This study examined how English-language newspapers in South Korea depicted "social distancing" during the COVID-19 pandemic, using text mining techniques such as word-correlation analysis and topic modeling. The data consisted of 2,463 news articles published from January 20, 2020 to January 13, 2021 in the two major English-language newspapers in South Korea -The Korean Times and The Korea Herald. For analysis, the researchers extracted sentences that contained "social distancing" to make a

more »

... rpus. The corpus contained 4,566 sentences. For topic modeling analysis, this study used Latent Dirichlet Allocation (LAD) which views each document as a mixture of topics, then considers each topic a mixture of words. The word-correlation was calculated at the sentence-level, not at the article-level. In other words, the correlation values between "social distancing" and other words were calculated based on how frequently "social distancing" and the words co-occurred in the same sentences. In the analysis, the data was grouped according to five time-periods that were proposed by the Korea Disease Control and Prevention Agency (KDCA) based on the major changes in COVID-19 situations. Results indicated that media coverage related to "social distancing" focused on two distinctive topics: "spread and prevention" and "social impact." Also, the word-correlation analysis showed that "social distancing" was more associated with "spread and prevention" than with "socialimpact" of the pandemic. The findings of the study suggest a couple of implications: (1) The paradoxical social meaning of "social distancing" is apparent in the media depictions of COVID-19 (2) The media depiction of "social distancing" needs to provide more in-depth interpretations of "social distancing" concerning its impact on society.