Abstract
Despite the plethora of research on computer-supported collaborative writing, previous studies have failed to provide an adequate analysis of the quality and quantity of multi-authored texts due to the vast amount of unstructured data. This study aims to expand the current research by contributing new perspectives to the understanding of students' collaborative writing interactions and strategies. Nineteen EFL learners, divided into five groups, were asked to collaboratively write a literature review on a selected research topic using Google Docs. The qualitative and quantitative analysis of the data, supported by text mining using DocuViz, revealed that the groups demonstrated patterns of interactions along the continuum of equality and mutuality. However, due to the nature of the required task, groups with the same collaborative interaction patterns exhibited some writing strategy variations. Text visualization analysis further revealed that the groups adopted more than one writing strategy to complete the task. They tended to start with an outline in which the task was divided, and students worked either in parallel or sequentially. This was followed by cooperative revisions or a synchronous hands-on style before the task's due date. By using multiple data analyses, this study distinguishes between interaction patterns and writing strategies, two concepts previously used interchangeably. Methodological and pedagogical implications are also presented.