This dissertation research investigated the role of online discussion in students’ experience in an academic writing class. As an intervention study, I implemented 20-minute-long online discussions at the end of every class period over a semester as part of required class activities and measured students’ subsequent timed writings and their first and final essays to trace some possible influence from online discussion to their writing development. Topics for online discussions were organized according to course objectives and the day’s lesson, with students developing subtopics reflecting their own interest according to the evolution of each discussion. These topics included theoretical concepts on academic writing as well as orthographical, lexical, grammar, and discourse-related inquiries. Participants included 10 treatment and 12 control students registered in two sections of a rhetoric and composition course designated for non-native English speaking students at a private university. This course was not an ESL class, but was part of the regular composition course offerings, except that it was restricted to international students specifically. Data sources included the treatment group’s 26 online discussion transcripts, 12 sets of timed writings, individual interviews, field notes, two types of essays, and surveys. The control group contributed essays, one set of timed writing taken in the middle of the semester, survey responses, five class recordings, and an instructor interview. Data analysis was performed by using a mixed method approach. Results from online discussion transcripts revealed that treatment students made use of online discussions for their learning, shown through types and characteristics of language-, content-, and writing-related episodes and the semester-long changes and pattern in such talk. Interviews and survey data showed students’ positive learning experiences and changes in their perception toward computer-mediated learning experiences over the semester. In terms of students’ writing, the treatment group made significant improvement in their timed writings over the semester and also outperformed the control group in essay writing significantly, in five of seven categories on a writing rubric. The most significant finding from this study was the improvement of treatment students’ writing scores over the semester. This study suggests the possible value of incorporating computer-mediated instruction in writing instruction as well as future research ideas that bridge research on traditional L2 writing and technology-enriched language learning. / text