Media coverage of the novel Coronavirus (Covid-19) in Kenya and Tanzania: Content analysis of newspaper articles in East Africa

Abstract Abstract: This study uses the framing theory to analyse dimensions being conveyed in the coverage of COVID-19 in Kenya and Tanzania between February 2020 and April 2020. A quantitative analysis of the Daily Nation and Citizen newspapers showed different patterns of framing of the virus. Specifically, this analysis focuses on multiple frames used by the two newspapers with respect to the following topical categories: context, basic information, preventive information, treatment information, medical research, Social context, Economic context, Political context, personal stories and other. Although the Daily Nation published more stories than the Citizen Newspaper, only the frame personal stories were significantly higher in the Daily Nation compared to Citizen Newspapers.


Introduction
The novel coronavirus pandemic took the world by surprise. The pandemic left devastation of varying degrees on the global stage. Regardless of who (whether an individual or institution) frames the knowledge, most of what we know about the coronavirus came from mediated sources. For Nulman (2022), the mediated channel was "perhaps our only window" to understand John-Bell Okoye is a PhD candidate in communication at Daystar University, Kenya. Media and health discourses are his areas of research interest.

PUBLIC INTEREST STATEMENT
The novel coronavirus pandemic took the world by surprise and left devastation of varying degrees on the global stage. Regardless of who frames the knowledge, most of what we know about coronavirus came from mediated sources. Presently, most countries in Africa are experiencing the second wave of the pandemic, while a few are stepping into the third wave. Despite these phases of COVID-19 experiences on the continent, this study focused on the early stage of COVID-19 in Kenya and Tanzania through the lenses of The Daily Nation Newspaper in Kenya and The Citizen Newspaper in Tanzania with an aim to identify how these two newspapers framed the pandemic and identify differences in depictions of the pandemic. Furthermore, this study is important as the public's behaviour, reaction or response to a situation is usually a reflection of government approach to the issue and media representations of the event.
the coronavirus crisis and our place in it (p. 3). Although news consumption may have tremendously changed in recent times, tilting towards accessing online news disseminated on the web, the qualities associated with how audiences appreciate the news remains durable and stable (Meijer & Kormelink, 2021). The present study investigates media framing of the coronavirus in Kenya and Tanzania, using the Daily Nation (Kenya) and The Citizen (Tanzania) newspapers.
Currently, most countries in Africa are facing the second wave of the pandemic, while a few are stepping into the third wave. Despite these phases of COVID-19 experienced on the continent, this study focused on the early stage of COVID-19 in Kenya and Tanzania. Given the complex challenge posed in the first wave of COVID-19 for journalists to report the health crisis with accurate data from official sources (Gearing, 2021), it is important to understand how newspapers portrayed this critical moment of the virus in Kenya and Tanzania. In the opening chapter of Communicating Science in Times of Crisis: COVID-19 Pandemic, O'Hair and O'Hair (2021) asserted that "COVID-19 is not a disaster simply because it is a pandemic, its widespread destruction is also due to the world's lack of preparation and ill-advised responses to it" (p. 9). Africa is known for its wobbling health systems, thus, researching the initial wave of the novel coronavirus in Kenya and Tanzania is pivotal to assess the role of the media and government responses. Southwell et al. (2020) acknowledged that a good understanding of people's behaviour during a pandemic is to investigate the initial wave of the health crisis. As we should know, the public's behaviour, reaction or response is usually a reflection of the governments approach to the issue and media representations of the event.
The choice of the Daily Nation and The Citizen newspapers are based on their audience base, agenda-setting influence and circulation reach in Kenya and Tanzania, respectively (Elliott, 2015;Media Council of Tanzania & Reporters without Borders, 2018). The study proceeded as follows: engagement with literature on framing and public communication during pandemics; methods and findings; and the last section discussed the implication of the findings in relation to framing of health crises. The phenomena of interest to study were investigated with these research questions and hypotheses.
H1a: The Daily Nation newspapers covered more stories related to the coronavirus than The Citizen newspaper.
H1b: Dominant frames in the two newspapers focus more on coronavirus basic information and prevention than on treatment, medical research, social, economic, and political impacts of the virus.
