From Corpus to Curriculum: Disciplinary Variation in Doctoral Abstracts and Its Implications for Writing Instruction

A recurring challenge in English for Academic Purposes (EAP) writing instruction is deciding how far a single, shared curriculum can serve doctoral writers from different disciplines, and where teaching must instead be calibrated to particular fields. Drawing on a corpus of 800 English doctoral abstracts submitted with Chinese doctoral dissertations in Biology, Chemistry, Physics, and Engineering, this exploratory study examined 20 lexical and syntactic complexity indices to identify where disciplines differ or show limited differentiation. The Entropy Weight Method (EWM) estimated each index’s distributional salience, while multivariate and univariate analyses assessed disciplinary differentiation. Disciplinary effects were broader and generally larger among the lexical indices than among the syntactic indices. Chemistry had the highest means on several lexical sophistication and diversity indices, while Physics showed a narrower pattern centered on verbal and modifier variation. Syntactic differentiation was more limited: coordinate phrases per T-unit (CP/T) was the only syntactic index significant after Holm adjustment, and conclusions from the non-inferiority analyses depended on a post-specified effectsize bound. The findings provide a provisional rationale for combining shared instruction in abstract organization and information packaging with sustained, discipline-specific vocabulary work; they do not establish a universally sufficient syntactic common core.