Abstract
This study explores how South Korean newspapers reported the issue of AI (avian influenza) by employing framing, and the concepts of media advocacy and mobilizing information (MI). Results indicate that news stories were more likely to attribute blame to the government. Government, scientist/researcher, and farmer sources were most prevalent in news coverage. Mentions of tactical MI for the preventive actions increased. Overall, findings indicate the increased media advocacy efforts during repetitive outbreaks of AI.
After the first avian influenza (AI) of 2016 was reported to the Ministry of Agriculture in South Korea on October 28, more than 30 million poultry from 800 farms were culled due to the worst bird flu epidemic across the country (“Avian influenza,” 2017). During the 3 months between November 2016 and February 2017, more than 20% of poultry from commercial poultry farms in South Korea were killed (Park, 2017). It was considered the worst outbreak of AI in South Korea.
AI is an infectious disease of birds caused by the influenza virus. AI was first identified in 1996 in China. The following year, AI had infected humans in Hong Kong (Chen et al., 2005). Since 2003, 256 human cases have been confirmed (Shih, Wijaya, & Brossard, 2008). In recent years, a severe outbreak in poultry has been detected in Southeast Asia due to a particular virus, H5N1, and has become a major threat to humans (Chen et al., 2005). Although human-to-human transmission has not been confirmed, direct contact with infected poultry causes human infections (Shih et al., 2008). The highly pathogenic H5N1 virus has become pandemic in poultry in Southeast Asia, and this has become a threat to humans (Chen et al., 2005).
As the outbreak of AI represents an increasingly important public health issue in South Korea, this study explores how the issue has been reported in news media. More specifically, this study examines how new media helped identify problems, stimulate public health awareness on the issue, and advance a solution during the AI crisis. Due to the high risk of human infection and economic loss caused by the AI outbreak, news media attention to AI has increased in accordance with the level of public health risk and economic loss caused by AI. Analyzing news stories of AI seems to be especially important at this early stage of developing into a pandemic in the country, given the important role of news media in influencing readers’ reactions to a perceived health crisis (Glik, 2007). In addition, while research examines news coverage on pandemic diseases, much research has focused on examining how news media dramatized the issue rather than helping to alleviate the public health crises (Klemm, Das, & Hartmann, 2016).
This study content analyzed news articles from national newspapers published between 2003 and 2017 during the seven specific time periods when AI broke out. Considering that news media can affect policy change by informing policy makers and stakeholders to better understand issues, (S. H. Kim, Thrasher, Kang, Cho, & Kim, 2017), this study’s purpose is to examine how South Korean newspapers have addressed this chronic contagious disease and helped facilitate policy changes by employing framing and practices of media advocacy and mobilizing information (MI).
Although previous studies on public health emergencies have considered news media as a communication tool for messages of warning, alert, and information (Kittler, Hobbs, Volk, Kreps, & Bates, 2004; Lin, Jung, McCloud, & Viswanath, 2014), this study sees news media not only as a communication component during health crises, but also as a potential advocate for health-related policy change (Len-Rios et al., 2009). Furthermore, this study can provide a useful case study, as South Korea has updated contagious disease control measures and regulations on public health at the national level due to a series of outbreaks of pandemic diseases such as MERS in 2015 and AI in 2016.
This case study may help us understand how news media play a role in advancing a solution during public health crises. Given repetitive outbreaks of AI in South Korea during the last 15 years, this study examines actors to whom news media attributed blame for the outbreaks, sources used in newspaper coverage of AI, and what causes and solutions (e.g., MI), if any, were offered. This study examines the evolution of these attributions over the lifetime of AI outbreaks in South Korea.
Literature Review
Framing Avian Flu in the News
Since the first outbreak of AI in 2003, scholarly investigations on news coverage of AI have been somewhat inconsistent. Some studies have focused on examining sensationalism in news stories (Dudo, Dashlstrom, & Brossard, 2007; Klemm et al., 2016). Others identified frame-building factors of news coverage and focused on journalists’ role as public health advocates in pandemic news coverage (Lee, 2014; Lee & Basnyat, 2013).
