In a society which is calling for more productive modes of collaboration to address increasingly complex scientific and social issues, greater involvement of students in dialogue, and increased emphasis on collaborative discourse and argumentation, become essential modes of engagement and learning. This paper investigates the effects of facilitator-driven versus peer-driven prompts on perceived and objective consensus, perceived efficacy, team orientation, discomfort in group learning, and argumentation style in a computer-supported collaborative learning session using Interactive Management. Eight groups of undergraduate students (N = 101) came together to discuss either critical thinking, or collaborative learning. Participants in the facilitator-driven condition received prompts in relation to the task from a facilitator throughout the process. In the peer-driven condition, the facilitator initially modelled the process of peer prompting, followed by a phase of coordinating participants in engaging in peer prompting, before the process of prompting was passed over to the participants themselves. During this final phase, participants provided each other with peer-to-peer prompts. Results indicated that those in the peer-driven condition scored significantly higher on perceived consensus, perceived efficacy of the IM methodology, and team orientation. Those in the peer-driven condition also scored significantly lower on discomfort in group learning. Furthermore, analysis of the dialogue using the Conversational Argument Coding Scheme revealed significant differences between conditions in the style of argumentation used, with those in the peer-driven condition exhibiting a greater range of argumentation codes. Results are discussed in light of theory and research on instructional support and facilitation in computer-supported collaborative learning.