H1c: The pandemic was framed as a global crisis rather than a local crisis.
H1d: As cases begun to rise in the two countries, the two newspapers focused more on personalizing the coronavirus.
Research question 1: How did the Daily Nation and The Citizen newspapers frame the COVID-19 pandemic between February and April 2020?
Research question 2: What differences exist about depictions of the pandemic in the two newspapers in terms of coverage of salient frames?
Overall, the study seeks to advance knowledge on the early coverage of the coronavirus disease in Africa from a framing perspective. Health communication studies on early representations of the virus in the media focussed mainly on the global north and China (see e.g., Wang & Mao, 2021;Yiu et al., 2021).

Global framing of pandemics
Media coverage of pandemics and health crisis have generated concentrated media coverage as the public's desire for information increases. Research on media framing of pandemics highlights similarities and differences in frames as well as common frames associated with pandemics. Tian and Stewart (2005) in their study of SARS, compared how BBC and CNN framed the crisis. The scholars found similarities and differences in the framing. Both news outlets were concerned with the spread of SARS, both discussed SARS impact on public health and the travel industry and both framed the crisis from a global perspective. However, CNN had a cluster about SARS economic impact, which the BBC did not. Furthermore, "control" was a more noticeable theme in the CNN text than in the BBC text.
Using the 2014-2015 Ebola outbreak as a case study, Pieri (2018) analyses how the risk of contagion in the US, Europe, and the UK has been constructed in UK media and policy discourse. The Ebola outbreak was framed as a localized African crisis, a regional crisis, a global security threat and an outbreak. These frames produced an obsessive fear of contagion spreading to the West. Lee and Basnyat (2013) trace the development of framing devices from a government public health agency's press releases to news stories about the 2009 H1N1 influenza pandemic. Through content analysis the researchers found significant changes in media frames, including the expansion and diversification in dominant frames and emotion appeals, stronger thematic framing, more sources of information, conversion of loss frames into gain frames, and amplification of positive tone, favouring the public health agency's position. Shih et al. (2008) examined how print media frame public health epidemics, such as Mad Cow Disease, West Nile Virus, and Avian Flu. They found that the "action" and "consequence" frames were the two frames journalists employed consistently to construct stories about epidemics in the New York Times. The authors also found different attention cycle patterns for each disease. Coverage of public health epidemics was highly event based, with increased news coverage corresponding to important events such as newly identified cases and governmental actions.

Public communication during pandemics
Some of the truths, albeit bitter, that emerged from the Covid-19 pandemic are "related to political authority and community spirit" (Caron, 2021, p. 2) in dealing with the disease. In other words, how are governments engaging with citizens about the virus, and how are the people responding to the conversations in order to bring the virus under control? Effective communication is often a central aspect of health crisis response during outbreaks. Efforts to communicate with the people can take different shapes and more so emanate from different sources. But in all this, the government is the strategic authoritative voice to reach out to the people for reassurance and direction. Noar and Austin (2020) echoed similar perspective: "A well-crafted national message has the potential to build unity around the goal of defeating the virus through behaviour change" (p. 1735). Public service announcements (PSAs) have been consistently used during outbreaks. Created mainly by non-governmental organisations (especially those with focus on health-related issues) and government parastatals (like ministry of health), the goal of deploying PSAs is to create public awareness that encourages behavioural change. This mass sensitisation approach has been used in previous pandemics like the Zika virus, SARs, Ebola, H1N1 and presently in the coronavirus health crisis. leveraging the power of PSAs entails appropriating the media in form of video clips, print contents, radio messages (Manganello et al., 2020), and in some instances use of town hall communication, to reach wide and diverse members of the society.
In recent health communication scholarship, the shift in the use of PSAs entails adding social media dimension through user-generated contents. Most scholars and health communication practitioners are indeed imagining that the future of communicating health information during outbreaks is likely to be user-generated social media disseminated contents (Manganello et al., 2020). For instance, Jiun-Yi Tsai and colleagues averred that exposure to user-generated messages made social networking sites important sources for coping with previous pandemics like the MERS and the current coronavirus disease (Tsai et al., 2020).