Analyzing previous studies about news coverage on swine flu by focusing on three indicators—volume of media coverage, mention of threat versus means of protection, and tone of coverage—Klemm et al. (2016) found that global news coverage of swine flu was more likely to treat the issue in a sensational way through excessive media coverage and frequent mentions of threats rather than mentioning means of protection. Analyzing 360 news articles in U.S. newspapers between 2000 and 2006, Dudo et al. (2007) similarly found that AI news was presented predominantly with episodic frames. Considering that episodic frames emphasize the emotional side surrounding the issue (Nitz & West, 2004), this study indicated that U.S. news media are more likely to present AI in a dramatized way.
With specific regard to frame building of AI, Lee (2014) found that news media are more likely to adopt news releases with a preventive, thematic, emotional appeal, gain framing, and a positive tone during an outbreak situation. Although previous studies indicate news media coverage during public health crises typically rely upon government sources and information, Lee and Basnyat (2013) studied how journalists selectively adopt press releases and produce news stories. Lee and Basnyat (2013) argued that pandemic news stories include diverse frames such as emotional appeals, thematic framing, and a positive tone, which were not apparent in original press releases from government entities. Lee (2014) and Lee and Basnyat (2013) showed the independent role of journalists in producing news during public health crises. During emergencies, journalists are more likely to consider themselves public health advocates or audience advocates, and their framing of news may align more with reducing the impact of the pandemic on the audience (Lee & Basnyat, 2013; Len-Rios et al., 2009).
Avian Flu in South Korea
Over the last 15 years, AI has broken out 7 times in South Korea (Ministry of Agriculture, Food and Rural Affairs [MAFRA], 2017a; see Figure 1). Although no human infections were reported in South Korea, outbreaks of AI have caused enormous economic loss and health threats to society whenever the issue arises (“Unfinished outbreak of AI,” 2017). Avian flu viruses differ distinctly genetically, depending on their hosts and geographical origin (Gauthier-Clerc, Lebarbenchon, & Thomas, 2007). Although more cases of low pathogenic AI, which represents low infectious hazards to humans, were found than high pathogenic AI in South Korea (H. Kim, 2012), prevention and quarantine measures against AI have always been challenging. Figure 1 shows the economic loss from AI over time, in terms of culled poultry and money spent during each outbreak.
From “Avian Flu in South Korea,” by the Ministry of Agriculture, Food and Rural Affairs, 2017a, in the public domain
The 7th time AI broke out in South Korea was in October 2016. Before March 2017, the government spent US$234 million to combat AI, and 32 million poultry (almost one third of poultry in the country) were culled for prevention (MAFRA, 2017b). Because the South Korean peninsula is located in the path of migratory birds, H5N1 virus is generally carried by migratory birds from other Southeast countries (H. Kim, 2012). Once the virus is transmitted to poultry in commercial farms, it continues to rapidly propagate in other regions because of the movement of domestic poultry and poultry products, and people (Gauthier-Clerc et al., 2007).
Although AI virus is simply the flu virus for migratory birds with strong immune systems, and does not lead to death, the dispersion of H5N1 virus becomes a significant threat to poultry with weak immune systems in packed factory farms, and it leads to massive death and culling (Earn, Dushoff, & Levin, 2002). For prevention measures, infected poultry and other animals in surveillance zones are necessarily culled (Gauthier-Clerc et al., 2007). Despite the constant emphasis on prevention against the virus from animal health authorities, AI continues to cause outbreaks in poultry in South Korea (H. Kim, 2012).
After the severe outbreak of AI from November 2016 to March 2017, the MAFRA in South Korea announced updated versions of measures including the regulation of factory farming, and government and farmers’ recommended actions for rapid response for outbreaks of AI in April 2017 (MAFRA, 2017b).
