Department of Media and Communications, Faculty of Arts & Social Sciences, University of Malaya, Kuala Lumpur, Malaysia.
Faculty of Information Science and Technology, Universiti Kebangsaan Malaysia, Bangi, Malaysia.
PLoS One. 2023 Oct 18;18(10):e0287367. doi: 10.1371/journal.pone.0287367. eCollection 2023.
Social media has been tremendously used worldwide for a variety of purposes. Therefore, engagement activities such as comments have attracted many scholars due its ability to reveal many critical findings, such as the role of users' sentiment. However, there is a lacuna on how to detect crisis based on users' sentiment through comments, and for such, we explore framing theory in the study herein to determine users' sentiment in predicting crisis. Generic content framing theory consists of conflict, economic, human interest, morality, and responsibility attributes frame as independent variables whilst sentiment as dependent variables. Comments from selected Facebook posting case studies were extracted and analysed using sentiment analysis via Application Programme Interface (API) webtool. The comments were then further analysed using content analysis via Positive and Negative Affect Schedule (PANAS) scale and statistically evaluated using SEM-PLS. Model shows that 44.8% of emotion and reactions towards sensitive issue posting are influenced by independent variables. Only economic consequences and responsibility attributes frame had correlation towards emotion and reaction at p<0.05. News reporting on direction towards economic and responsibility attributes sparks negative sentiment, which proves that it can best be described as pre-crisis detection to assist the Royal Malaysian Police and other relevant stakeholders to prevent criminal activities in their respective social media.
社交媒体在全球范围内被广泛用于各种目的。因此，评论等参与活动吸引了许多学者的关注，因为它能够揭示许多关键发现，例如用户情绪的作用。然而，对于如何通过评论来检测基于用户情绪的危机，还存在一个空白，为此，我们在本研究中探索了框架理论，以确定用户情绪在预测危机中的作用。通用内容框架理论包含冲突、经济、人类利益、道德和责任属性框架作为自变量，而情感作为因变量。从选定的 Facebook 发布案例研究中提取评论，并使用情绪分析应用程序接口 (API) 网络工具进行分析。然后使用积极和消极情感量表 (PANAS) 对评论进行进一步分析，并使用 SEM-PLS 进行统计评估。模型表明，44.8%的情感和对敏感问题发布的反应受到自变量的影响。只有经济后果和责任属性框架与情感和反应相关（p<0.05）。新闻报道对经济和责任属性的方向引发了负面情绪，这证明它可以最好地描述为危机前检测，以协助马来西亚皇家警察和其他相关利益相关者在各自的社交媒体上预防犯罪活动。
