Spelling errors in the preliminary English B1 exam: Corpus-informed evaluation of examination criteria for MENA contexts

While the “corpus revolution” has advanced beyond lexicography and into the area of reference materials, coursebooks and data-driven learning for language learning, the role of corpora in informing language testing is still in its infancy. To address this dearth in the literature, this chapter centres on a corpus linguistic analysis of written assessments to evaulate learner language in different MENA contexts. The chapter presents an analysis of the error-annotated component of the Cambridge Learner Corpus, which is a 30-million-word corpus of learner language from Cambridge examinations. It focuses on comparing a subcorpora of Preliminary B1 written examinations by all Arabic speakers and all non-Arabic to those by students from Libya and the United Arab Emirates in order to test for variation across the MENA region. In the context of the Preliminary English B1 exam, errors are contrasted in order to illustrate that while similarities do occur in exam takers across MENA contexts, there are also important differences among Libyan and Emirati learners. Spelling errors are then discussed in greater detail and are contextualised in two case studies, concerned with misspelled verbs and misspelled words beginning with ‘wh/w’. The findings are then considered in terms of the assessment criteria for the specified examination with the aim to produce and deliver guidance on how assessment criteria could be localised.