Abstract
This study examines the online niches of English-language newspapers in Bangladesh by analyzing stories and user engagement on the Facebook pages of five highest circulated newspapers. Based on the theory of the niche, this study analyzed data on 17,687 Facebook posts downloaded programmatically. Results indicate that users read these newspapers for in-depth news stories and features on diplomacy, national security, lifestyle, science, technology, and international events. The study identified a gap between what these newspapers offer and what their users want.
When traditional news media across the world were struggling for survival, the newspaper industry—particularly English-language newspapers—in Bangladesh, a developing country in South Asia with a population of over 163 million, seemed to be growing (Doctor, 2018; Rolt, 2012). As of July 2017, there were 32 English-language dailies in Bangladesh, and most of these newspapers were established after 1990 (Anam, 2002; “Unknown Newspapers Sell,” 2017). The number has been growing steadily as any major English-language newspaper shutting down was rare during this period. However, little systematic knowledge exists about the content and users of Bangladeshi newspapers as scholars rarely paid attention to them although the country with the world’s eighth largest population has a long history of news media (U.S. Census Bureau, 2018). This study is an attempt to fill part of this gap in knowledge. It takes the advantage of publicly available data on Facebook pages where English-language newspapers of Bangladesh post stories and engage with users (Carey, 2014; Josephson & Miller, 2015; Trilling, Tolochko, & Burscher, 2017). Facebook pages offer a way to conduct various analyses of contents and user reactions to them and, thus, provide some insights into what these newspapers produce and what their users want. The theory of the niche is the basis here to translate this interaction into a functionality of the newspapers (Dimmick, 2003).
This study examines Facebook pages of the five highest circulated English-language dailies in Bangladesh. It analyzes news stories, headlines, and Facebook status messages to understand newspaper content aimed at online readers. It also examines user engagement on the Facebook pages to understand needs and wants of online readers. Taking advantage of the Facebook Graph application programming interface (API), the authors downloaded all Facebook posts and engagement data—subject to the API limit—from the official pages of those dailies (n = 17,681) in early January 2018.
A deeper understanding of the English-language newspaper audience would be valuable to both local and international media organizations seeking to find global audiences as the Internet transformed audiences “from almost exclusively local to communities of interest that transcend geographic and political boundaries” (John Pavlik in Silverstone, 1999, p. 58; Thurman, 2007). Thurman (2007) found that websites of some U.K. news media (e.g., The Guardian) attract more unique visitors from the United States than the websites of local newspapers such as USA Today. The legacy news media industry is growing steadily in several South Asian countries such as Bangladesh and India at a time when this industry is in sharp decline in most parts of the world (Tharoor, 2017; Vaidyanathan, 2011). This growth—often attributed to increasing rate of literacy, connectivity, and economic growth—offer opportunities to international media organizations to build new partnerships and audiences. Huffington Post and Buzzfeed have set their eyes beyond the Western countries for new users (Lichterman, 2015; Wang, 2015). Huffington Post started a partnership with a news portal in Afghanistan called Sahar Speaks to create exclusive stories for local audience (Wang, 2015). One major problem that such platforms face is lack of knowledge about non-Western audiences (Lichterman, 2015). Scott Lamb, BuzzFeed’s vice president for international growth, told Nieman Lab that Buzzfeed took longer than expected to launch its service in Japan because of lack of internal knowledge about the country (Lichterman, 2015). Research on English-language newspaper audiences in countries such as Bangladesh will, therefore, be a valuable addition to the literature that would benefit both the industry and academia.
Also, as the struggle of traditional news media keeps worsening across the world, it is more important than ever for newspapers to have a clear understanding of user needs and allocate resources accordingly (Dimmick, 2003; Mitchell, 2014). This is also a key proposition of media economist John Dimmick’s theory of the niche, which explains conditions for media organizations’ survival (Dimmick, 2003). The theory informs this study as it seeks to understand the functionality of English-language newspapers in Bangladesh on the Internet.
