Error Analysis in a Learner-Written Corpus from Spanish Speakers EFL Learners : A Corpus-Based Study

Abstract

ABSTRACT:This thesis was conducted to investigate the relationship between the main errors found in written compositions from students at university level and the socio-demographic factors that could have incidence in the development of the writing skills at Universidad del Norte in Barranquilla, Colombia. The participants in the present research are 515 university students classified through a placement test in accordance with the parameters from the Common European Framework of Reference for Languages (CEFR). The data was compiled following the steps of Corpus Linguistics methodology, correspond to the written compositions of the participants, and was the final task for the course of English as a Foreign Language. The analysis of the data was based on Corder’s ( 1981) theory about Error Analysis (EA) and following the error description guide given by James (1998). This thesis presents a comprehensive framework of error analysis that aimed to reveal the frequent error patterns in a leaner corpus of university students as well as the relationship of these errors with some socio-demographic aspects that may have an impact on the development of writing. 14,631 errors in eight categories were identified and analysed by categories according to the EA methodology. The errors were labelled using Louvain University Error Tagging Manual version 1.2 (E. Dagneaux et al., 2005) in order to obtain comparable results with similar work worldwide. The analysis finds that among the eight error categories proposed by Louvain’s tagger grammatical errors are the most recurrent. The results highlight the difficulties the learners present in the use or omission of the definite article. A comprehensive account and description with examples of the main errors in the different categories along with possible causes for future pedagogical intervention are presented.