Analyzing large text data for vocabulary profiling in corpus-based studies of academic discourse

This article introduces a protocol designed to analyze large corpora for vocabulary profiling, aimed at enhancing corpus-based studies of academic discourse. Given the complexity and volume of data typical in academic fields, this protocol integrates advanced corpus compilation techniques with lexical analysis tools to effectively identify and categorize vocabulary suitable for academic use. The study details the systematic process of compiling a large corpus of academic texts, and describes the adaptations made to corpus linguistics tools to handle and analyze a corpus with 278 million running words efficiently. Validation of the mid-frequency word list demonstrated its strong relevance to chemistry, with 6.4% coverage in chemistry research articles and 2.5-3% coverage in related disciplines like biology and life sciences. However, the coverage was much lower in general corpora, highlighting its specialized nature. This methodology not only provides a framework for academic vocabulary profiling but also offers scalable solutions for educators and researchers dealing with extensive text datasets. The findings contribute to advancing vocabulary research in chemistry and related fields, offering practical applications for improving educational resources and designing more effective curricula for academic English. The resulting vocabulary lists have significant implications for the design of curricula and educational resources, aiming to improve both the precision and effectiveness of language instruction in specialized academic settings.•Developed a scalable protocol for analyzing large text data for vocabulary profiling.•Applied advanced lexical analysis to a 278-million-word academic corpus.•The mid-frequency vocabulary list produced offers pedagogical value in academic discourse.