时间：2023年9月22日（星期五）15:30-16:30
地点：经管大楼A楼 四楼第二会议室报告厅
主题：描述你的音频专辑：探索文本特征对在线学习者参与行为的影响（Describing Your Podcasts: Disentangling the Effects of Linguistic Features in Bolstering Online Learners’ Engagement with Podcasters）
主讲人：刘晓慧（上海理工大学管理学院）
简介：刘晓慧，博士毕业于上海交通大学安泰经济与管理学院，丹麦哥本哈根商学院（Copenhagen Business School）联合培养博士，主要研究兴趣包括用户行为、平台经济、人机交互、人工智能与机器学习等。在Information System Frontiers、Technology inSociety、Industrial Management& Data Systems等国际权威期刊发表SCI/SSCI论文5篇，并参与多次国际学术会议，发表信管领域顶级国际会议论文6篇。
Xiaohui Liu have graduated with a doctoral degree fromthe Antai School of Economics and Management at Shanghai Jiao Tong Universityand co-trained with the Copenhagen Business School in Denmark. Her mainresearch interests include user behavior, platform economy, human-computerinteraction, artificial intelligence, and machine learning. She has published 5papers on authoritative SCI/SSCI journals, such as Information SystemFrontiers, Technology in Society, Industrial Management & Data Systems. Shealso has participated in many international academic conferences and published6 top-level international conference papers in the IS field.
摘要：尽管音频主播之间存在激烈的竞争关系，但目前的文献缺乏研究描述音频专辑的文本特征在促进学习者参与音频主播作品方面的作用。基于认知评价理论，我们不仅假设了客观文本特征（即词汇丰富度和主观性）与在线学习者参与度之间的曲线关系，而且进一步假设可读性这一主观文本特征对上述关系具有调节作用。为了实证验证我们的假设，我们使用自然语言处理算法从中国领先的音频平台上获得的2280个教育主播中提取描述音频专辑的三个文本特征。分析结果表明，词汇丰富度/主观性与参与度之间存在倒U型关系。对调节效应的分析进一步表明，可读性使词汇丰富度与参与度之间的倒U型关系变的陡峭，而使主体性与参与度的关系变的更平坦。
Despite fiercecompetition among podcasters, there is a dearth of research that has sought toelucidate the role of podcast description’s linguistic features in fosteringlearners’ engagement with podcasters on audio platforms. Building on cognitiveappraisal theory, we not only posit curvilinear relationships between objectivelinguistic features (i.e., lexical richness and subjectivity) and onlinelearners’ engagement, but we further postulate readability, a subjectivelinguistic feature, as having a moderating influence on the abovementionedrelationships. To empirically validate our hypotheses, we employ NaturalLanguage Processing algorithms to extract the three linguistic features ofpodcast descriptions from 2280 educational podcasts obtained from a leadingaudio platform in China. Analytical results point to inverted U-shapedrelationships between lexical richness/subjectivity and engagement. Analysis ofmoderating effects further revealed that readability steepened the invertedU-shaped relationship between lexical richness and engagement while flatteningthe relationship between subjectivity and engagement.
