Topic Modeling and Semantic Network Analysis of Themes and Trends of News Reports on Drugs in South Korea

Abstract

Taking into account the news media’s agenda-setting and framing functions, this study examined the prevalent themes and topical focus of Korean news reports about drug-related issues. Using big data analysis techniques such as topic modeling and semantic network analysis, we reviewed a total of 124,969 articles provided by 49 media outlets from 2010 to the present. These articles included 43,524 on illegal drugs, 58,146 on psychotropic pharmaceuticals, and 23,050 on marijuana. Topic modeling results show that over the past 14 years, coverage of drug-related issues has tended to increase with each occurrence of a drug scandal involving celebrities. Different topics were extracted across types of drugs: international problems concerning illegal drugs, drug abuse, treatment and prevention for psychotropic pharmaceuticals, and legalization and industry for marijuana. Notably, semantic network analysis indicates that, even in news stories about psychotropic pharmaceuticals, keywords such as 'treatment' and 'hospital' had low centrality and relational significance. Such a trend suggests that news stories lack adequate discussion of addiction and treatment. Based on our findings, we argue that media coverage should shift from anecdotal frames, which focus mainly on individual drug incidents involving celebrities, to more comprehensive thematic frames that provide accurate information about drug use and address the systemic and preventive dimensions of drug-related issues. Furthermore, we propose that the government should devote efforts not only to supply-control policies that target and criminalize the illegal drug trade but also to demand-control policies that facilitate drug use prevention, management, and treatment.