The Press in Bangladesh
The press in Bangladesh, a part of the Indian subcontinent ruled by the British Empire for almost 200 years until 1947, has a long history dating back to the regimes of Muslim rulers who had an efficient newsgathering system centuries ago (Khurshid, 1971). Modern print journalism came to the area during the British colonial rule with the launch of the English-language weekly Hickey’s Bengal Gazette published from Kolkata (formerly Calcutta) in 1780 (Mahmud, 2013; Mukherjee, 1990). All newspapers in the Indian subcontinent were in English until 1818 when the first newspaper in Bengali, the native language of most people in Bangladesh, appeared (Haider, 2010; Mahmud, 2013; Rahman & Ahmed, 2006). Newspapers first appeared in the territory of the current Bangladesh in the mid-18th century (Khurshid, 1971). Before independence in 1971, many newspapers published in Bangladesh, including some English-language dailies such as Morning News and Pakistan Observer. There was a national television station and a radio station with some substations at that time.
After liberation, Bangladeshi media started to grow significantly but founding president Sheikh Mujibur Rahman declared one-party rule in the country in 1975 and shut all newspapers except four state-owned dailies (Bhuiyan, 2011). Until 1990, Bangladeshi media were under tight control. Media restarted growing after a mass uprising toppled military dictator, Hussain Muhammad Ershad. Bangladesh’s journey toward development began with the outset of democracy in 1991 when “media were at least given freedom” (Anam, 2002, p. 267). Since then, media became stronger and promoted development objectives and democracy. Anam (2002) wrote, “The boldness, the depth, and the frequency with which some of the leading newspapers write about corruption in government and in the private sector, abuse of political power, nepotism, and irregularities in general is something previously unknown in Bangladesh” (p. 268).
The overall media system of Bangladesh has a pattern similar to Western media (Bhuiyan, 2011). Although most people in Bangladesh live in rural areas, its print media focus on urban areas as rural people can hardly afford to read newspapers (Anam, 2002). Mass media of Bangladesh are mostly privately owned. With the restoration of parliamentary democracy in 1991, Bangladesh’s press experienced a wave of tremendous growth. The foreign currency revenue from the apparel products exported mainly to the countries in North America and Europe and remittance from its workers working in Middle Eastern countries contributed significantly to its economic growth. People started migrating to urban areas to find employment. With that socioeconomic growth, news media experienced a boom both in quantitative and qualitative terms. Thus, newspapers readership of both Bengali and English-language newspapers began rising. Almost all the English-language newspapers popular with the readers today started publishing after the 1990s. For example, just after restoration of parliamentary democracy, The Daily Star, the largest circulation English-language daily, began its journey in January 1991. Two other newspapers named The Financial Express, a business news daily, and The Independent began publishing later in 1993 and 1995, respectively. New Age came in the market in 2002. During the last two decades Bangladesh got a few more English-language newspapers, with the latest newspaper, The Asian Age, beginning in 2016.
English-language dailies now publishing in Bangladesh include The Daily Star, New Age, The Daily Observer, The Independent, The News Today, Dhaka Tribune, Daily Sun, The Bangladesh Today, The Financial Express, The New Nation, and The Asian Age. Besides newspapers, some online news agencies, including state-owned Bangladesh Sangbad Sangstha (BSS), United News of Bangladesh (UNB), bdnews24.com, and banglanews publish news in English. Bangladesh Television (BTV), Radio Bangladesh, and most private television channels broadcast daily small-scale English-language bulletins. Many international news agencies such as Reuters, Associated Press (AP), Agence France-Presse (AFP), Xinhua, British Broadcasting Corporation (BBC), and Voice of America (VOA) have bureau offices in Dhaka, Bangladesh’s capital. Many regional and international news outlets, including Press Trust of India (PTI), Islamic Republic News Agency (IRNA), Islamic International News Agency (IINA), Antara, Cable News Network (CNN), The Guardian, Al Jazeera, and The New York Times also have representations in Bangladesh in the form of stringers.
The English-language newspapers play critical roles in serving the public, covering the issues of political, sociocultural, financial, geopolitical, national security, environmental, and entertainment. Bangladesh’s largest English-language daily, The Daily Star, has a circulation of about 50,000 and its official Facebook page has about 2.7 million followers. The Financial Express also covers issues relating to business and economy. This newspaper has significant reach to the business communities in Bangladesh with about 1.6 million followers on its official Facebook page.
However, only a small portion of Bangladesh’s 163 million people read English-language newspapers. There is no systematic study about the size of English-language newspaper readership in the country, but some assessments suggest that English-educated elites, expatriates, some university students, many business people, and decision makers are the main consumers of English-language newspapers (Banu & Sussex, 2001; Genilo, Asiuzzaman, & Osmani, 2016). The readership is limited mostly to cities. Majumder and Hasan (2013) found that around 37% students of a local university read English-language newspapers while a study in another university put the figure at 7% (Akanda & Haque, 2013). It remains ambiguous how the English-language newspaper industry in Bangladesh is surviving and thriving.
