Recently three of my students from the Fall semester of 2018 presented at the Writing and Research Conference hosted by the University Writing program at George Washington University. This conference is designed to showcase the research and writing that first-year students do in their UW courses with current first-year writing students, librarians, and faculty. I asked three students from our class to come showcase their work as a panel. Even though they all had very different research topics, they all used Twitter datasets for their projects. I introduced the idea of studying Twitter data to my students last term based on my own dissertation research. In 2015 I connected with the librarians at Gelman Library (GWU) after sharing my project idea in a lightening round at a THATCamp to study the entire Twitter feed of Salman Rushdie. The librarians approached me with exciting news – they had developed a tool called Social Feed Manager that could help me. Up until then I had been simply screen-capturing Rushdie’s feed, but as you can imagine this proved to be an unwieldy way to go about capturing thousands of tweets. The Social Feed Manager (SFM), on the other hand, would be able to harvest the available tweets in Rushdie’s feed and then continue to harvest tweets as long as I had the “seed” active in the program. The SFM works with the Twitter API and, as such, is subject to certain rules and limitations, but the librarians were enthusiastic about my research and helped me navigate the platform. I could then export the tweets and available metadata into a spreadsheet (Fig. 1). This was vastly preferable to my screen-capture project because it was searchable and because of all the data it provided in addition to the content of each tweet. After completing my analysis of the data, I published it in the Journal of Commonwealth Literature as “‘The Man Who Would Be Popular’: An Analysis of Salman Rushdie’s Twitter Feed.”
It wasn’t until the summer of 2018 that I reconnected with the Gelman librarians to ask them about their process in creating the SFM and to inquire about the possibility of using it in a first-year writing course. In thinking about how the students might interact with the SFM, I had to keep in mind that I was teaching three sections of 17 students. Initially I pictured grabbing a dataset and having them do a small-group or class-wide hypothetical research proposal based on the data that is provided in the spreadsheet. They could develop potential projects based on, for example, the favorite count column which shows the favorite count of each tweet at the time it was harvested to analyze what kinds of tweets garner the most likes. Or the hashtag column to analyze the relationship between using certain hashtags and favorite counts. Or the tweet type column to analyze a user’s overall pattern of Twitter use to see if the user creates more original content or mainly retweets existing content.
After speaking with the librarians, however, I expanded my thinking about how the students might interact with the SFM more fully. They were willing to come to our classes one Friday to talk with the students about the tool and to teach them how to use it to generate different datasets. We also discussed the possibility of students using existing Twitter datasets if they wanted to bypass using the SFM for this project. They could do this through another Gelman Library platform, Tweetsets. Tweetsets are broad datasets that the students could filter using specific criteria based on their research ideas. The librarians assured me that handling 51 students for these projects would not be an issue, and so I expanded the initial project idea to allow students to use either the SFM or Tweetsets to collect a dataset with which they would produce a formal research proposal.
Because I wanted the students to be ambitious in their research ideas, I decided the proposal would be the best option so that they wouldn’t feel restricted by having to actually carry out the research on the dataset. They did, of course, have to do some preliminary research on their chosen dataset to figure out how they would approach the dataset in its entirety for the hypothetical research project. We discussed qualitative and quantitative research methods. I used my own work with the Rushdie dataset as an example. While the proposal writing proved to be challenging for them, the hardest part initially was creating a viable dataset. If they chose to work with the SFM, they had to work within the restrictions of the Twitter API which dictates how many “past” tweets can be collected from a given feed, for example. They also had to work within the rules of the SFM itself which could not, for example, collect tweets for a specific hashtag going back in time –only going forward. Many who chose Tweetsets struggled with the filters in order to produce a narrow enough dataset that it could be exported into a spreadsheet. We quickly discovered the maximum number of rows you could have in a single Excel file. We also realized that Tweetsets could only handle so many users at once before it broke. For some students the technical hurdles were much harder to clear than for others. It bears repeating, but just because we see students using smartphones like limbs doesn’t mean they are digital natives let alone technology-savvy. It was an exercise in persistence and patience—a potentially valuable byproduct that I hadn’t anticipated when I designed the assignment.
Seeing my students present at the conference last week, you wouldn’t guess that we struggled as much as we did. Their presentations focused on developing the research idea, on the methodologies they developed, and how to use writing to convince someone to fund your research. They spoke with authority about their datasets and collection methods—an element that I later learned intimidated my current students some of whom were present at this panel. I just beamed with pride the entire time. I had, of course, read their proposals already and knew how good they were, but their presentations were excellent composition products in their own right. You can read their project abstracts here (just search the program for “Azar”). Along with their classmates, these students took my loosely-formed idea for using Twitter datasets in the classroom and surprised me with their creativity and rigorous standards. One student proposed a study that would trace rhetorical patterns in how certain politicians tweeted about climate change. One student’s project seeks to help NASA amend its Twitter use to garner more public support for its programs. Another student’s project looks at how leading news media Twitter accounts shape public perceptions of social media. Others looked at tweets related to the women’s march, to the #metoo movement, and hurricane Harvey.
The good news is that the SFM is open source, and so potentially available at a library near you. Even if it is not possible for your institution to develop the SFM, I hope this post inspires you to get in touch with your librarians to see how students might be able to use the library resources more fully. It could be a nice supplement to what you already have in play or, like me, it could fundamentally alter your course design.
Some might question my decision to have first-year writing students use a tool that up until now has been used mainly for senior projects and graduate/faculty work. Or why I had them develop original research ideas and learn proposal writing since these are more often skills taught to upperclassmen. The truth is that I was bored of assigning traditional research papers in first-year writing. The longer I teach, the less it makes sense to have the first-year writing capstone project be regurgitating bits of scholarly articles back to me in highly predictable patterns. What exactly is the real-world (outside of the classroom) genre equivalent to this type of writing? A literature review? An executive summary? A journal article? Then why not just teach those genres instead of the pale imitation that the first-year research paper seems to be. Why not have first-year students start thinking about how to do proposal style writing which they can use in a wide variety of ways even as undergraduates (scholarship, internship, transfer, and job application materials to name just a few)? Why not focus on helping them locate research gaps and develop methods to contribute to different areas of knowledge? We live in an age of big data—most of which has yet to be analyzed. Never has teaching students how to think critically about knowledge production/consumption and how to analyze data (in an ethical, socially responsible way) been more important. It is time to shake things up, and if my students are any indication, they are more than up for the challenge.
If you’re interested in hearing more on alternatives to the first-year writing research paper, come see me present at the NeMLA Conference next week on “Replacing the Composition Research Paper with Wikipedia Writing.” I’ve also written about teaching with Wikipedia if you’re interested in learning more.