Correspondence analysis: Comparing wordlists across specialised corpora

Correspondence Analysis (CA) is one of the multivariate analyses, which can be adopted to investigate statistical similarities and differences across corpora. This study demonstrates the application of CA in a corpus-based study of lexical items of specialised corpora, focusing mainly on highly frequent and significant words in the context of English for Electronics and Computer Engineering. The CA conducted with the most frequent, most keyed, and most key-keyed wordlists of specialised corpora reveals the consistency in the findings, in terms of differentiating the corpora from each other as well as highlighting words which characterise each corpus in the study. Thus, this confirms the potential use of the CA as a tool for language description. The results provide interesting insights into the complex interrelationship between the corpora. In addition, the identification of words which make up a specialised corpus allows further investigation into their lexical behavior in the specialised texts for English for Specific Purposes (ESP) pedagogical considerations.