(89) Tracing the influence of game elements on learning via physiological and behavioral process measures
Date:
Contributors: Huber, S. E. , Lindstedt, A., Kiili, K., & Ninaus, M.
Venue: EARLI 2025, Graz, Austria, August 25-29, 2025
Abstract: Evaluation of game-based learning is often based on learning outcomes. However, the latter depend on cognitive and affective processes during learning which cannot be observed directly. Their investigation requires indirect methodological access through associated process measures like psychophysiological or behavioral signals. The current study thus set out to investigate if and how these process measures can be utilized to capture the influence of game elements on learning. To do so, heart rate, heart rate variability, electrodermal activity, eye movements, and eye blinks of 89 participants were recorded during an associative learning task. In this task, participants needed to learn associations between 20 symbols and numbers over 5 trials. A non-gamified and a gamified version of the task were implemented to investigate the effects of game elements by randomly assigning participants to one of those versions. Using logistic regression with 10-fold cross-validation, we find that all process measures except heart rate allow correct classification of participants into their learning conditions (non-gamified or gamified) significantly above chance level (50%, p < .005) if measures are aggregated over all trials. For classification in individual trials, we find that there is no combination of measures that performs best in each of the 5 trials. This indicates that different processes are relevant for learning in different phases of the task. This change of relevant processes and measures thereof needs to be considered when multimodal assessment of psychophysiological and behavioral process measures is envisaged to provide a mapping of the dynamics of cognitive-affective processes during learning.