Roed's (2003) study and imporvement

Roed’s (2003) exploratory study on the text-based online communication is informative because a lot of merits can be found with such new approach of teaching English. One of the strength with virtual communication is obviously its absence of accents of the speakers. I well remember my fellow students in high school told me that they felt their accented English ‘unpleasant’ to the listeners, and thus unconfident to speak. Also, they felt that grammatical errors in their speech can be mocked at by their classmates. That is why many Chinese EFL learners spend most of the time listening in a spoken English training session.
While traditional classroom teaching may offer little help with this respect, Roed’s online communication treatment may be effective in eliciting more response from introvert students. What we can learn from Roed’s experiment is that teachers can provide the students with tasks that require information gap filling by text-based online communication.
However, several points need to be considered in the conducting of such activities. First, the students might easily go off the topic and talk about something irrelevant. Although a slight deviation from the task may be acceptable, this should not last for too long. The teacher should set a proper time limit to the task in this respect. Second, it is uncertain from Roed’s study as to whether more advanced students would try to focus the communication on form, that is, pointing out grammatical errors in the responses by the peers. Corrective feedback on forms can be helpful for the slower students to learn grammar, but it can also be a frustrating impediment to the smooth interaction. What the teacher can do before hand is to establish the rule that grammar errors may not be treated as long as they do not impede the meaning. Third, there will certainly be some freeloaders who simply read but not engage in the activity. It might be practical to make it compulsory for all students to participate by telling them that a minimal level of engagement is required.

Reference:
Roed, J. (2003) Language learner behaviour in a virtual environment. in Computer Assisted Language Learning, 16(2), 155-172.