Identifying academic phrases in self-compiled corpora: A case study from mathematical sciences

Multi-word combinations are a key component of a variety of academic genres. Student writers thus need to master control of these expressions to communicate effectively in their specific contexts. However, multi-word patterns often present a challenge to students at various levels of academic study. This paper reports on a case study exploring how a data-driven learning approach drawing on self-compiled corpora can enhance student writing in preparation for their doctoral study at a British university. The participant was an international student in mathematical sciences who utilized a self-build corpus throughout a ten-week pre-sessional programme to acquire and incorporate multi-word expressions in their writing. The effectiveness of this approach was investigated through a ‘talk around text’ technique in an interview on completion of the pre-sessional programme. The findings show that a self-compiled corpus is a valuable resource for students mastering writing in specific disciplinary contexts.