Designing collaborative resources in data-driven learning to accommodate learner diversity

Data-Driven Learning (DDL) usually relies on existing corpora like BNC and COCA, but these may not align with students' needs when writing an abstract. To address this issue, designing and implementing a tailored corpus could be a solution. This study explores students’ perceptions during the collaborative creation of an abstract corpus (RQ1) and its strengths and weaknesses in an academic writing process (RQ2). This research was conducted as a case study involving 80 final-year engineering students who built a corpus of abstracts to serve as a writing reference. Despite varying interests, the participants collaborated to compile a corpus of relevant abstracts. They worked collaboratively to collect, annotate, and analyze the corpus. To support corpus development and the DDL process, Google Drive was used as online storage, and AntConc software was used as an analysis tool. The data were collected through interviews and video recordings. The video recordings were used to capture the DDL process starting from preparation to corpus utilization. The retrospective interviews with five randomly selected participants were conducted to explore participants’ perceptions. The findings suggest that collaborative corpus development enhances students’ understanding of abstract writing genres and promotes engagement. Different research interests did not hinder the corpus development and utilization. Two observable weaknesses were the corpus tool installation problem and the interface. Despite this, the collaborative corpus development fostered inclusivity and deeper involvement, highlighting the potential of DDL to meet diverse students’ needs.