowship. Sometimes, their DH instruction centers on dataset creation and demonstrating the potential utility of DH techniques. For example, for an English class titled the American Counterculture: Systems, Spirits, Science, one goal was to provide students with a sense for how major counterculture figures in the 1960s and 1970s were connected. This objective suggested network analysis could be a useful approach to introduce. The harder question was how best to do so given class constraints. The class was a large lecture course that consisted of several sections, each of which had its own traditional teaching assistant assigned to it. Kimberly Quiogue Andrews, the DHTF for the course, provided DH support across the sections. Because it was not possible for the DHTF to provide hands-on training to all students in the class, we decided to incorporate network graphs at a high level: what are they, why might they be useful, what kinds of data do they require.

Andrews collaborated with DHLab staff to imagine two network graphs that would help highlight the breadth of the counterculture movement: one would link all of the individuals represented on the class syllabus and another would add links to those who were left off the syllabus but were, nonetheless, part of the movement. For the technical workshop, we discussed how to create node and edge lists by hand. How should we define an edge, and what are the implications of that decision? What information about the individuals and their relationships could be visualized in a network graph, and what would be left out? We went over the mechanics of Gephi and the effects different layouts, color schemes, and centrality measures have on how we interpret a graph, and we considered how to explain graphs conceptually to students.

For the second graph, Andrews created an early writing assignment that asked students to research figures who were connected in some way (personal friendships, professional relationships, etc.) to one or more of the people they were reading about in the class. This information was collectively gathered using Google spreadsheets. Because Andrews would be generating the network in Gephi herself, she was able to focus her time with students on the less glamorous but crucial data creation and evaluation processes, the bedrock to all DH projects. The student-created dataset led to a network that depicted a more expansive view of the American counterculture movement than could be captured in a syllabus. It also yielded deeper discussions about the contours of the movement. In a final reflection she wrote for the DHTF program, Andrews explained that the data input by students resulted in “a network where the strength of the ties between figures was wildly inconsistent.” Far from making the experiment a failure, however, she maintained that the inconsistencies served as “an accidental boon for the course, as we then spent time productively discussing what ‘counts’ as a connection between two people, and how we think about cultural formations (and syllabi) through an interpersonal lens. Ultimately, these discussions added a dimension to the course that would have been nearly unthinkable without the use of networking software.”9

Students’ newfound interest in networks carried forward in unexpected but welcome ways as the class continued. For the final project, which was open-ended as to what methodologies might be employed, one student wondered whether networks could illuminate David Foster Wallace’s Infinite Jest. To explore this possibility, he created an analog network using red and yellow yarn. The written text that accompanied his network walked readers through a deft analysis of the subcommunities within the novel, describing along the way what the transformation from novel into network reinforced or helped him see afresh. In his reflection on the process, the student mentioned creating the network in the common room of his dormitory, where he discussed the project with fellow undergraduates who were curious about the network and what it represented. This ripple effect also occurred among the other teaching assistants in the course who worked alongside the DHTF; one later sought out the lab’s support in order to add a DH component to her dissertation. In addition to providing graduate students with guided opportunities to learn and teach DH (which in turn gives their students exposure to DH), locating pedagogical training within labs also increases awareness of the labs themselves. Labs have much to offer and gain by expanding to become spaces where students can purposefully and mindfully develop their DH pedagogy.

Broad Challenges, Local Responses

What if the capacity-building work of DH labs were extended beyond technical skills training for projects to include pedagogical support as well? Labs are already hubs for formal and informal instruction through workshops, consultations, pair programming, and guest lectures. In addition to providing instruction on tools and methods, we could also advise on how to teach those tools and methods, increasing the number of people on campus who could teach DH. A lab-led pedagogical program also lowers barriers for getting involved in DH. Students generally have teaching appointments to fulfill; turning those appointments into moments where they can also learn digital skills with the support of a dedicated mentor and cohort provides an entry point into learning DH that does not require travel or a teaching-free semester. It also does not require students to have a DH faculty member in their department.

Empowering graduate students to teach DH feeds into other dimensions of their work. From a research perspective, DHTFs acquire—through teaching—a better understanding of the applicability of DH to their respective fields. From a technical standpoint, they have increased their familiarity with particular software and data management strategies. From a communications outlook, they have gained experience explaining technical concepts in ways that are legible to broader populations. All of these skills prepare students for positions within academia but also, importantly, for careers outside it.

Notes

1. 1. Fraistat, “Data First.”

2. 2. Goldstone, “Teaching Quantitative Methods,” 210.

3. 3. Cordell, “How Not,” 471, 472.

4. 4. See Stommel, “Dear higher.” Stommel’s series of Twitter polls in 2018 suggests a range in pedagogical training, with 51.9 percent of the respondents indicating that they received “basically nothing” through their programs, 28.7 percent stating they have had a “semester-long seminar,” 12.7 percent stating they have had “2+ seminars / certificate,” and 6.7 percent indicating that they participated in “week-long intensive” training.

5. 5. When it came to restructuring the program, DHLab staff benefited from early conversations with Yale’s Poorvu Center for Teaching and Learning, which has a cohort method for their teaching fellows as well.

6. 6. Cordell, “How Not,” 466.

7. 7. Russell and Hensley, “Beyond Buttonology,” 588.

8. 8. Fyfe, “Digital Pedagogy Unplugged,” paragraph 7.

9. 9. Andrews, “Reflections.”

Bibliography

1. Andrews, Kimberly Quiogue. “Reflections on My DH Fellowship.” January 11, 2017. In possession of the Yale University Digital Humanities Laboratory.
2. Cordell, Ryan. “How Not to Teach Digital Humanities.” In Debates in the Digital Humanities 2016, edited by Matthew K. Gold and Lauren F. Klein, 459–74. Minneapolis: University of Minnesota Press, 2016. https://dhdebates.gc.cuny.edu.
3. Fraistat, Neil. “Data First: Remodeling the Digital Humanities Center.” In Debates in the Digital Humanities 2019, edited by Matthew K. Gold and Lauren F. Klein, 83–85. Minneapolis: University of Minnesota Press, 2019. https://dhdebates.gc.cuny.edu.
4. Fyfe, Paul. “Digital Pedagogy Unplugged.” Digital Humanities Quarterly 5, no. 3 (2011). http://www.digitalhumanities.org/dhq/.
5. Goldstone, Andrew. “Teaching Quantitative Methods: What Makes It Hard (in Literary Studies).” In Debates in the Digital Humanities 2019, edited by Matthew K. Gold and Lauren F. Klein, 209–23. Minneapolis: University of Minnesota Press, 2019. https://dhdebates.gc.cuny.edu.
6. Russell, John E., and Merinda Kaye Hensley. “Beyond Buttonology: Digital Humanities, Digital Pedagogy, and the ACRL Framework.” College and Research Libraries News 78, no. 11 (2017): 588–91, 600. https://doi.org/10.5860/crln.78.11.588.
7. Stommel, Jesse (@jessifer). “Dear higher education teachers, a poll. Answer below, reply with stories, and pass along. How much training in teaching or pedagogy was/is included in your graduate program?” Twitter, May 15, 2018. https://twitter.com/Jessifer/status/996355913756893184.