Project Lead: Madiha Anjum
Mental models are informal representations of how the world works—they shape how we perceive the world around us and influence everything we do. People build and maintain mental models through direct observation, learning, and experience, and are then relied upon to reason, explain, design, communicate, decide, act, predict, and explore. Mental models thus play a fundamental role across all areas of human agency. Practitioners from fields as diverse as behavioral science, psychology, economics, education, and sustainability would, therefore, gain much insight from tools that enabled them to derive and formalise computer simulations from the mental models expressed by a group in relation to a specific question or problem situation.
The goal of this research is to develop a real-time, moderated steering environment for better understanding and formalising mental models from online discussions. Users are presented with a topic, question, or problem to debate, on which they can express their opinions/thoughts in the form of comments. Users can also respond or expand on other user’s opinions and/or initiate a new line of discussion. The discussion chain is mined in real-time, using an ensemble of algorithms (including but not limited to concept mining, topic modeling, and sentiment analysis) to extract opinions, keywords, and concepts. This information is transcribed into semi-quantitative simulations using Fuzzy Cognitive Maps, Causal Loop Diagrams, and Networks Diagrams as the discussion evolves. The system presents users with dynamic visualizations of the collective mental model which they can use as contextual information to refine and update their individual mental models.
The system is also designed to allow a moderator to “inject” feedback via targeted comments, aimed at steering or nudging the discussion towards a desirable collective mental model, outcome, consensus, or agreement. The framework will be validated and tested in a number of case studies, but it could be used in any social setting or problem situation where the steering of collective mental models plays a critical role in enhancing the functioning, resilience and/or sustainability of a socio-technical or socio-environmental system. More broadly, government, research, and advocacy institutions could rely on our software to build more direct, transparent and meaningful engagement with their stakeholders.
Develop a framework that captures representations of a group’s mental model from an online web forum and translates these into semi-quantitative computer models. The framework should allow the moderator (a researcher, government agency, or non-profit organisation) to extract and analyse cognitive maps that provide relevant and actionable insights into questions about the structure and evolution of a group’s mental model concerning a particular topic. The framework should provide both users and moderators with the information needed to debias the group’s mental model, and in doing so, produce new and creative solutions to complex and controversial societal issues.
- Compare and analyze various techniques/approaches for data mining and identify approaches which are most suited for finding the opinion of users.
Develop a dynamic, automated framework to translate the opinions expressed in a discussion forum into semi-quantitative models.
Develop a real-time and moderated online communication/discussion tool to involve/engage people in discussions.
Explore how users restructure knowledge and update their mental models based on their interactions with the proposed framework.
The web-based platform named DISCUSSOO has been developed where users are presented with a topic, question, or problem to debate, and where they can express their opinions/thoughts in the form of comments. Users can also respond or expand on other users' opinions and/or initiate a new line of discussion and they can share this discussion on other social networking websites. We are analyzing the ongoing discussions at this stage.
- Anjum, M, Voinov, A, Castilla Rho, J & Pileggi, SF 2019, 'Understanding mental models through a moderated framework for serious discussion', 23rd International Congress on Modelling and Simulation, Canberra.