The Theory of the Niche
The theory of the niche, Dimmick’s ecological theory for studying media competition and coexistence, explains how media populations or industries survive, thrive, or fail in competitive markets (Dimmick, 2003). The theory suggests that each population in a market must have a niche that supplies resources. Monge, Heiss, and Margolin (2008) defined niche as “a finite resource space” that helps a limited number of organizations sustain themselves (p. 453). Dimmick noted that the niche of a media population is “its relationship to its environment,” referring to news media’s relationship with their users and advertisers. Dimmick (2003) identified six resource dimensions—gratifications, gratification opportunities, content, user time, user spending, and advertising money—on which media’s relationships to their environment rests. All these resource dimensions, except for content, are related to users and advertisers. Only the content dimension is related to the media themselves, which determines the extent to which users and advertisers will be engaged with media. A mass medium is likely to have a niche in a market if its contents are perceived by users to be fulfilling their needs.
Resource dimensions identified in the theory of the niche serve as measures of the relationship between a mass medium and its environment. For instance, gratifications measure the extent to which users perceive contents of a mass medium to be fulfilling their needs. Gratification opportunities measure the extent to which users can access media contents at different times from different locations. For example, smart phones provide higher gratification opportunities than televisions because a user can use her phone at almost any time in any place. At the same time, one has to be in a specific location to watch television. Dimmick (2003) suggested that gratifications and gratification opportunities “are the basis on which media are selected by their patrons” (p. 29). Other dimensions, such as the time users spend with media and the money users pay for contents, are more direct measures of the niche.
The resource dimensions can also be categorized as (a) organizational offerings (e.g., content types), (b) uses (e.g., gratifications, gratification opportunities, and user time), and (c) financial resources (e.g., user spending and advertising money). For instance, news organizations can differentiate their niches along various topics (e.g., politics, entertainment), geographical areas (locality), and formats (e.g., hard news, feature story) of content (van der Wurff, 2011). In terms of uses, media organizations may differentiate their niches along different gratifications, such as information seeking, surveillance, and entertainment (Vincent & Basil, 1997). Media niches may also vary in accordance with their resource use patterns (Dimmick, 2003). For instance, almost all U.S. newspapers require both paid subscriptions and paid advertising (mostly the latter) to survive while U.S. movie theaters sell tickets and refreshments (and make most of their profits on the latter). In sum, the niche of a media organization may be found by looking at the content types it offers, uses of its content, and its resource use patterns. Dimmick noted that the theory of the niche is “abstract and general” (p. 24), leaving the option open to researchers to explore new resource dimensions.
The resource dimensions are also the basis on which niche breadth, niche overlap, and competitive superiority are calculated. Niche breadth refers to the range of services that an organization offers to users. For media organizations, service refers to various attributes of media content that fulfills user needs and wants. Niche breadth can be calculated on various resource dimensions as Dimmick described niche breadth as a “measure of the area of a niche along a particular resource dimension or axis” (p. 37). Niche overlap is defined as “ecological similarity between two [media] populations” (Dimmick, 2003, p. 37). It measures similarities or differences between two media in terms of their resource use patterns; media are assumed to be in competition if they have similar resource use patterns. Niche superiority explains which media are more likely to survive in case of strong competition as resources shrink. Theoretically, niche breadth, niche overlap, and niche superiority can be calculated on each resource dimension.
Many scholars used the theory of the niche to identify niches of various media populations and examine competition in different media markets in the United States and other countries (e.g., Dimmick, Chen, & Li, 2004; Dimmick, Kline, & Stafford, 2000; Dimmick, Patterson, & Albarran, 1992; Li, 2001; van der Wurff, 2011). Among others, studies examined niches of video entertainment industries (Albarran & Dimmick, 1993), mobile media (Dimmick, Feaster, & Hoplamazian, 2010), interpersonal media, such as cell phones, land line telephones, messaging, and email (Dimmick, Feaster, & Ramirez, 2011), cable and broadcast industries (Dimmick et al., 1992), email and telephone (Dimmick et al., 2000), news on television, electronic news and newspapers in Taiwan (Li, 2001), and Internet and traditional news media (Dimmick et al., 2004).