Given the diffusion of smartphones and increasing access and spread of the internet on the global stage, social media is an important communication platform to communicate health crises (Yu et al., 2020), has democratised public health communication through its affordances and functionality (Schillinger et al., 2020) and has accelerated the speed of disseminating behavioural change information about outbreaks (Dong & Zheng, 2020). Despite the seemingly obvious gain, this approach to communication during pandemics is fraught with challenges.
These challenges have manifested to some extent in previously experienced outbreaks, but in more problematic ways in the present Covid-19 health crisis. The most worrisome of these likely challenges is the ubiquity of disinformation, untruths, and misinformation in our contemporary society. Besides the thriving system of misinformation, there are other problems associated with use of user-generated content and social media for health communication, which include the use of conspiracy theory and trolling (Manganello et al., 2020).
As some researchers have pointed out, depending on how the messages are presented and the audiences ability to see through the contents, fake news or misinformation could feel so real and sway the audience into believing what they read as authoritative health information (Schwarz & Jalbert, 2021). In fact, "in many ways, one might think that social media were designed to make questionable messages seem true" (p. 17). In what Viswanath and co-researchers described as a "pandemic of the social media age," coronavirus communication over the web has bred opportunity for misinformation and disinformation to thrive unchecked in an atmosphere of limited gatekeeping. The implication, therefore, is the dissemination of unreliable and unscientific messages that furthers the spread of the virus and endangers the lives of the people in unimaginable ways (Viswanath et al., 2020(Viswanath et al., , p. 1744. However, it is important to note that the role of social media and user-generated content for public health promotion is still contentious and unresolved debate (Schillinger et al., 2020).
Scholars have argued that an effective preparation model or strategy to combat infectious diseases of pandemic proportion requires paying adequate attention to how people respond to such health crises. While panic tends to be the default reaction from the public, Southwell et al. (2020) remind us of the need to better understand how people approach emerging infectious diseases as objects of public opinion, especially at the initial phases of the manifestations of the diseases. Privileging such mindset fits within one of the core strategies of communication during health crisis. Kaur-Gill and Dutta (2021), for example, are of the view that the central purpose of deploying strategic health communication is to improve health outcomes. For sure, emerging infectious disease (EID) is invested with high level of risk and uncertainty (Holmes et al., 2009).
This state of uncertainty that characterises EIDs, if not properly handled with effective communication strategies, enables misinformation and infodemics to thrive and, in extreme cases, potentially eclipse accurate health information about infectious diseases. The metamorphosis of the COVID-19 pandemic is a site of contestation of meanings, an experience riddled with conspiracy theories, resulting in the coining of the term "infodemic" by the World Health Organisation, WHO, to describe the situation and the danger it poses to communicating accurate health information. In its simplest understanding, infodemic represents an overabundance of misleading information with the consequence of confusing and harming people who encounter it (Culloty & Suiter, 2021;Wright, 2021).
As part of its 2020 Misinfodemic Report: COVID-19 in Emerging Economies, Africa Check (2020) reported evidence of COVID-19 related misinformation circulated via social media in East Africa, with emphasis on Kenya. Wright (2021) noted that the implication of infodemic is usually unhealthy. Relatedly, Levy and colleagues stressed that infodemics undermine the capacity and credibility of experts to communicate science accurately and whittle the publics' ability to make informed decisions (Levy et al., 2021) in the face of health challenges. For this reason, health communication researchers have suggested the use of an emerging infectious disease framework to communicate about health crisis and engage with the people. The application of an EID communication approach pivots around its importance to build trust with the people while leveraging the efficiency of the trio of crisis communication, risk communication, and health communication. It also acknowledges the centrality of mass media in the communication process (Holmes, 2008;Holmes et al., 2009). This helps to further explain the focus of the present research which seeks to understanding the role of newspapers in Tanzania and Kenya respectively in increasing public awareness about coronavirus pandemic in both countries, as well as providing reliable COVID-19 news items to the people. Manganello et al. (2020) and colleagues noted the urgency to communicate information in a timely manner during a pandemic. Beyond the need for timely information is the need for credible health communication, one in which citizens can believe, rely on and act upon. This is where traditional media becomes a useful public sphere to engage with the audience and discuss the way forward in terms of progress made, challenges encountered and plans in place on how to defeat the virus with concerned members of the public. Hence, in order to shape public opinion about health risks and enforce behavioural changes, the mass media, no doubt, plays a very significant role (Hart et al., 2020).

