Comparing AI chatbot, tiered, and textual support in physics problem solving: Effects on cognitive load and affective-motivational outcomes

Providing appropriate support is important for helping students manage conceptual difficulties in physics. Established scaffolding methods, such as structured tiered support, can be effective but are often resource-intensive to develop and implement broadly. AI chatbots may offer a more easily scalable alternative, but their potential as learning support tools needs to be examined carefully and in direct comparison with established forms of support. As a first step, this study compares a custom-configured AI chatbot, a tiered support system, and traditional textbook-style textual support with respect to students’ perceived cognitive load and affective-motivational outcomes during physics problem solving. In a cluster randomized field study, 268 ninth-grade students in Germany worked individually on a buoyancy problem using one of the three support formats. We measured intrinsic and extraneous cognitive load and affective-motivational outcomes (enjoyment, hope, hopelessness, self-efficacy, and situational interest) via research-validated questionnaires. Results showed that the AI chatbot and the tiered support system were associated with significantly lower perceived intrinsic and extraneous cognitive load than textual support. Post-test comparisons showed that students in the chatbot condition reported more favorable affective-motivational outcomes than students in the textual-support condition, while differences between the chatbot and tiered-support conditions were not statistically significant. Overall, the findings suggest that interactive support formats may shape students’ short-term perceived cognitive and affective-motivational experiences during physics problem solving; future research should examine how these experiences relate to solution quality, conceptual learning, and longer-term outcomes.