Although some key premises of Dimmick’s theory of the niche rest on niche overlap and competitive superiority measures, this study focuses on niche breadth because it studies only one media population in a market rarely studied by scholars. The concept of niche breadth is particularly suitable for the current study as it seeks to understand the niche of English-language newspapers in Bangladesh. Most studies testing or based on this theory were conducted in well-researched media markets. Available empirical evidence enabled researchers to identify similar populations and apply niche measures suitable for studying competition. As little or no empirical evidence exists about Bangladesh’s English-language newspaper population, having a deeper understanding of its niche breadth would be a significant addition to the literature. This would amount to drawing a part of the map of Bangladesh’s media industries.
Although content is central to the resource dimensions identified in the theory of the niche, Dimmick (2003) stressed the importance of analyzing user perceptions about the extent to which different media (e.g., television, newspaper, or Internet) or media outlets (e.g., a newspaper website) fulfill their needs and wants. One approach that the theory of the niche employed to identify niches and analyze competitiveness among media organizations or industries is the uses and gratifications approach. Gratifications “represent the needs served by media and are the basis on which media are selected by their patrons” (p. 29). However, finding a representative sample for a survey of newspaper users in a developing country is very difficult. Moreover, finding a sample of English-language newspaper readers in Bangladesh, who are scattered and account for a small portion of the overall population, is impossible. In addition, scholars identified several shortcomings of the uses and gratifications approach. McQuail (1994) argued that this approach could not predict or explain media use because much of media consumption is accidental. (This is a major reason why uses and gratifications cannot be called a social scientific theory.) The uses and gratifications approach is individualistic, pays little attention to content, and assumes media are functional (CCMS-Infobase, 2003). The approach also faced criticism for its reliance on surveys or other self-reports, which have limitations such as biased reports because of weak memory and social desirability (Kreuter, Presser, & Tourangeau, 2008; Wright, 2005). Therefore, finding an alternative to the traditional uses and gratifications approach to analyze media uses would broaden the analytical scope of the theory of the niche.
The Internet and new technologies that allow researchers to collect large data sets about media uses offer an alternative to self-reported data. For instance, user engagement data (e.g., reaction, comment, share, tweet, or reply) on a social media page can indicate which types of media content are popular among users. These data are programmatically recorded in real time as users interact with websites. Such data sets allow researchers to study at least some aspects of media markets such as Bangladesh, which are otherwise impossible to study. For markets where studying user perception of media gratifications through surveys may be a reliable approach, online user engagement data can work as complement and help make better niche predictions. Therefore, this study proposes to add a seventh resource dimension—user engagement—into the theory of the niche.
User Engagement: A New Dimension in Study of Media Niche
Scholars defined engagement in various ways. Engagement is often referred to as interactivity, connectivity, or interaction of users with webpages, media products, or content creators (Bartlett, 2013; Gade & Lowrey, 2011; Ksiazek, Peer, & Lessard, 2014; Napoli, 2011). As Anderson, Swenson, and Gilkerson (2016) put it, “Multiple overlapping definitions exist for related terms like engagement, interactivity, and responsiveness, causing potential confusion for researchers and professionals” (p. 4095). User engagement in the context of the current study refers to user interaction with content or another component on a webpage, which can be recorded and has the potential to reveal information about users’ media preferences. Such interaction may come in various forms, such as providing reactions to an online news story, commenting on it, and sharing it. Researchers and media organizations have used such user engagement data, known as web analytics, to predict what users want (Attfield, Kazai, Lalmas, & Piwowarski, 2011).
The Internet and new communication technologies empowered media users in an unprecedented way. Users now play an important role in the process of content creation and distribution (Mitchell & Holcomb, 2016; Tewksbury & Rittenberg, 2012). Many scholars argue that news organizations and journalists must be engaged in discussion with users to learn about their needs. Gade (2011) noted that users can “interact with journalists about stories, pass on news tips, and actually report and create their own news” (p. 26). Lowrey and Gade (2011) suggested that news organizations follow online social networks, particularly users who actively search news for consumption as well as share news with others. Social media activities of these people provide vital information to news organizations. In other words, reaction to a social media post, commenting on it, or sharing a post are examples of user engagement.