Coronavirus outbreak in the news media
For Mahon (2021), the Covid-19 pandemic generated an increase in news reports, created a lot of untruths, and thus bred a challenge for citizens in search of accurate information in the midst of an unknown threat. Research in health communication has shown a correlation between media types and audience trust of the news. Much of these studies found that legacy media, especially television and newspapers, are positively associated with higher level of trust among the people. The same experience is re-enacted with the coronavirus health crisis, whereas the crisis evolves not all media types will be trusted equally (Tsai et al., 2020).
Pivotal as the media are in reporting health crises, Sharma (2021) noted that the focus of the media is influenced by several extraneous factors. However, whatever factors determine news selection and positioning, the media role in framing pandemics has the potential of spawning an (in)direct effect on audience behaviour change and perception. Dong and Zheng (2020) study revealed that Covid-19 coverage in China sensitised the target audience while "exaggerated news reports" engendered panic and stress among the people. In the U.S., early coverage of the outbreak of coronavirus was riddled with polarisation and politicisation. The researchers indicated that the political leanings of mainstream media contributed to the polarised attitudes about the virus among the public in the U.S. (Hart et al., 2020).
The point is, depending on the framing pattern, media coverage of Covid-19 has implications on audiences mental wellbeing (Zheng et al., 2020), financial market stability (Haroon & Rizvi, 2020), tourism sector (Yu et al., 2020), stress disorder on the people (Dong & Zheng, 2020), mitigating the spread of the disease (Liu et al., 2021) and above all have the likelihood of eroding trusts on the institution of science (Mandelbaum, 2020) and medical professionals (Motta et al., 2020).
While the coronavirus pandemic has generated a lot of news interest (Mahon, 2021) and scientific publication (Mandelbaum, 2020), much of the studies around mediated communication of the disease has been focussed and conducted in the western world and China. The role of the media and communicative discourse of the Covid-19 outbreak in Africa is missing, necessitating the need for such research to provide knowledge on how the continent is dealing with the pandemic. It is within this research gap that this study is situated to understand how newspapers have initiated and engaged with discussion around the disease in two African countries (Tanzania and Kenya) that are the two strongest power hubs in Eastern Africa with tourism as one of their foreign exchange earners.
The manner in which the media in Kenya and Tanzania approached reporting the virus will impact on audience attitudes, perceptions and opinions in these countries, especially given the centrality of the media in the socio-political fabrics of Kenya and Tanzania. Researchers have linked the uncertainties, emotional fatigue (Sylvia & Bundenz, 2020), a sense of panic (Haroon & Rizvi, 2020) and the racist underpinning of Covid-19 pandemic (Lwin et al., 2020;Ruiz et al., 2020) to mediated representation of the disease.

Materials and methods
Framing is a normal part of the communication process where communicators act to construct a particular point of view that encourages the facts of a given situation to be viewed in a particular manner, with some facts made more or less noticeable than others (Kuypers, 2010;Kiptinness & Kiwanuka-Tondo, 2019). According to its proprietor Goffman (1974), frames enable individuals to discover, recognize, identify, and label occurrences or information. The entire study of mass communication is based on the principle that the media have significant effects (McQuail, 1994). These effects determine how the public views situations, notably the coronavirus pandemic. When highlighting some aspect of reality over other aspects, frames act to define problems, diagnose causes, make moral judgments, and suggest remedies.
A quantitative content analysis was conducted to study the framing of the coronavirus by the Daily Nation newspaper and The Citizen newspaper during the period of February 2020 to April, 2020. We conducted an online search of articles covering coronavirus in the Daily Nation newspaper and The Citizen newspaper during the specified time period which yielded 3811 results in the Daily Nation newspapers and 503 results in The Citizen newspaper online portals. A constructed week was used for data collection in this study. The overall goal of constructed week sampling is to create maximum sampling efficiency while controlling for cyclical biases such as weekly news patterns (Luke et al., 2011), furthermore, a constructed week sampling is more efficient than simple random sampling or consecutive day sampling. Hester and Dougall (2007) recommend at least two constructed weeks are needed to accurately represent online news content gathered. All articles in the first and the last weeks of February, March and April, 2020 within the search criteria of "coronavirus" were sampled.
