社会研究中心系列讲座第55期（线上）

【主题】Social Big Data Analysis of Future Signals for Bullying in South Korea: Application of General Strain Theory（韩国霸凌现象预测的社会大数据分析）

【主讲人】Yoonsun Han（韩国首尔国立大学社会福利系助理教授）

【时间】2020年9月8日9:30—11:00

【地点】zoom线上

【工作语言】英文

【主办单位】北京大学社会研究中心北京大学社会科学学部

The utilization of traditional social survey data to approach today’s bullying problems presents some limitations. In response, a new approach to investigate and subsequently intervene is required for understanding the bullying phenomenon. Therefore, this study analyzed the big data generated by social media to identify Future Signals of bullying that can more effectively clarify the problem and suggest targeted interventions to address the bullying phenomenon in South Korea. Also revealed were topics that may provide future insight regarding social circumstances that require further public attention to address the bullying phenomenon in South Korea. A new approach to investigate and subsequently intervene is required for understanding the bullying phenomenon. By using social big data analysis, 350,314 web documents were collected per hour each day from January 1, 2013 to June 30, 2017, from 279 subject channels based on an ontology of bullying-related topics. Term frequency, document frequency, degree of visibility, and degree of diffusion were computed to identify Future Signals. A substantial overlap of findings between studies based on social big data and traditional survey results was observed for family (e.g., parental divorce, domestic violence, child abuse), peer (e.g., transfer, friend violence), economic (e.g., economic problem), and strain domains, whereas strains concerning the media (e. g., movie, celebrity) and cultural (e.g., materialism, hell Korea) domains seemed to be more salient in social big data. Weak Signal topics in social big data representing media and cultural strain domains (e.g., Youtube, class society, bullying culture) related to the bullying phenomenon appear to be emerging in significance. These topics and their respective strain domains represent potentially important new areas that warrant further investigation by practitioners and policymakers. These findings may allow the early detection of crucial information by providing data to support better informed insight and intervention related to the complex problem of bullying in South Korea.