Although writing centers have not always made the most—or best—uses of data visualization, our work demonstrates the important role it can play in our research and our centers. In this study, we found that there are multiple uses for data visualization ranging from the exploratory to the confirmatory. Similarly, there are different kinds of audiences and datavis tool users ranging from expert to novice. It is our hope that researchers in writing studies start to explore and define the different rhetorical functions that data visualizations can play in our scholarship, assessment requirements, and administrative work. We also hope that audience and use feature into considerations when determining what visualizations best suit specific rhetorical situations.
In this webtext, we created four pages that function as a heuristic for data visualization in writing and writing center studies. The first page identifies how tables assist in hypothesis formation. Tables also represent complex numeric outcomes resulting from quantitative analysis, such as frequency distribution; therefore, tables can be useful visualizations for developing arguments about findings as much as they can be useful visualizations for conveying findings to readers. The second page, about word clouds, identifies different purposes for data visualization—namely, the importance of taking into consideration the rhetorical context when selecting particular data visualization models. For example, although word clouds may be aesthetically appealing, they ultimately lack the quantitative rigor of other forms of data visualization; therefore, they are better suited to situations where affect is more important than significance. This section calls for a widening of data visualization heuristics for writing studies to include both amateur and expert users and tools. Although word clouds might be less precise than charts (in terms of showing frequencies or relationships between phrases), they also impact audiences viscerally because their aesthetic features can be examined through visual rhetoric, such as font, color, size, and layout. Word clouds, then, require more restraint on the part of researchers, and, perhaps, more use in pedagogical and administrative (e.g., trainings and workshops) settings. The third page, about word trees, traces the lineage of this particular visualization back to the evolutionary tree as well as to the language tree. Word trees, as we note, show preceding and following continuations for the word, thus allowing words to be viewed in their contexts. Although word trees are exploratory, they are also interactive, serving a dual purpose to form hypotheses as well as to explain information. The fourth page utilizes the Collocates graph in corpus analysis. The Collocates graph shows key terms, with high frequencies of occurrence in context. We argue that quantifying and locating key terms within common contexts allows for an inductive method of analysis. The Collocates graph, in other words, allows researchers to draw conclusions about language in context. Visualizations, then, are deeply embedded in the research question, but are not necessarily determined prior to being created. For us, the kind of visualization influences how research questions are formed and tested.
In addition, as Dana Driscoll and Sherry Wynn Perdue (2014) concluded in “RAD Research as a Framework for Writing Center Inquiry,” writing center scholarship has frequently focused on the uniqueness of writing center contexts; therefore, many of the research studies lack replicable methods and generalizable results. One way of moving toward replicable research is for writing center professionals to participate in multi-institutional research projects, examining data collectively, as we have done by using a corpus containing session notes from all of our institutions. Although the fields of science, technology, and business have a long history of participating in multi-institutional—or multi-party—research (Corley, Boardman, & Bozeman, 2006), writing centers do not, but there are many benefits to it, including the chance to pool expertise and resources, produce larger data sets, and produce findings that are not limited to one institution. We found many benefits to multi-institutional research, including creating a larger corpus than one institution was capable of producing, moving beyond the idea that each writing center is “unique” in order to make arguments about the field as a whole, and learning from each other, as our team comprises early and middle career administrators and scholars. Additionally, this process has helped transfer knowledge between institutional researchers and helped us to consider, and reflect on, our own practices.
Through our research we have found one area ripe for exploration. Part of the impetus for publishing this piece, which is interactive and digital, in Kairos is the value of and need for more dynamic data visualizations. Although Voyant offers many adjustable tools—the word cloud, for example, can include a range of words that varies from 25 to 500 and the WordTree allows for dynamic investigation of words as they stem from other words—there is little research in the field of writing studies about the importance of dynamic data visualization models or tools. There are even fewer studies on audience engagement with dynamic data visualization models, though this is an area ripe for exploration and is currently being studied in STEM fields (Cui et al., 2010). Audience engagement with visualized data creates new possibilities for discovery and challenges notions of what textual analysis can be. As Geoffrey Rockwell (2003) encouraged audiences to “think of our tools [datavis] as creating possibilities for interpretation” (pp. 213–214), we also want our readers to think of the tools we utilized as creating possibility for writing centers to rethink their data and practices. In addition, an increased awareness of how visualized data structures and organizations themselves influence thought or invoke questions has often been the purview of STEM fields and needs to be explored in writing studies. Visualizations are sometimes seen as shiny objects, but they need to be interrogated in the same way text itself is interrogated; left to right, top to bottom, large to small—form and function are inseparable. The ways data are presented and organized determine how those data are then interpreted (Gopen & Swan, 1990).
Although we showed the limits of data visualization in forming conclusions about our work, we also demonstrated the way in which it can provide valuable insights in and across writing centers, as well as in the field of writing center studies. As Colin Ware (2004) noted, data visualization can facilitate hypothesis formation; that is, data visualization often does not provide the end product of research but sparks inquiry and tentative answers that need to be verified by other means. In addition to provoking meaningful questions and critical thinking about our field, our data visualizations helped us to understand language in context and how individual words function as part of a larger network of meaning. Given that many writing centers are robust data warehouses, the time is ripe to bring text mining and data visualization tools to the conversation to better understand these data and to present them intentionally and clearly.