L2 writers' revisions in a computer-based academic English writing test task: a keystroke logging study

Abstract

This study investigated the revision patterns and purposes of forty native Urdu speaking English L2 writers in a computer-based academic English writing test task. The chosen writers were divided into a more successful and less successful group based on the score obtained in the given IELTS academic writing task 2. A keystroke logging program Inputlog recorded each writing session including all revisions, and it could be re-played after the session. From the recorded sessions, idea or meaning related revisions were further investigated. The writers were shown these revisions immediately after the session as a stimulus, and verbal protocols were obtained to investigate their revision purposes. The revisions obtained from keystroke logging program were analysed using a revision taxonomy based on Stevenson, Schoonen and Glopper (2006) and Lindgren and Sullivan (2006) studies. Findings suggested that in general, both groups focused on form revisions more than the concept revisions. But, more successful writers made conceptual revisions more than form revisions. Also, more successful writers revised larger text chunks, such as clause, sentence and paragraph. Both the groups reported seven revision purposes; however, eighth category 'writers did not remember' also emerged from the data. The seven revision purposes included clarity, explicitness or emphasis, coherence, hedging, modify an idea, eliminate an idea, creating a new global content unit and making text reader-friendly. However, more successful writers expressed more awareness of their revision purposes than less successful writers. The findings of this study have theoretical, pedagogical and methodological implications, particularly in the Urdu L1 context. These findings provide empirical evidence about the revision patterns and purposes of L2 writers which may be useful for the teachers or instructors teaching English writing to L2 writers, particularly to the Urdu speaking English L2 writers. They can use these findings to help them improve their writing skills by focusing on the revisions. However, future research in this area is suggested.