This mixed-methods study examines how pre-service teachers select instructional videos on YouTube for physics teaching. The study focuses on the role of surface features that YouTube provides (e.g., likes, views, thumbnails) and the comments underneath the videos in the decision-making process using videos on quantum physics topics as an example. The study consists of two phases: In phase 1, N = 24 (pre-service) physics teachers were randomly assigned to one of three groups, each covering a different quantum topic (entanglement, quantum tunneling, or quantum computing, respectively). From eight options provided, they selected a suitable video for teaching while their eye movements were tracked using a stationary eye tracker in a laboratory setting, and think-aloud data was collected. In the subsequent phase 2, participants were allowed to freely choose one YouTube video on a second topic of the above-mentioned ones while thinking aloud. The results reveal a significant emphasis on video thumbnails during selection, with over one-third of the fixation time directed towards them. Think-aloud data confirms the importance of thumbnails in decision-making, e.g., as evidenced by a categorization of the study participants’ arguments and thoughts voiced. A detailed analysis identifies that participants did not rely on (content-related) comments despite they have been found to be significantly correlated with the videos’ explaining quality. Instead, decisions were influenced by surface features and pragmatic factors such as channel familiarity. Retrospective reflections through a questionnaire including rating scale items support these observations. Building on the existing empirical evidence, a decision tree is proposed to help teachers identify high-quality videos considering duration, likes, comments, and interactions. The decision tree can serve as a hypothesis for future research and needs to be evaluated in terms of how it can help systematize the process of selecting high-quality YouTube videos for science teaching.

这项混合方法研究探讨了职前教师如何选择 YouTube 上的物理教学教学视频。该研究重点关注 YouTube 提供的表面特征（例如点赞、观看次数、缩略图）以及视频下方的评论在决策过程中的作用，以量子物理主题的视频为例。该研究分为两个阶段：在第一阶段，N = 24 名（职前）物理教师被随机分配到三组之一，每组涵盖不同的量子主题（分别是纠缠、量子隧道或量子计算）。他们从提供的八个选项中选择了合适的教学视频，同时在实验室环境中使用固定眼动仪跟踪他们的眼球运动，并收集有声思考数据。在随后的第二阶段中，参与者可以自由选择一个关于上述主题的第二个主题的 YouTube 视频，同时大声思考。结果显示，在选择过程中，人们非常重视视频缩略图，超过三分之一的注视时间都集中在视频缩略图上。有声思考数据证实了缩略图在决策中的重要性，例如，研究参与者的论点和想法的分类就证明了这一点。详细的分析表明，参与者并不依赖（与内容相关的）评论，尽管这些评论被发现与视频的解释质量显着相关。相反，决策受到表面特征和渠道熟悉程度等实用因素的影响。通过包括评级量表项目的调查问卷进行的回顾性反思支持了这些观察结果。基于现有的经验证据，提出了一个决策树，以帮助教师考虑时长、点赞、评论和互动来识别高质量视频。决策树可以作为未来研究的假设，并且需要根据它如何帮助系统化选择高质量 YouTube 视频进行科学教学的过程进行评估。