Abstract
The year 2021 marked the 10th anniversary of the Great East Japan Earthquake (GEJE), which caused a meltdown in the Fukushima Daiichi nuclear power plant. News coverage related to the GEJE’s 10th anniversary centered around March 2021. This study identified the discourses prominent in the Japanese sphere of the social networking service Twitter from February to April 2021 and used quantitative content analysis to investigate the extent to which information from legacy media was referred to in these discourses. Employing content analysis and human coding of supervised data, an open-source machine-learning model for natural language processing evaluated the content of ∼2 million tweets, which were then classified into eight categories. Memorial and remembrance messages related to the GEJE, information related to disaster preparedness (from both institutional and individual sources), and political discourses on reconstruction (involving situations related to nuclear accidents) were extracted as the major narratives. The temporal variation of memorial messages varied considerably, and while the recall of people’s memories may be due to media events, the information presented by the media was rarely used. The disaster preparedness category’s content remained consistent throughout the study period. Legacy media information was quoted heavily, mainly in political messages, and had a high retweet rate. We also observed a tendency for certain political websites to be linked in numerous tweets promoted by a small number of people. There was a distinct division between users citing conservative and liberal media sources, and it can be inferred that political discourse on nuclear power plant policies was characterized by an ongoing, irreconcilable stalemate. In political discussions, the media appeared to provide an agenda for society, stimulating people’s political discourse. This study empirically identifies legacy media’s uneven impact on public opinions on each topic and discusses how this relates to collective memory formation.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have