Videos and advertisements were not included as part of the samples. Therefore, the Daily Nation yielded 189 articles while The Citizen yielded 127 articles. Total articles were 316.
Two coders were trained for the data collection process. The coders undertook a pilot coding exercise of 10% of the 316 articles. The inter-rater coding activity yielded a percentage agreement of 70% using Cohen's Kappa. The unit of analysis was the entire story. Coders coded all stories in the first and last weeks of February, March and April, 2020 within the search criteria of "coronavirus" in both newspapers. The coding sheet included aspects such as the date the article was published, length of the article, the by-line as well as the presence of the frames in the article. Semetko and Valkenburg (2000) define two approaches to framing analysis, the inductive and deductive approach. The inductive approach starts with broad presuppositions of frames with an aim of identifying all possible frames. The deductive approach has stronger presuppositions of frames and examines the occurrence of these frames in the news. After analysing the literature on media framing of pandemics (Lee & Basnyat, 2013;Pieri, 2018) the author chose the deductive approach. The predetermined frames which were determined from a review of literature on media reporting on pandemics (Pieri, 2019;Lee & Basnyat, 2013;Shih et al., 2008;Collinson et al., 2015) were: (1) Context-Articles portraying the coronavirus as a National, Regional (East Africa/African) or Global (outside Africa) crisis.
(2) Basic information-Includes information or updates on Covid-19 in terms of numbers of fatalities or confirmed cases, or WHO actions.
(3) Preventive information-Articles with information of good behaviors that prevent Covid-19 spread, such as hand washing, wearing a mask and temperature taking.
(4) Treatment information-Includes treatment options including what to do, and where to go for treatment of the coronavirus.
(5) Medical research-Articles with new medical findings, new technology related to the virus e.g., vaccine development and efficiency or virus mutations.
(6) Social context-Includes articles with juxtaposition of Covid-19 into a societal context, e.g., impact on social activities such as social gatherings.
(7) Economic context-Juxtaposition of Covid-19 into an economic context, e.g., tourism or support from local businesses, job losses.
(8) Political context-juxtaposition of Covid-19 into a political context, e.g., local politics or diplomatic ties) including donations from countries.
(9) Personal stories-Includes articles with experiences of patients/families.

(10) Other-Anything else of interest
With regard to ethical considerations, all articles were freely accessed via both newspapers websites and no identifying information of the article authors or names of individuals or companies mentioned in the articles were recorded. Chi square was used to analyse data.

Findings
As Table 1 notes, the Daily Nation newspaper covered more stories on the coronavirus than The Citizen newspaper between February to March 2020 (N = 189; N = 127). The dominant frame in the Daily Nation newspaper was the social frame with 55.6% articles (Table 1), while the dominant frame in The Citizen Newspaper was the Basic frame with 59.2% articles ( Table 1). The Daily Nation Newspaper framed the coronavirus as a National crisis with 51.3% news stories portraying a national context frame while The Citizen Newspaper framed the pandemic as a Global crisis, with 38.6% stories (Table 2). From Table 1, news stories with personal frames did not feature highly with Daily Nation scoring 17% while The Citizen Newspaper scoring only 9%. The difference between the two was found to be significant (0.008).
As seen on Table 1, tin most of the frames there is no significant difference between the depictions of the pandemic in the two newspapers in terms of the salient frames used in the coverage except for the framing of personal stories. Figure 1 shows that the Daily Nation had significantly more articles with the personal stories frame than the Citizen newspaper (0.008).

Discussion
From the findings of the study, the Daily Nation newspaper in Kenya mainly employed the social frame and depicted the COVID-19 pandemic as a national crisis. On the other hand, The Citizen newspaper in Tanzania employed predominantly basic frames and portrayed the challenge as a global problem. An initial approach to understand the variance of framing patterns between both newspapers is appreciate the remarks of health communication and journalism scholars who note that the coronavirus disease revealed differences in responses of different governments in addressing the problem (Gearing, 2021;Nulman, 2022), some of which are ill-advised according to O'Hair and O'Hair (2021).