News Framing
News media define and frame issues and offer possible causes and solutions for the public (S. H. Kim, Carvalho, & Davis, 2010). Entman (1993) stated that news framing necessarily involves selection and emphasis of certain aspects of an issue for attention. Ultimately, news frames “can exert a relatively substantial influence on citizen’s beliefs, attitudes, and behaviors” (Tewksbury & Scheufele, 2009, p. 19). Because news media makers produce various frames, and not all of them are available to audiences when building a social reality, frames serve as an interpretation and a point of particular interest (Entman, 1993).
One component of framing research is the attribution of blame, which indicates the responsibility of causing and resolving social problems (Holton, Weberling, Clarke, & Smith, 2012). For example, Holten et al. (2012) found a pattern in recent years of news media increasingly presenting societal-level causes in framing issues. In health and risk communication, Hallahan (1999) argued that attribution of responsibility is an important frame, and that blaming individuals for health problems is common practice in American news media coverage. But South Korea may or may not be similar to the United States in this regard. Similarly, Iyengar (1991) stated that assignment of responsibility is the key factor in influencing public opinion on social issues because it ultimately influences individuals’ perceptions of acting and solving social problems.
Reynolds and Seeger (2005) wrote that, following a crisis, news media and the public become more critical and raise questions about who should take the blame and responsibility. Attributing blame to the government can be easier for journalists to create a simpler message in news stories (S. H. Kim & Willis, 2007).
Research has indicated that thematic framing is more influential in terms of policy changes (Major, 2009), but episodic framing, individual stories or “human frames” also help with media advocacy efforts and may influence policy change over time (McKeever, 2013).
Based upon a review of framing theory, the following research question was posed for this study:
In framing news stories, the use of certain sources affects advancing a specific angle or aspect of a story (Zoch & Molleda, 2006). Due to the complexity of health issues, journalists often quote expert sources when presenting health stories (Len-Rios et al., 2009). Based on these ideas, the following research question is proposed:
According to Entman (1993), framing involves the process of defining a problem, diagnosing the causes, and suggesting remedies. Identifying the causes of problems and justifying particular solutions provided by news media leads the audience to interpret issues differently (S. H. Kim, Carvalho, Davis, & Mullins, 2011). Iyengar (1991) demonstrated that framing influenced audiences’ perceptions of attributing responsibility for causing and solving the problem. Therefore, the following research question is proposed:
Media Advocacy and MI
Three decades ago, the U.S. Department of Health and Human Services (1989) defined media advocacy as the strategic use of mass media as a resource for advancing a social or public policy initiative. Wallack, Dorfman, Jernigan, and Themba (1993) referred to media advocacy as “influencing public debate and putting pressure on policymakers by increasing the volume of the public health voice and, in turn, by increasing the visibility of values, people, and issues behind the voice” (p. 2).
In general, media advocacy refers to how an issue is portrayed in the media and how this media coverage influences public opinion (Winett & Wallack, 1996). In addition, according to media advocacy, news media influence policy change by informing policy makers to better identify the issues (S. H. Kim et al., 2017). By focusing on social causes and solutions of an issue, advocates seek to emphasize issues as public health concerns (Wallack et al., 1993).
As one aspect of media advocacy, MI enables the public to act beyond simply reading news articles (Holton et al., 2012), although previous research on framing of social problems indicate that news media do not typically provide solutions (Hoffman, 2006), Lemert (1984) stated that providing MI is imperative in leading the public to act. In the case of a public health crisis, MI could enable the audience to protect themselves. For instance, a news article including an 800-number people can call or a website people can visit for more details about a health issue are examples of MI.
