Computer mediated discussions: effects of the previous messages' evaluations, knowledge content, social cues and personal information on the current message

This study of the flow of online discussions examined how previous messages affected the current message along five dimensions: (1) evaluations (agreement, disagreement, or unresponsive actions); (2) knowledge content (contribution, repetition, or null content); (3) social cues (positive, negative, or none); (4) personal information (number of visits); and (5) elicitation (eliciting response or not). Using dynamic multilevel analysis (DMA) and a structural equation model (SEM), this study analyzed 131 messages of 47 participants across seven topics in the mathematics forum of a university Bulletin Board System (BBS) Website. Results showed that a disagreement or contribution in the previous message yielded more disagreements and social cue displays in the current message. Unlike face-to-face discussions, online discussion messages that disagreed with a previous message elicited more responses. Together, these results suggest that teachers can use and manage online discussions to promote critical thinking, facilitate discussion of controversial topics, and reduce status effects.