Prof. Sam Pazicni (University of Wisconsin-Madison/USA): From Ideas to Networks: Cognitive Models of Students’ Chemical Reasoning

Cognitive models are central to meaningful assessment and instruction because they describe how knowledge is represented, activated, and coordinated in a domain. In chemistry, such models are especially important because students must connect observable phenomena with particle-level mechanisms, symbolic representations, and formal models. Without robust cognitive models, assessments risk measuring only whether students produce expected answers rather than how they organize and use ideas when reasoning.

This presentation describes work toward cognitive models of students’ reasoning about observable phenomena. Introductory chemistry students completed open-response prompts that asked them to illustrate, describe, and explain phenomena related to kinetic molecular theory. Responses were analyzed by identifying distinct ideas and links among them, then represented using Epistemic Network Analysis. This approach makes it possible to examine not only which ideas students express, but how those ideas are organized, as well as how different prompts elicit different reasoning strategies. Findings illustrate how networked cognitive models can support more responsive assessment and instruction by making visible the structure of students’ chemical reasoning.

Dr. Sebastian Tassoti (Universität Graz/Austria): Navigating cognitive debt and (un-)productive usage patterns: AI-associated challenges in teaching university-level chemistry

Integration of Generative Artificial Intelligence (GenAI) into university teaching of chemistry currently knows a limited number of applications that were proven to be fruitful for learning. On the contrary, integration is sometimes involuntary and one-sided, for example when students decide to use a GenAI-chatbot like ChatGPT to accompany their learning process. Recently, there has been a range of research on the impact of GenAI on learning, elucidating cognitive debt and acceptance of AI-output as well as productive and unproductive usage patterns when learning chemistry. This talk shows a two-sided approach towards more productive scenarios involving GenAI-chatbots in higher chemistry education: first, this talk will discuss how students can be prepared for the use of GenAI, for example by teaching strategies for prompting that work particularly well for some fields of reaction prediction. Second, there will be a focus on how GenAI chatbots can be prepared and developed to be more useful in chemistry education settings, with examples of chatbots designed to diagnose and address student conceptions on topics of higher chemistry education.