One approach of literature on media advocacy focuses on constructing frameworks of the media advocacy concept (Wallack & Dorfman, 1996). The other approach of research on media advocacy is to apply the concept to actual cases. Previous scholarly investigations, research on news coverage of several health issues such as sugar-sweetened beverages (Elliott-Green, Hyseni, Lloyd-Williams, Bromley, & Capewell, 2016), tobacco (Niederdeppe, Farrelly, & Wenter, 2007), and autism spectrum disorders (McKeever, 2013) illustrated how the concept of media advocacy might apply to various public health issues.
For example, Wallack and Dorfman (1996) examined media advocacy through two case studies to understand how media advocacy can affect policy change for public health issues. The first case, a California violence prevention initiative, showed how news media coverage of the issue contributed to reframing youth-related violence as an important public health issue. The other case, about Henry Horner Homes, illustrated the effective use of news media to gain legitimacy and credibility for an issue.
McKeever (2013) conducted a content analysis of news coverage of autism in U.S. newspapers from 1996 to 2006 to understand how news media coverage played a role in moving the health issue to the public agenda. Policy frames in autism stories increased while science frames decreased over time, suggesting that news media played a part in making the issue salient in terms of public health and promoting policy change. However, McKeever’s study also suggested the need for greater media advocacy from health communicators and nonprofit organizations, finding that MI was limited overall in the news articles examined.
Lemert (1981) identified three types of MI: identificational (names and contact information for people or groups), locational (time and place of an activity), and tactical (detailed instructions for certain actions). Subsequent research has also used these categories in exploring media content (e.g., Holton et al., 2012; McKeever, 2013). Our study adopted these concepts of MI to understand if South Korean media coverage included MI, which may have enabled the public to act related to AI risk and prevention. Based on this idea, the following research question was posed:
Method
In selecting the newspapers to analyze, three national daily newspapers with the largest circulation figures in South Korea (Chosun Ilbo, Joongang Ilbo, and Donga Ilbo) were selected. Due to conservative orientations of these three newspapers, the two largest liberal newspapers were also included (Kyunghyang Shinmoon and Hankyoreh Shinmoon). Among national daily newspapers, these two liberal-leaning newspapers have comparable circulation sizes to the three conservative newspapers. Political orientation of newspapers was considered because news coverage might be biased toward the government’s policies according to political orientation. Since AI was first reported in 2003, AI has broken out 7 times during the last 15 years (MAFRA, 2017a). This study’s sample included news stories from each time period of the seven outbreaks: (a) December 2003 to April 2004, (b) November 2006 to April 2007, (c) April 2008 to June 2008, (d) December 2010 to June 2011, (e) January 2014 to December 2015, (f) March 2016 to May 2016, and (g) November 2016 to March 2017. Each period was selected based on the data from MAFRA, a government division in charge of AI quarantine. To see the postcrisis effect (Reynolds & Seeger, 2005), 1 month following each outbreak was included in the final sample.
News articles were retrieved from the Korean Integrated News Database System (an online news archive) and the Internet homepage of each newspaper, using the keywords, “AI,” “Avian Influenza,” “Avian Flu,” or “Joryu Dokgam” (AI). The keyword search yielded a total of 4,319 articles about AI. An initial screening filtered out 2,016 (46.7%) duplicates and unrelated stories where AI was simply mentioned in passing or AI was used to refer to “Artificial Intelligence.” During the sampling process, coders briefly skimmed 2,303 articles. Because many news articles simply mentioned the breakout or severity of AI without more details, an additional 1,694 articles were excluded during the coding process. News articles that were long enough to include attributions of blame, causes, solutions, and/or MI were included in final samples. The remaining 609 news stories were included in the final data analysis. Analysis included all news items including feature stories about AI, but not including letters to the editor.
Coding
Two graduate student coders coded AI stories after having conducted a series of training and pilot-test sessions. Intercoder reliability was calculated by double-coding a random subsample (n = 92, 15%) of the data. Intercoder reliability corrected for agreement by chance (Krippendorff’s alpha) ranged between .81 and 1.00 with an average reliability of .93.
