Multimodal learning analytics in computer-supported learning

Abstract:

MMLA in CSCL investigates the collaborative process and the learning outcome using multimodal learning analytics (MMLA) in a CSCL (computer-supported collaborative learning) practice e.g., knowledge-building activities. In this empirical research, we adopt the concept of evidence-centered design to develop new mode of formative assessments that can be embedded and used in collaborative learning environments.

About the project

MMLA in CSCL will employ multimodal technology to uncover visible and invisible dynamics of collaboration and interaction patterns. This project is executed in two phases over a 3-year period from Sept 2019 to Aug 2022:

Phase I: Design, Enact and Analyse: Identify collaborative activities in Knowledge Building classroom. Explore dispositional, discourse and social analytics to triangulate physiological data and process data in learners’ interaction and discourse.
Phase II: Engage and Sustain: Engage practitioners through a series of teachers’ professional development & training. Design and enact a series of teachers’ professional development courses to engage teachers in sustained CSCL practice with MMLA for formative assessment.

Approach

Using mixed-methods study of a single CSCL case study (with embedded units in primary, secondary, and tertiary), we focus on studying each collaborative scenario in the embedded case study in a series of design research harnessing advanced multimodal technology:

Ethnography approach integrated with Design-based approach focusing on each collaborative scenario, e.g., the interaction between teacher and students or among students;
Evaluate different types of multimodal indicators, multimodal data and analyses; as well as feedback types (real-time or post-hoc).

Expected Outcomes

Following are the projected key deliverables:

Enrich the semantics with learners activity traces from dispositional & discourse analytics;
Enable triangulation of indicators of collaboration from multiple data sources: text, video, data logs, audio, physiological;
Enable real time and post-hoc feedback in a multimodal way;
Engage teachers in practicing, interpreting and analysing CSCL with MMLA.