A computer-aided error analysis of Saudi students’ written English and an evaluation of the efficacy of using the data- driven learning approach to teach collocations and lexical phrases

The thesis reports on three corpus-based studies in a Saudi university context. The first study is a computer-aided error analysis (CEA) of a corpus of Saudi English majors’ writing. The second and third studies employ the DDL approach to teach collocations and lexical phrases. The errors in the Saudi learner corpus (SLC) were tagged following the Louvain Error Tagging Manual 1.2. The CEA revealed that the ten largest error subcategories were (Form, Spelling), then (Grammar, Verb Tense), (Lexical, Single), (Grammar, Articles), (Grammar, Verb Number), (Grammar, Noun Number), (Word Redundant, Singular), (Word, Missing), (Lexical, Phrase) and finally (Punctuation, Missing). These error types are analysed qualitatively to identify the linguistic features that seem to be problematic for Saudi EFL learners. Multiword units are notoriously difficult for L2 learners, and the Saudi EFL context is no exception; in the second and third studies a number of collocations and lexical phrases were selected from the SLC to be taught using DDL paper-based and dictionary-based materials. The results showed that learners in general learn better under the DDL treatment. Learning gains as a result of the DDL instructional condition in short-term delayed posttests were not significantly better than the dictionary-based instructional condition in the case of collocations, but they were significantly higher for the lexical phrases. The DDL long-term delayed posttests results were significantly better than the dictionary results for both the collocations and lexical phrases. A questionnaire and retrospective interviews were used to investigate students’ and teachers’ attitudes and the results encouragingly revealed that they felt positive about the DDL materials. The data shed light on strengths and weaknesses of the DDL and the traditional approaches. The thesis closes with a discussion of the pedagogical implications, particularly with reference to the use of corpus tools and corpus-based materials in the Saudi EFL context.