Genre analysis in the academic writing class: with or without corpora?

Abstract

This article presents the results of a study that compared two English for academic writing classes designed for international graduate students. Both classes focused on guiding students in the analysis of research articles in their disciplines to identify linguistic and organizational conventions frequently used by published authors. One of the classes was computer-based and corpus-based: students analyzed research articles in a corpus they collected for this class. The other class was taught in a regular setting, having student investigate only four hardcopies of articles extracted from disciplinary journals.

Student production from both classes were evaluated by a group of trained raters and showed that students could reproduce the schema organization and linguistic conventions they had identified in the research articles they learned to analyze. The quality of the written production of the two groups, however, did not present significant differences. Student surveys and interviews showed that the use of corpora was perceived as positive by most students in the corpus-based class. On the other hand, students in the non-corpus class considered that the number of articles they could analyze was a strong limitation preventing them from making generalizations.