Table 1 shows how the coding instrument specified what may compose blamed actors, sources, possible causes and solutions, and MI. Two coders first determined the presence of actors to whom blame was attributed in each news article. For this study, a blamed actor was defined as a central actor who was frequently and dominantly blamed within an article. More specifically, there must have been at least two or three sentences criticizing an actor. Because blamed actors were mostly evident in the last paragraph, coders read the whole article to determine whether and where blame was placed. Each story was coded as having one of the five actors: government, farmer, migrant birds, immigrant employees, and others. Only the most prominent actor was coded from each story.
Coding Sheet and Intercoder Reliability
Note. AI = avian influenza.
Second, coders recorded the sources used in news stories. Any individual or organization cited for delivering information in AI articles was defined as a source. Among government, scientist/researcher, industry, small business owners, farmers, lay person, and others, each article was coded as having none, one, or more of these sources. Individuals affiliated with any government agency were coded as government. Although the same source may have been cited multiple times, it was coded only once per article.
Coders then recorded the presence or absence of possible causes and solutions for AI. Causes were categorized into government and individual causes, uncontrollable, and others. Government-level causes included policy and infrastructure factors that might contribute to the public health risk: Delayed bird flu quarantine and poor disinfection standards, compensation policy loophole, insufficient sanitation experts. Individual-level causes, on the contrary, included individual behaviors that might have caused or helped spread AI: Factory farming, farmer’s delayed report and cooperation on quarantine, unethical farmers. Uncontrollable factors were defined as the factors out of human control: Unidentifiable and mutated viruses, and others. Solutions were also divided into government-level solutions (developing AI vaccine, new law and policy enactment related to compensation and efficient quarantine, more human resources, strong and effective quarantine led by government), individual-level (responsible farmers, animal-friendly farming), and others. Each article was coded as having none, one, or more than one of these causes and solutions. No matter how many mentions were made, it was counted as one mention if they came from the same article.
Finally, coders determined the presence of three types of MI (names of individuals or organizations), locational MI (phone numbers, web links), and tactical MI (specific instructions on how to get involved beyond reading news stories; Lemert, 1984). For example, identifying information might have included naming a government agency or department that is in charge of a quarantine; a location could be a phone number of a government agency to which AI should have been reported; and a tactic might have directed the reader not to go to Southeast countries, poultry farms, or migrant bird habitats.
Results
In terms of the quantity of news coverage over time, the data indicate that the number of news articles substantially increased in the seventh time period of AI in South Korea. Although only three AI stories were published in the sixth period, the amount exponentially increased in the seventh period to 267 articles (Table 2). Then, though the amounts vary, the first, third, and fifth time periods had relatively more news coverage of AI than other remaining periods, except for the seventh period, which had nearly 3 times as much coverage as the time period with the next largest quantity of articles.
Number of Articles in Samples for Each Time Period and From Each Newspaper
Note. Duration for each time period is included in parentheses as month.
RQ 1 asked how news media have framed attribution of responsibility for AI. The government was the most frequently blamed, appearing in a total of 208 articles (34.2%). Government actors were dominant in the first (n = 21, 30.4%), third (n = 33, 42.3%), sixth (n = 2, 66.7%), and seventh periods (n = 106, 39.7%). During the third and seventh periods, in particular, the distribution of blame in the newspapers was not equal in terms of chi-square tests, and this difference was significant, third: χ2(5) = 41.84, p < .001; seventh: χ2(4) = 97.96, p < .001. Overall, findings indicate that the more damage was caused by AI, the more blame was placed on the government.
Following the government, the next most frequently blamed were migrant birds. Migrant birds were blamed in a total of 168 articles (27.6%). Migrant birds were blamed prominently in the second (n = 20, 42.6%), fourth (n = 24, 48.0%), and fifth periods (n = 41, 43.2%). During the second, fourth, and fifth periods, the distribution of blamed actors in the newspapers was not equal, second: χ2(5) = 33.29, p < .001; fourth: χ2(5) = 43.36, p < .001; fifth: χ2(4) = 61.47, p < .001. The other four actors, no blame (n = 103, 16.9%), others (n = 64, 10.5%), farmers (n = 58, 9.5%), and immigrant employees (n = 8, 1.3%), were mentioned far less frequently.
