| This research paper addresses the new field of online learning, using a system of discourse analysis created by Arno Bellack and further explored in face-to-face mathematics classes by James T. Fey. Discourse analysis provides a powerful tool for examining the fundamental structure of an online class, a way of examining what makes such classes revolutionary for some, while an anathema to others. Thirteen online mathematics classes, involving seven experienced online instructors, were observed and coded using Bellack and Fey's system. The majority of codes had excellent levels of inter-coder reliability, as measured by Cohen's kappa. The discourse structure of each course and the types of discourse that encourage students was examined. Findings suggest that the ratio of teacher to student discourse is far lower in online classes than in face to face classes, and varies widely from one instructor to another. An instructor posting proportion of about 22%, posting about every fifth post in the discussions, seems to be ideal---both more or less than that is linked with lower levels of student posts. An increase in instructor posts is associated with an increase in posts per student. In addition, requiring students to post may be important. A large number of instructor responding posts (posts that responded directly to a student request) correlated negatively with posts per student, while posts that were evaluative (by rating, clarifying and expanding on what students said) had a positive correlation with posts per student. (Both theses terms, responding and evaluative, were codes defined by Bellack and Fey and refined in this research.) In other words, when instructors ask questions within the discussion and encourage other students to answer each other's questions and at the same time maintain an instructor presence by evaluating student posts rather than responding directly to student questions, conversation among students increases. This posting style may also mitigate the tendency of an instructor's post to end a thread. Making sure all posts are specifically about mathematical problems may be linked to a decrease in student conversation. Finally, there is some evidence that the instructional platform (the computer learning management system) may matter far less than the instructional strategies. |
