9. Discussion and conclusion
This study offers one potential response to Garrison and Arbaugh's (2007) research and concerns about the lack of empirical research using the CoI framework in different fields, particularly in language learning and teaching. More specifically, Shea, Hayes, Vickers, et al. (2010) identified that what is “[l]acking in this research is an attempt to test hypotheses generated by previous work documenting online knowledge construction through the interaction of all elements of the model” (p. 11). Thus, this study attempts to further previous research by reporting PTs' deep and meaningful knowledge construction of feedback practice in a virtual learning community. Theoretical implications of applying CoI to teacher learning can be drawn from this study by extending the existing CoI framework to teacher education in an Asian context, and practically, by delineating teacher learning processes at the interplay of the three presences. It can be concluded that when the PTs were provided with online participatory practice wherein they provided feedback for student writers and served as subject matter experts between each other, they were intensively engaged in identifying each other's feedback practice problems and generating pedagogical knowledge, which supported their re-conceptualizations of effective feedback. This finding is in line with some researchers' assertions (e.g., Borthwick & Gallagher-Brett, 2014; Guskey, 1986; Knezevic & Scholl, 1996) that teachers can learn to conceptualize effective teaching practices when they can visualize their own teaching behaviors and review student learning processes. By monitoring and analyzing their teaching behaviors and student learning processes, teachers can probe into the reasons why their feedback do not lead to improved student learning and help identify instructional support that students might need at different learning stages. It is suggested that PTs should be encouraged to change their role to serve as subject matter experts, in order to observe peer feedback practice, identify problems in peers' feedback practice, and provide suggestions to improve their practice. This study suggests that rather than by solely relying on teacher educators as the only sources to help identify the problems and provide advice, the PTs should share the division of labor when monitoring, reflecting, and scaffolding both their own and their peers' teaching practice. In this way, PTs can identify problems in their teaching presence and focus the discussion on common and pertinent pedagogical issues. Such findings are in contrast with most studies (e.g., Gallagher, Griffin, Ciuffetelli Parker, Kitchen, & Figg, 2011; He, 2009; Hung & Yeh, 2013) and the literature which stresses the importance of teacher educators in facilitating discussions between teachers in a teacher community. Most studies stress that without facilitation from the teacher educator in a teacher learning community, discussions between teachers often remain at low cognitive levels.
