ScholarWorks@숙명여자대학교

초록

This study aimed to analyze the overall trends and specific characteristics of multimodal data used in recent empirical studies focused on measuring collaborative competencies through Multimodal Learning Analytics (MMLA). To achieve this research objective, a systematic literature review methodology was applied, and 29 studies published in English from 2022 to 2024 in the Scopus database were selected as the final dataset. The analysis framework involved coding and categorizing variables such as publication year, research subjects, instructional environments, types of multimodal data, and analytical methods. The main findings indicated that multimodal collaborative research primarily involved small groups (2-4 participants), mainly consisting of college students. Various learning environments, such as simulation-based and game-based online platforms, were utilized. Key components analyzed in collaborative competencies included verbal and non-verbal communication, emotional and cognitive aspects, and roles and participation. Among the multimodal data types, discourse and language data were most frequently employed. Recent studies showed a particular trend toward integrating multiple data sources (e.g., discourse, audio, facial expressions, and physiological data) to conduct more in-depth analyses of collaborative states. Analytical methods used encompassed statistical analyses, predictive modeling, sequence analysis, and network analysis. Based on the findings, this study identified limitations in current multimodal collaborative research and suggested directions for future studies.

키워드