Scholarlink

<p>This study examines how video duration—specifically short-form (under 1 min) versus long-form (over 10 min)—affects learner engagement and sentiment in YouTube-based educational content. Drawing on 623 videos retrieved via the YouTube Data API using the search term “Business English”, we analyzed how learners interact with content across different video lengths. Engagement was operationalized using comment frequency. Thematic and affective dimensions of viewer responses were assessed through Latent Dirichlet Allocation (LDA) topic modeling and deep-learning BERT-based sentiment analysis. A generalized linear model (GLM) revealed that long-form videos tend to generate significantly more user comments, even after controlling for view count and the number of likes. LDA results showed that long-form videos elicit more content-focused discussions and expressions of gratitude, whereas short-form videos prompt a higher proportion of evaluative feedback. Sentiment analysis indicated that comments on long-form videos were more positively skewed, while short-form content attracted a greater share of neutral and critical sentiment. These findings suggest that long-form videos support deeper engagement and positive reinforcement, while short-form videos facilitate quick evaluation and immediate learner reactions. The study contributes to a growing understanding of asynchronous video learning by highlighting how duration influences learner interaction and emotional response. Implications are discussed for educational content design in digital learning environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Duration matters: comparing learner engagement and sentiment in short- and long-form business English educational videos on YouTube

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月21日

查看全文
  • HO YOUNG YOON,
  • Sungyeun Kim

摘要

This study examines how video duration—specifically short-form (under 1 min) versus long-form (over 10 min)—affects learner engagement and sentiment in YouTube-based educational content. Drawing on 623 videos retrieved via the YouTube Data API using the search term “Business English”, we analyzed how learners interact with content across different video lengths. Engagement was operationalized using comment frequency. Thematic and affective dimensions of viewer responses were assessed through Latent Dirichlet Allocation (LDA) topic modeling and deep-learning BERT-based sentiment analysis. A generalized linear model (GLM) revealed that long-form videos tend to generate significantly more user comments, even after controlling for view count and the number of likes. LDA results showed that long-form videos elicit more content-focused discussions and expressions of gratitude, whereas short-form videos prompt a higher proportion of evaluative feedback. Sentiment analysis indicated that comments on long-form videos were more positively skewed, while short-form content attracted a greater share of neutral and critical sentiment. These findings suggest that long-form videos support deeper engagement and positive reinforcement, while short-form videos facilitate quick evaluation and immediate learner reactions. The study contributes to a growing understanding of asynchronous video learning by highlighting how duration influences learner interaction and emotional response. Implications are discussed for educational content design in digital learning environments.