Jump Analysis: An approach to analyzing continuity in written text production
Psycholinguistics is a field that investigates the mental processes involved in speaking, listening, reading, and writing. Of particular interest to this project is the field of writing research, which focuses specifically on the cognitive processes involved in text production and how they can inform literacy, second language learning, and reading/writing disorders. While the use of biometric technology (such as eye-tracking and keyboard logging) has been utilized to study patterns in typing, pausing, and re-reading, current research on writing linearity (how a writer “moves around” when producing and editing a text) is more limited, and largely restricted to simply prescribing writing sessions as “linear” or “nonlinear”. This has traditionally been done by determining how frequently a writer moves away from the “leading edge” (the furthest cursor point in the document where the majority of text production occurs). However, the “leading edge” approach has limitations, both in automation (it may be misidentified if there are a few spaces or characters at the end of the document) and in oversimplifying the actual behaviors happening (such as assuming that true text production is only happening at the leading edge and editing behavior is occurring at all other points). This project adds to the current research on linearity in text production by proposing a novel approach: “jump analysis.” Instead of focusing on a single “leading edge,” this approach indexes all “edges” (or cursor locations where text is deleted or produced) and collects data on each of these “jumps” and the text chunks between them (“blocks”). By implementing this approach on a dataset of student writers, we conducted an exploratory analysis of linearity in writing with greater detail than afforded by traditional methods.
To cite this study: Godbersen, J., Dux Speltz, E., & Chukharev, E. (2024, May 1). Jump Analysis: An approach to analyzing continuity in written text production. ISU Honors Poster Presentation. doi.org/10.6084/m9.figshare.25739205