RQ 2 asked about the types of sources included in AI news coverage in South Korea. Government sources were included most frequently in news coverage of AI. The presence of government sources (n = 377, 61.9%) was significantly more frequent than other sources: scientist/researcher (n = 86, 14.1%, McNemar’s χ2 = 229.15, p < .001), farmers (n = 61, 10.0%, χ2 = 269.63, p < .001), industry (n = 19, 3.1%, χ2 = 337.16, p < .001), layperson (n = 14, 2.3%, χ2 = 349.45, p < .001), and small business owners (n = 6, 1.0%, χ2 = 367.02, p < .001).
Trends of sources were also examined over time periods. Table 3 shows the presence of sources for each period. Although the government was consistently the most often cited source, scientist/researcher was the second most presented source across the seven periods. At the same time, farmers were frequently cited in the first, third, and fifth periods, but were not mentioned at all in the fourth and sixth period. However, they increased again in the seventh period.
Presence of Sources in Each Time Period
RQ 3 asked how news coverage of AI presented possible causes and solutions. Overall, causes were more frequently mentioned than solutions, and the difference was statistically significant (t = 4.43, p < .001). Causes were presented more than solutions during the first, second, third, fifth, and seventh time periods. In the first (t = 5.094, p < .001) and third (t = 5.135, p < .001) periods, in particular, the differences between mentions of causes and solutions were statistically significant.
Of the nine causes of AI, delayed bird flu quarantine and poor disinfection standards were mentioned most often (n = 172, 28.2%). Delayed bird flu quarantine and poor disinfection standards were mentioned significantly more often than other remaining causes: factory farming (McNemar’s χ2 = 37.71, p < .001), compensation policy loophole (χ2 = 131.16, p < .001), farmer’s delayed report and cooperation on quarantine (χ2 = 50.13, p < .001), insufficient sanitation experts (χ2 = 100.51, p < .001), unidentifiable (χ2 = 71.64, p < .001), mutated viruses (χ2 = 42.39, p < .001), unethical farmers (χ2 = 86.87, p < .001). Overall, government-level causes were mentioned more than individual-level causes, but the difference was not statistically significant (t = 1.711, p = .88).
Of the seven solutions for AI, strong and effective quarantine led by the government (n = 179, 29.4%) was discussed most often, followed by new law and policy enactment related to compensation and efficient quarantine (n = 76, 12.5%), responsible farmers (n = 66, 10.8%), and animal-friendly farming (n = 60, 9.9%). Strong and effective quarantine led by the government was mentioned significantly more often than new law and policy enactment related to compensation and efficient quarantine (McNemar’s χ2 = 52.28, p < .001), responsible farmers (χ2 = 63.03, p < .001), and animal-friendly farming (χ2 = 64.16, p < .001).
Overall, government-level solutions were presented more frequently than individual-level solutions, and the difference was statistically significant (t = 9.211, p < .001). Government-level solutions were mentioned significantly more than individual-level solutions in the third (t = 6.997, p < .001) and seventh periods (t = 4.591, p < .001).
RQ 4 examined the prevalence of MI. MI appeared in 9% (n = 55) of articles in the sample. In particular, tactical MI significantly increased in the seventh period (Table 4).
Mobilizing Information in Each Period
Discussion and Conclusions
News coverage of AI increased corresponding with the severity of AI, suggesting media advocacy related to AI, a result consistent with Niederdeppe et al., (2007).
News coverage of AI was consistent with Reynolds and Seeger (2005) on the news media and public looking to assign blame after a crisis, and with S. H. Kim and Willis’s (2007) argument that it is easy for journalists to blame governments.
