演講主題: Effect of the Rounded Rating Display on Online Rating and Review Behavior
主 講 人: 劉 唱，香港中文大學商學院信息系統專業博士生
主 持 人: 趙學鋒，管理科學與信息管理系教授
網絡直播平臺:騰訊會議，會議ID：510 780 222
劉唱，女，1996年生，2016年本科畢業于華中科技大學管理學院信息管理與信息系統專業，現為香港中文大學商學院信息系統專業博士生。主要研究領域包括互聯網用戶行為和口碑傳播，人工智能，社交媒體和社會網絡，大數據分析和價值挖掘等方面。主要研究方法包括數據挖掘，深度學習，自然語言處理，社會網絡分析，應用計量經濟學等。以一作及通訊作者身份在Pacific Asia Conference on Information Systems (PACIS)上發表會議論文2篇。受邀參加PACIS，SCECR（Statistical Challenges in eCommerce Research），ICEC（International Conference on Electronic Commerce）等國際學術會議共6次?，F有數篇論文正在Information Systems Research（UTD24），Journal of Marketing（UTD24）等期刊審稿中。
As online review platforms have allowed people to easily access others’ opinions, potential consumers use the information displayed on these platforms to form pre-purchase expectations. In order to help consumers quickly categorize service providers into multiple levels (such as the 5-star categorization), it is now common for platform providers to display the rounded average rating. However, this rounding strategy may create a discrepancy between pre-purchase expectations and the experienced true quality when the displayed rating is different from the true average rating. Such a discrepancy may affect the consumer’s satisfaction level, which would eventually influence subsequent rating decision and review behavior. This study aims to examine the causal impact of the display of the rounded rating on subsequent rating decisions and the provision of emotional text in a subsequent online review, especially when the displayed rating differs from its true average. Using online review data collected from Yelp, we utilize a regression discontinuity design by taking advantage of Yelp’s rounded rating display. Our primary results suggest that a rounded-up rating (compared with a rounded-down rating) leads to a sharp decrease in subsequent ratings, as well in as the provision of pleasant emotions, but a sharp increase in the provision of unpleasant emotions. Furthermore, we examine the moderating effect of prior reviews’ rating dispersion and information diversity. When prior ratings are consistent, or when the content in prior reviews covers diverse topics, the main impact of the displayed rating would be intensified.