| Sumario: | Background: Online learning has become increasingly prominent in education, necessitating a deeper understanding of its impact on students' cognitive load (CL). Existing studies often focus on single data sources and overlook multimodal assessments, limiting a deeper understanding of how CL and emotional states manifest and interact during online learning. Objectives: This study examines the relationship between CL and emotional responses in asynchronous online learning environments to address existing research gaps. Using a multimodal approach, it aims to develop and establish the convergent validity of a composite index for estimating experienced cognitive strain during online learning. Method: We analysed the CL of 24 students during online learning sessions on regression analysis and logistic regression. Data sources included false task rate, emotional responses from voice and video data, and brain signal analysis. The comprehensive Cognitive Load Index (CLI) was developed using these measures, with the Analytic Hierarchy Process (AHP) assigning weightings to each data source. Results and Conclusions: Our findings indicate significant correlations between CL, emotional responses, and performance metrics. There were strong positive correlations with NASA Task Load Index (NASA‐TLX) scores, providing convergent validity evidence for the proposed index, and with false task rate, indicating that higher cognitive strain is associated with poorer task performance. Notably, emotional responses, particularly negative emotions, served as the strongest modality‐level contributors to the overall CLI. These multimodal data were effectively integrated by CLI, offering a convergent‐valid composite estimate of experienced cognitive workload during online learning. The study provides a methodological foundation for passive, observer‐based CL assessment in online learning environments for educators and instructional designers to enhance online learning experiences. Practitioner Notes: What is already known about this topic ○E‐learning is a prevalent educational method, but persistent challenges like low learner engagement and high dropout rates remain significant concerns.○Cognitive load (CL) and emotional states—specifically negative affective responses—are key factors that impact learning efficiency and outcomes.○Traditional assessment of CL often relies on single data sources or retrospective self‐reports, which may fail to capture the dynamic, multidimensional nature of mental effort.What this paper adds ○This study introduces a Comprehensive Cognitive Load Index (CLI) that integrates objective performance data (false task rate) with negative emotional responses derived from audio, video, and brain signals (EEG).○This research utilizes the Analytic Hierarchy Process (AHP) to assign expert‐informed, transparent weights to different data modalities.○Findings reveal that negative emotions are the strongest modality‐level contributors to the overall estimation of a learner's cognitive strain.○The CLI demonstrated strong convergent validity through significant correlations with established benchmarks like the NASA Task Load Index (NASA‐TLX).Implications for practice and/or policy ○Affect‐Aware Design: Educators and instructional designers should prioritize emotional engagement, as negative emotions compete for the same limited working memory resources required for learning.○Passive Monitoring: The study provides a methodological foundation for passive, observer‐based assessment, allowing for real‐time monitoring of cognitive strain without interrupting the learning process.
|
|---|