A Corpus-Based Study of Academic Writing Patterns Among EFL Learners

Abstract

English as a Foreign Language (EFL) learners face persistent challenges in producing academically acceptable written texts. While pedagogical interventions exist, empirical evidence on systematic patterns of lexico-grammatical and rhetorical features in learner corpora remains limited. This study investigates the academic writing patterns of Saudi EFL learners at the university level, focusing on lexical bundles, collocational errors, and rhetorical organization.  A specialized corpus of 200 argumentative essays (approximately 85,000 words) was compiled from intermediate to advanced EFL learners. Using AntConc and LancsBox, frequency lists, keyword analysis, and concordance lines were generated. The corpus was compared against the British Academic Written English (BAWE) corpus as a reference.  Findings reveal three dominant patterns: (1) over-reliance on high-frequency lexical bundles (e.g., on the other hand, as a result, in my opinion), often misused in formal contexts; (2) significant collocational deviations, particularly verb-noun (e.g., make a research instead of do/conduct research) and adjective-noun combinations: and (3) rhetorical patterns showing topic-fronting and informal discourse markers absent in native academic writing. EFL learners systematically transfer spoken discourse features and L1 rhetorical structures into academic writing. The study recommends explicit corpus-informed instruction targeting collocational precision and register awareness.