In presenting sources, government and scientist/researcher sources are considered “experts” by news media (McKeever, 2013). When it comes to small business owners and industry sources, it is noteworthy that these two sources decreased in news coverage of AI over time. Considering that these two source types were included in framing stories about economic loss caused by the outbreak of AI, one can conclude that news stories focused more on public health during the crises, rather than the financial loss it might cause.
Individual farmer sources outnumbering industry sources over time, and news media thus framing the issue episodically is consistent with Dudo et al.’s (2007) findings about U.S. news coverage of AI presenting mainly episodic frames.
Articles mentioned causes significantly more than solutions during the first and third periods. This is particularly related to the first outbreak of AI in the first period and the first outbreak of AI during spring in the third period. AI is considered a winter illness in the country, and it usually disappears as the temperature increases. Outbreaks of AI in the spring indicate that diseases have possibly become year-round threats in South Korea (“Bird Flu in Midsummer,” 2014). When a new social problem arises, news media often focus more on finding causes than solutions (at first). Considering that depicting certain aspects of causes is an important role of news media in terms of framing (S. H. Kim et al., 2011), presenting more causes than solutions during the emergence of a new social problem like AI is not surprising.
Given the dominant blame toward the government, it makes sense that government-level causes were more prevalent than individual-level causes in news coverage, consistent with Holton et al. (2012). Government-level solutions were presented significantly more during the third and seventh periods, two of the three worst periods for AI, consistent with Wallack and Dorfman’s (1996) point that news media advocacy may promote public health issues for to push for policy change.
Specific tactics of MI increased in the fifth and seventh periods, in particular; the two worst outbreaks of AI in South Korea, possibly indicating increased media advocacy efforts. This apparent increase in tactical MI is potentially contrary to research that found very limited tactical MI (Holton et al., 2012; McKeever, 2013). This finding could be related to the particular issue at hand (AI), as previous research examined news media coverage of autism spectrum disorders and vaccines, while AI is a contagious disease that requires urgent actions. On the contrary, MI found in this study focused on short-term steps (such as not visiting poultry farms or immediately reporting AI outbreaks to the government), which may suggest that news organizations and health/risk advocacy groups need to focus more on long-term solutions and prevention (McKeever, 2013). Similarly, perhaps government or health organizations need to put extra efforts into proactive crisis preparation and risk communication.
Finally, results showed news media cited few health advocacy groups and did so infrequently (the content analysis initially coded them as “other” sources), suggesting existing groups need to make greater media advocacy efforts (McKeever, 2013). More importantly, the limited number of health advocacy organizations in South Korea (I. Kim & Hwang, 2002) might explain news media’s reliance on government sources. In any case, both government personnel and news media could put extra effort into communicating with the public during both prevention routines and public health crises.
Limitations and Future Research
This study has several limitations. First, sampling of newspapers is limited. Although samples included five national South Korean newspapers with the biggest circulations, they do not represent all South Korean news media. Second, analyzing only traditional news media limits the external validity of this study. Although previous studies indicate that news content in traditional and new or social media are closely interwoven (Jang, McKeever, McKeever, & Kim, 2019; Sayre, Bode, Shah, Wilcox, & Shah, 2010), the way people consume news is always changing with the advent of new technologies. Future research can investigate whether the findings from this study are valid across various new media.
Finally, this study contributes to the body of literature about media advocacy, news framing of public health crises and risk communication, and emphasizes the important role of media in terms of communicating blame and suggesting solutions or changes to public health policies and/or practices. Although previous studies on news coverage of AI focused on sensational framing of the issue, frame-building factors, and news media as a communication tool during crises, this study sought to highlight the news media’s role as public health advocates through the case of AI in South Korea. However, more research is needed regarding global public health risks and crises, including prevention and possible policy changes, throughout Asia as well as in other parts of the world.