Using multimodal analytics to systemically investigate online collaborative problem-solving

Abstract

Abstract The purpose of this research was to apply multimodal learning analytics in order to systemically investigate college students’ attention states during their collaborative problem-solving (CPS) in online settings. Existing research on CPS relies on self-reported data, which limits the validity of the findings. This study looked at data in a systemic manner by collecting and analyzing multimodal data including electroencephalogram data, knowledge tests and video recordings. The study found students’ attention was positively correlated to their knowledge gains. Also, students’ attention varied across different conditions of collaborative patterns as the highest attention level was recorded in the centralized condition. A hidden Markov model was then applied to explain the difference across various conditions by identifying both the hidden states and the transitions among the states during CPS. The findings of this research advanced theoretical insights and provided practical implications on understanding and supporting CPS in online college-level courses.