Characterising postgraduate students’ corpus query and usage patterns for disciplinary data-driven learning

摘要: Data-driven learning (DDL; Johns, 1991), involving students’ hands-on use of corpora for self-guided language learning, is a methodology now increasingly used in many tertiary contexts to enhance the teaching disciplinary postgraduate thesis writing. However, there are still few studies tracking actual engagement with DDL. This mixed-methods study reports on corpus via purpose-built query and data visualisation platform integrated into large writing program at university Hong Kong. Data usage history (e.g. times access, duration use), syntax lexis/phraseology wildcards part-of-speech tags), function frequency lists/distribution, concordance sorting collocation) filters searches by faculty, discipline, or section) were collected from 327 students spanning over 11,000 individual queries. The results show significant interdisciplinary inter-/intra-user trends variation particular functions adopted users. Students varied type knowledge domain-specific, language-specific) they accessing, frequently went beyond exemplars DDL course materials generate unique queries under their own initiative. Qualitative case three users’ activity logs also distinctive function. These provide clearer insight what actually do during different directions trajectories that users take as result All accompanying tasks included supplementary materials.