Wpływ obrazu nauczyciela w edukacyjnych materiałach wideo na przyswajanie wiedzy
PhD Student
Nicolaus Copernicus University
I am a PhD student at Nicolaus Copernicus University in Toruń, Poland, working at the intersection of AI in education, distance learning, and linguistics.
My research focuses on how people learn from digital materials—especially educational videos and language-learning resources. In the Pearls of Science project presented on this page, I examined how different instructor-on-screen formats influence learning and attention in beginner Mandarin vocabulary learning.
In parallel, I am developing speech-recognition approaches that could provide learners with automated feedback on both sound and tone correctness in Mandarin Chinese, with the longer-term goal of supporting scalable online learning.
Educational videos are central to blended learning, flipped classrooms, and MOOCs. Yet, the field still debates whether showing the instructor helps learning.
There is a theoretical conflict between the Cognitive-Affective-Social Theory (CASTLE), which suggests a visible teacher enhances motivation, and Cognitive Load Theory (CLT), which argues a visible teacher acts as a distraction.
This research project investigates this conflict in the specific context of Mandarin Chinese vocabulary learning. By utilizing eye-tracking technology, we measured visual attention patterns to determine if the "talking head" format helps students learn tonal languages or simply splits their attention.
We synthesized prior work on instructor presence in educational videos, focusing on learning outcomes, attention processes, and cognitive load-related findings.
We prepared three video variants (no instructor / continuous / intermittent) and conducted a matched-groups, two-session study with eye tracking and speaking tests.
We linked learning outcomes (speaking test) with objective gaze data to assess whether instructor presence changes performance, attention allocation, and cognitive-load indicators.
We tested whether a cloud speech-to-text service could support automatic scoring of recorded responses. Agreement with expert ratings was moderate and uneven across participants.
We did not observe a statistically meaningful advantage of continuous or intermittent instructor presence over the no-instructor format for short-term vocabulary learning outcomes.
Across formats, learners primarily attended to Pinyin. The instructor tended to be a secondary visual element, supporting the idea that attention shifts do not automatically lead to better performance.
Eye-tracking-based indicators did not provide strong evidence that any instructor format consistently reduced cognitive processing demands compared to the others in this specific lesson setting.
For short vocabulary videos, an instructor-absent design is a robust default: it maintains effectiveness while reducing production complexity and simplifying post-production edits.