Keyness Analysis Of Lexical Bundles In ESL Learners’ Writing: A Corpus-Based Study Of CEFR Levels

Abstract

The keyness analysis of lexical bundles from ESL learners’ academic writing at CEFR B1, B2, and C1 levels is studied by means of its corpus-based approach. Hammond, J. (April 29, 2014), “The Big Picture on Lexical Bundles: Recurring sequences of three or more words (lexical bundles) are an indispensable feature of writing proficiency in the domain of academic writing because the use of bundles typically increases writers’ fluency, coherence, and syntactic complexity.” The most frequent lexical bundles and their structural and functional variations were identified in essays written by ESL learners at different proficiency levels. Keyness analysis (a method for identifying the prevalence of word bundles within a corpus relative to reference data) is then used to explore how these bundles vary by CEFR level and against a corpus of native academic writing. The ESL learners’ argumentative essays were collected based on CEFR levels and compared to British Academic Written English (BAWE) corpus. The quantitative corpus techniques (such as frequency and dispersion metrics) are combined with qualitative assessments of the bundle’s structural and functional characteristics. Results show distinct developmental patterns in lexical bundles, with more advanced students using a more varied, accurate, and functional bundle. Moreover, the study reveals specific lexical bundles that are either overabundant or need more if a particular proficiency level is reached regarding academic writing competence. In so doing, this research advances the corpus linguistics study of lexical bundles as measures of writing proficiency, second language acquisition, and academic writing pedagogy. These findings provide some practical implications for teaching English for Academic Purposes (EAP) because they suggest that the design of instructional materials and targeted interventions on writing improve ESL learners’ writing skills. This dissertation integrates keyness analysis and lexical bundle research to better understand language development across proficiency levels.