Towards a Context-Dependent Framework for Visualizing Mental Models

Abstract

Mental models describe an internal representation of knowledge, which can be used to describe, explain, and predict a person’s interactions with their environment. While there is no shortage on mental model research, a recent literature review revealed the majority of research on mental models does not include an actual model of any kind. Mental models are able to capture complex information, which is difficult to effectively convey through written word alone. A set of visual metaphors are needed to effectively convey the disparate information that can be identified in mental models. We have conducted initial research to develop a notional framework for visualizing mental models based on the context of the elicitation methods used (and the data collected).

A Framework for Eliciting Players’ Mental Models of Game Mechanics

“…Harper and Dorton believe that mental models can describe knowledge structures, both from in individuals and in groups, the latter of which they refer to as "shared mental models", and they assert that mental models are important for understanding and successfully executing complex tasks that require co-ordination and communication (Harper & Dorton, 2020a). Having acknowledged that there are effective ways to elicit mental models that apply to individuals and groups, and having developed a framework for selecting the most appropriate elicitation method for any given situation, they followed up this work with a sequel paper on mental model visualization (Harper & Dorton, 2020b). In the process, they explored visualization techniques such as pathfinder networks, ordered tree diagrams and concept maps, along with other prominent methods for representing the moving parts that make a system, or any entity in general, tick.…”

“…The third stage (high experience) featured a low amount of subnetworks but an even higher amount of procedural connections, as well as higher amounts of everything else. Mental model exploration, when applied to my research questions, is likely to facilitate the investigation of Harper and Dorton's take on elicitation and visualization (Harper & Dorton, 2020b), about mental models being internal representations of reality that develop over time. In my research, the goal is to understand how people understand complex systems, and ways that understanding can be applied; whereas in the study on experience as a factor in mental model development, the research question hinges on studying the relationship between different facets of mental models and how they differ across three levels of video game experience.…”

“…This adaptability could allow AI agents to continuously learn from user interactions, improving their performance and better meeting user expectations over time, which could then lead to a more fulfilling user experience. Human mental models facilitate comprehension of ambiguous or incomplete information and enable effective communication (Harper & Dorton, 2020b), and by aiming for the same in AI agents, designers could enhance the agents' natural language processing capabilities and enable more sophisticated and context-aware communication with users. For AI agents to gain user trust and acceptance, they need to be transparent and capable of rationalizing and explaining their decision-making, and designing them to do so could lead to less friction in their application in a variety of systems.…”