Using Theory to Drive Multimodal Research on Learning

Using Theory to Drive Multimodal Research on Learning

📅 Friday 22/05 11:00-12:30h

📍 Workshop Space B

🔎 Needs Analysis

Research on learning processes has increasingly moved beyond traditional methods such as self-reports toward multimodal approaches that integrate diverse data sources (e.g., linguistic data, behavioral traces, interaction logs) to capture learning as it unfolds (Järvelä & Hadwin, 2024). However, significant methodological challenges remain in multimodal data collection and analysis (Sharma & Giannakos, 2020). These include identifying which data modalities capture specific learning processes, aligning theoretical constructs with observable data traces, analyzing heterogeneous data types, and determining how and when to triangulate modalities (Azevedo & Gašević, 2019). Multimodal approaches are most meaningful when modalities are theoretically aligned and jointly interpreted to support defensible inferences about learning. This workshop guides PhD researchers in designing, analyzing, and interpreting multimodal studies in technology-enhanced learning (TEL).

📒 Session Description

Part I: Designing Multimodal Studies in TEL
This part focuses on multimodal data collection and study design. Participants explore how theory guides the selection of data modalities to capture learning processes in TEL. Examples include linguistic data (e.g., written responses, revisions, AI prompts) reflecting cognitive and metacognitive processes; behavioral traces such as clickstreams and navigation patterns (Gašević et al., 2015); interaction data such as feedback uptake or human–AI dialogue (Wise et al., 2014); and temporal data capturing process sequences over time. Participants map learning constructs to observable indicators.

Part II: Analyzing and Interpreting Multimodal Data
This part focuses on analysis and interpretation. Topics include construct validity, triangulation, multimodal analytics, risks of over-interpreting trace data, and inference boundaries.

💡 Learning Objectives

At the end of this workshop, participants will be able to:

• Design theory-informed multimodal studies by selecting appropriate
  data modalities and instrumentation tools to capture learning
  processes.
• Align theoretical constructs with multimodal data traces, identifying
  which observable indicators meaningfully represent learning
  behaviors.
• Analyze and interpret multimodal datasets using triangulation
  strategies and computational approaches while maintaining
  construct validity.

🧩 Pre-activities and prerequisites

There are no specific prerequisites for the workshop. PhD researchers who have had exposure to collecting and/or analyzing more than one data modality in previous or current work are particularly welcome. Prior to the workshop, participants may receive some pre-reading materials (e.g., 1-2 papers) about multimodal approaches to investigating learning.