How Do EFL Learners Regulate Corpus Consultation? A Qualitative Study of Self-Regulated Learning in Data-Driven Writing

Abstract

Corpus-based language pedagogy has shown some positive results in EFL writing, but how learners organise the consultation process itself has rarely been studied. This paper investigates how learners regulate corpus consultation in data-driven learning under an examination-oriented writing task. Twelve Chinese undergraduates preparing for the CET-4, sampled from the higher and lower ends of the proficiency range, participated in eight Sketch Engine-supported argumentative writing sessions. Guided learner logs of 412 search episodes and retrospective interviews were analysed using hybrid thematic analysis, combining the stages of self-regulated learning as an a priori framework with inductive coding in the phases. The analysis found seven subthemes that describe consultation as a cycle in which learners frame searchable problems around multiword units, monitor and revise queries based on concordance evidence, and convert search results into modified routines. Higher-proficiency learners planned queries at the phrase level and weighed evidence across lines; lower-proficiency learners searched in fragments and readily accepted evidence, and individual trajectories showed that these patterns were modifiable habits rather than fixed attributes. Based on the above results, phraseological competence and regulatory skills develop together; therefore, teaching should focus on practising search decisions rather than tool operations and establish consultation strategy as enduring content for corpus-informed writing pedagogy.