Online Writing Behaviors, Writing Outcomes, and Task Complexity in Computer‐Assisted Collaborative Writing

ABSTRACT This study examined how online writing behaviors aggregated at the pair level were related to writing outcomes in computer‐assisted collaborative writing and how task complexity moderated these relationships. The participants were 56 Chinese university students with advanced English proficiency. They were randomly organized into 28 pairs. Each pair performed two versions of a reading‐to‐write task in a face‐to‐face collaborative writing setting using Google Docs. Task complexity was manipulated as whether there was a clear text structure to follow when writing a summary. The simple task version involved summarizing a complete text, whereas the complex task version required summarizing information from three separate texts. Each participant's online writing behaviors recorded by the keystroke logging software were coded for speed fluency, pause length and frequency, and revision frequency. Individual data were then aggregated as pairs’ writing behaviors. Pairs’ writing products were assessed for linguistic quality and text quality. Linear mixed‐effects regression analyses identified that, longer P‐bursts, shorter pauses between sentences, and more frequent pre‐contextual revisions were associated with improved text features in some respects. Task complexity played a small but non‐negligible role in the relationship between between‐sub‐sentence pause frequency and contextual revision frequency, and certain text characteristics. These results highlighted the diverse nature of the relationship between pairs’ online writing behaviors and writing products in collaborative writing of tasks with varying complexity.