A computer-aided error analysis of Omani students' written English and an investigation of their grammatical competence: A corpus-based study

This thesis presents a detailed account of compiling, analysing and utilising a learner corpus to strengthen the connections between learner corpus research (LCR), second language acquisition (SLA), and English language teaching (ELT). First, strict design criteria to compile the Omani Learner of English Corpus (OLEC) are described. OLEC contains 210,283 tokens, and was error-coded following the Cambridge Learner Corpus error coding system. Second, this thesis (i) focuses both on learners’ errors, by conducting a computer-aided Error Analysis (CEA), and (ii) grammatical competence, by applying the English Grammar Profile (EGP) methodology (O'Keeffe & Mark 2017). In order to attempt to follow up on the implications of the findings, I compare the findings of this study against some teaching materials (e.g., textbooks, in-house materials, etc.). As a practical aspect of the thesis, I designed exercises from the OLEC to show how a learner corpus can be used in ELT in general and, most importantly, with the same type of learners in the OLEC. The CEA study reveals that learners have frequent errors in punctuation, spelling, word order, determiner, among others, largely attributable to the cross-linguistic influence from their L1 (Arabic). Grammatical competence investigation findings, on the other hand, establish that OLEC learners match the description of basic users (between A1 and A2) according to the Common European Framework of Reference (CEFR) classification. Last, the comparison between the findings of this thesis (in spelling, punctuation, verbs, determiners, and prepositions) and the used teaching materials by the institution reveals that the investigated teaching materials do not tackle most language areas those learners need to improve their language production. If they do (teaching materials), they tend to be incomplete. These findings provide a clearer picture of learner language, facilitating the designing of teaching materials that meet learners’ needs instead of relying heavily on global textbooks designed for many contexts. For textbook publishers, this study calls for the need of using any existing learner corpora to inform the materials in global textbooks; otherwise, they are destined to be neglected ultimately. Finally, this study urges teacher education programmes to prepare teachers to use learner corpora to understand the type of learners they deal with and encourage them to use them in their classroom.