Computer-aided error analysis

Abstract

This article introduces the technique of Computer-aided Error Analysis (CEA), a new approach to the analysis of learner errors which it is hoped will give new impetus to Error Analysis research, and re-establish it as an important area of study. The data used to demonstrate the technique consist of a 150,000-word corpus of English written by French-speaking learners of intermediate and advanced level. After the initial native speaker correction, the corpus is annotated for errors using a comprehensive error classification. This stage is a computer-aided process supported by an ‘error editor’. The error-tagged corpus can then be analyzed using standard text retrieval software tools and lists of all the different types of errors and error counts can be obtained in a matter of seconds. In this study, we use the corpus to analyze progress rate between intermediate and advanced level students along a range of grammatical and lexical variables, with interesting results. Initial research suggests that this approach provides one way of discovering important features of learner writing, in particular areas of persistent difficulty. As such, we believe it could play an important role in the development of the next generation of pedagogical tools, ones which incorporate more ‘learner aware’ information.