On 14th December, Duygu Bektik defended her thesis successfully, and now only minor corrections stand between her and her doctorate.
Learning Analytics for Academic Writing through Automatic Identification of Meta-Discourse
When assessing student writing, tutors look for ability to present well-reasoned arguments, signalled by elements of meta-discourse. Some natural language processing systems can detect rhetorical moves in scholarly texts, but no previous work has investigated whether these tools can analyse student writing reliably. Duygu’s thesis evaluates the Xerox Incremental Parser (XIP), sets out ways in which it could be changed to support the analysis of student writing and proposes how its output could be delivered to tutors. It also investigates how tutors define the quality of undergraduate writing and identifies key elements that can be used to identify good student writing in the social sciences.