While the Kenyan government was rallying citizens, sensitising them on the existential threat posed by coronavirus disease, the Tanzanian government under the late president was sceptical of the virus and thus treated information about it with levity. These different strategies have implications on how the media reported the health crisis.
As John and fellow researchers pointed out, framing in health communication should be conceived as a macro and micro level constructs. In other words, framing studies and interpretation of the findings should be cognizant of the source and receiver of the health information (John et al., 2014). What these scholars are trying to explain is that frame definers impact how the receivers understand and potentially respond to mediated messages. Researchers have argued that political institutions, actors, and elites are significant definers of frames in the news (Gabore, 2020; Okoye, 2020). In early newspaper coverage of the COVID-19 crisis in China, Wang and Mao (2021) study revealed that local newspapers portrayed the virus as a controllable disease and thus blunt the edges of the threat associated with it. But as soon as the national government stepped in, the news depicted the health situation as an uncontrollable event at the local level. The framing slant positioned the pandemic as a national issue that required solidarity to fight the common enemy. When compared with the findings of the present study, it clearly showed the role of government to evolve strategic communication to confront health crisis. While the Tanzanian government refrained from daily update of the situation, the Kenyan government was persistence with communicating updates that highlighted the infection rates and casualty figures. Since most media depend heavily on the government for such news, it is quite understandable why the framing of the coronavirus crisis by the Daily Nation newspaper was more nationalistic and The Citizen presented the problem as a global one with the potential risk of confusing Tanzanians into believing they are immune to the disease.
In terms of frame salience, in most cases there was no significant difference between the coverage of both newspapers except for the depiction of personal stories. Arguably, this finding might be attributed to the ownership ties shared between the Daily Nation and The Citizen newspapers. We must admit that this is the major limitation of the study. However, the lesson for understanding framing of health issues is that ownership structure can impact outcome, regardless of whether the newspapers bear different names and operate in different sovereign countries as is the case with the newspapers studied. Furthermore, the Citizen newspaper may have covered less personal stories during the study period due to Covid-related stigma concerning this "Global" pandemic, rather than the National frame portrayed in the Daily Nation which localized the pandemic.
The values that the findings of the study hold for framing of health challenges are two-fold. First, it is important to investigate the early framing pattern of such events. This is vital to notice changes in the role of the media in terms of the framing directions and how the target audience are responding to the messages. It helps to measure the preparedness of the people about emerging infectious diseases as a product of a mediated narrative that impacts public opinion (Southwell et al., 2020). In the age of disinformation, misinformation and infodemics amply facilitated by social media, it is necessary to engage in researching the initial phase of health crisis like the COVID-19 pandemic. For example, even though the study focussed on the first three months of the disease in Kenya and Tanzania, the findings indicated fewer personalised frames across both newspapers. One interpretation to advance is that coronavirus disease was attached with some elements of stigmatisation and stereotyping early enough (Mutanu, 2020) as was the case with other emerging infectious diseases like HIV/AIDS and Ebola virus. This stigmatisation, therefore, may prevent people from sharing their stories with journalists.
A second point worth noting regarding what the findings mean for framing health crises is that it shows the extent to which citizens trust the government and scientific experts as frame definers in the news on the one hand, and the degree of their trust of the media on the other hand. Cairney and Wellstead (2021) expounds on this: "In a [health] crisis, almost-instant choices about who to trust or distrust could make a difference between life and death" (p. 1). More specifically, health practitioners are equipped with empirical knowledge of the preparedness of policymakers in responding the early signs of infectious diseases. As the findings revealed, the Tanzanian government paid little attention to the pandemic, and this is well reflected in the framing of the news about the virus as a global crisis. This, no doubt, recalls the message in third-person effect hypothesis, in which case Tanzanians may view themselves as less vulnerable to coronavirus diseases, whereas their counterparts in Kenya would, based on framing of the news and the daily updates from the government, understand their vulnerability in the face of the health concern.