Advancing the Field of Writing Analytics: Lessons from "Text-as-Data" in the Social Sciences

Abstract

The field of writing analytics is currently in a state of growth, redefinition, and refinement. In this essay, I review the trajectory of a related subfield, text-as-data in political science, as a lens through which to understand the present and future of writing analytics. I first describe how text-as-data has evolved over several eras, before transitioning to a review of some of the most exciting contemporary areas of political text-as-data. I then draw parallels between these developments and the work being done by the community of writing analytics scholars across the globe. I conclude by proposing several considerations for current practitioners seeking to emulate the "success story" of political text-as-data, including investment in the collection of new, high-quality corpora, development of shareable and open-source methodological tools for text analysis, and the strengthening of a community of scholarship.

Arriving at the everyday: building the NUWPArc public digital writing program archive

“…Complicating the archive's silence is the selection of collection records into "sets" that can be used as "data" for a variety of research purposes. Anson (2020) takes up the term "textsas-data" from the social sciences (political science, specifically) in discussing how "scholars have borrowed tools from computer science and computational linguistics, applying them to the large-scale examination of texts using techniques like sentiment analysis, document classification, causal analysis, and textual network analysis," which together have "led to novel theorizing and major methodological advances" (2020, p. 2). 53 The concept of texts-as-data is important to consider for NUWPArc and for writing studies research because, as Connors (1992) writes, "the overwhelming bulk of data from the past that the historian of composition studies must deal with is in written and printed form" (p. 51).…”