Article overview
Educational video transcripts contain rich information about concepts, people, organizations, and relationships, but turning that material into structured data and interpretable qualitative evidence is difficult at scale. This article examines how large language models can support two connected tasks: improving entity recognition and extending qualitative insight from transcript data.
The work contributes to a broader research agenda on responsible, transparent uses of LLMs for educational data analysis. It focuses on what language models can add to transcript-analysis workflows while keeping human interpretation central.
My role
I am the first author of the article, working with Cody Pritchard and Joshua Rosenberg.
Citation
Wang, W., Pritchard, C., & Rosenberg, J. (2026). Analyzing educational video transcripts with LLMs: Accuracy, named entities, and interpretability. Journal of Educational Data Mining, 18(2), 143–168. https://doi.org/10.5281/zenodo.22689987
Publication details
- Journal: Journal of Educational Data Mining
- Status: Published
- Publication date: September 10, 2026
- Volume, issue, and pages: 18(2), 143–168
- Authors: Wei Wang, Cody Pritchard, and Joshua Rosenberg
- Section: Human-AI Partnership for Qualitative Analysis
- DOI: 10.5281/zenodo.22689987
- Official article: JEDM publication record