Internal: Create Two, Remember More – Brief Learner-Generated Multimodal Video Creation Boosts L2 Vocabulary Retention

SEMINAR
Date: Thursday 23 April 2026
Time: 14:00 - 15:30
Location: P473-474

How can Chinese learners of English as a foreign language – EFL – use videos to learn new words? Video-based methods may lead to longer-lasting vocabulary learning, as visiting PhD student Chao Xu shows in his study.

This study tests whether brief learner-generated multimodal production improves L2 vocabulary retention beyond viewing the same materials and beyond teacher-led instruction. 164 Chinese university EFL learners learned 39 unfamiliar B2–C1 English words in one of three conditions: (1) video creation: each learner created only two short videos (under one minute each) for two assigned words, combining on-screen text and images with spoken narration and embodied cues (e.g., gestures and facial expressions) in situational contexts (e.g., occasional role-play or short skits); (2) video viewing: learners watched videos but did not create any; and (3) traditional instruction: teacher-led pronunciation, explanations, and examples. In class, both video groups learned all 39 words by watching the same instructor-selected 39 student-generated videos (one per word), while the traditional group received teacher-led explanations. In-class instructional time was matched across all three conditions.

Retention was measured with immediate and 5-day delayed multiple-choice tests of form–meaning recognition. Items self-reported as known prior to instruction were excluded. To minimise item-specific production effects, we distinguished self-produced from non-produced items and ran the primary analyses on non-produced words (i.e., words creators did not produce and learned in class through viewing). Trial-level accuracy was analysed using binomial generalized linear mixed-effects models (GLMMs) with random intercepts for participants and items, controlling for English proficiency and working memory.

Results showed a robust multimodal advantage over traditional instruction and an incremental creation advantage. On non-produced items, creators outperformed viewers at both time points (Immediate: 82.6 percent vs 73.2 percent; Day 5: 73.8 percent vs 59.2 percent), and both video groups exceeded traditional instruction (Immediate: 50.9 percent; Day 5: 30.9 percent). Forgetting was attenuated in the video conditions, especially in the creation condition. These findings suggest that, even with minimal production, learners may adopt a more generative, embodied processing stance during subsequent viewing, supporting more durable form–meaning mappings than passive viewing alone.

About Chao Xu

Chao Xu is currently a doctoral student at the University of Padua and a Lecturer in English at Chaohu University in China. His research focuses on digital pedagogy, multimodal learning, and learner engagement in language education. In his doctoral research, he explores how creative digital tasks can support language teaching and learning in higher education. He will be a visiting scholar at SU in April, 2026.


Last updated: 2026-04-09

Source: Department of Teaching and Learning