Towards Effective Structure-Based Assessment of Arguments and Proposals in Online Deliberation

Deliberation, i.e., discussing and ranking different proposals and making decisions, is an important issue for many communities, be they political, be they boards of experts for a scientific issue. Online deliberation however has issues, such as unorganized content, off-topic or repetition postings, or aggressive and conflicting behavior of participants. To address these issues, based on a relatively simple argumentation model and on feedback of different type, we propose to weight community members in an elaborate manner; this in turn is used to score arguments and proposals. Given such a scoring scheme, it is important to examine to which extent individuals have understood and accepted the approach, to identify characteristics of ‘good’ discussants and of strong arguments and proposals, and to study the robustness of the approach with regard to minor changes. To this end, we have carried out an experiment with a real-world community which had to make subjective decisions on issues relevant to them, and we have analyzed the data generated by it systematically, covering the different layers of our approach. Our takeaway is that the approach proposed here is promising to improve deliberation in many settings.