To understand the niche of a media organization or an industry, it is important to know what the interests and goals of the users are. This was the justification for choosing gratifications as a major resource dimension (Dimmick, 2003). User engagement online addresses this issue. Network scholars agree that online engagement—particularly engagement on social media—is motivated by mutual interests and goals (Castells, 2011; Hunter, 2008; Rheingold, 2000). A user won’t actively engage (i.e., put a comment on a post) in a social network until the user finds it serving her or his interests. User engagement on social media has strong potential in measuring user preference because such engagement in networks are collective, dynamic, open, and scalable. Anyone can join a network and leave if it does not serve one’s interests. Only the people whose interests are served by the network will stay (Castells, 2011).
This, however, does not mean news organizations should publish more trivial stories, or merely what people want, and abandon serious stories (e.g., public affairs news). This means it is not enough for journalists “to get the truth and print it.” They also “have to process that truth in ways that make readers want to go to the effort of receiving it” (Meyer, 2004, p. 229). Research shows that people are interested in serious news stories, but the way they consume them is different (Bialik & Matsa, 2017). Today’s media users are diverse, fragmented, heterogeneous, and specialized as they have abundant choices available on the web (Dimmick, Powers, Mwangi, & Stoycheff, 2011; Lowrey & Gade, 2011). In other words, the web created numerous specialized niches for news from generalized niches. The theory of the niche provides a theoretical framework to explore new online niches while social networking websites provide the data.
Engagement on Facebook as a Niche Measure
Facebook has proven itself as an effective platform for news organizations and journalists to engage with users and learn about their needs. Several studies provide evidence in support of this argument. Garcia-de-Torres et al. (2011), who examined how journalists from seven South American and European countries use social media, concluded that Facebook “gives a richer, more personalized collaboration; and it engages readers” (p. 18). These journalists from Argentina, Colombia, Mexico, Peru, Portugal, Spain, and Venezuela believe that Facebook helps them find the right audience (Garcia-de-Torres et al., 2011). Lauren McCullough, a former social media manager at the AP, suggested that Facebook became a place for conversation between journalists and users (Craig & Yousuf, 2013). McCullough was quoted as saying, “It’s important for journalists to always be going to the places where the users are, and not just expecting users to come and find us” (Craig & Yousuf, 2013, p. 101). One example of journalist–user engagement was the ProPublica Patient Safety Community, a Facebook page created by ProPublica to have a conversation with victims of medical harm (Victor & Allen, 2012). As of February 12, 2018, the group (https://www.facebook.com/groups/patientharm/about/) had more than 5,300 members.
Studies also examined techniques and strategies media outlets had applied to engage readers online. Josephson and Miller (2015) found that adding a summary headline, while sharing a news story on Twitter and a question headline on Facebook, engages more readers. Drawing on an eye-tracking approach and a short survey with participants, the study found the techniques worked well in these two social networking sites. Trilling et al. (2017) found that news stories about Dutch domestic issues got the highest interaction on Facebook. On Facebook, domestic issues got 1.8 times as many shares as stories about nondomestic issues; on Twitter the same issue got 1.29 times as many. The study also found that cultural distance also matters: Stories about non-Western countries receive only 0.83 (Twitter) and 0.69 (Facebook) times as many shares as the baseline of Western countries. A topic that is closer to home is shared more often because it involves another Western, and thus culturally similar, country. Al-Rawi (2017) examined the types of Facebook messages on news media pages that are more engaging to Arabs. Results show that Arab Facebook users are more engaged with posts on broad topics than on breaking news. Interesting quotes also generate interaction and conversation among Arab users. Carey (2014) examined Facebook pages of community newspapers to find subjects that generate more user engagement. Findings showed that weather-related posts generate twice the amount of engagement in the forms of comments and likes than average post. Posts related to crimes, disasters, politics, public health, and welfare also generate more engagement than an average post.
The literature above demonstrates that Facebook does not work only as a platform where many users get news, but also provides opportunities for news organization to learn more about what users want from them and how they want it. Facebook, like many other social networking websites, offers public APIs through which researchers can access publicly available data. Previous studies indicate that data extracted from Facebook pages can provide deep and reliable understanding of the niches of news organizations.
Research Questions
Three major concepts stood out in the literature that would help understand the niche of the Bangladesh’s English-language newspaper industry. The concepts are news content, engagement techniques employed by newspapers, and user engagement. This study, informed by these three concepts, also explains the extent to which English-language newspapers were catering to the preferences of their audiences. Based on the literature, this study asks the following four research questions.
The first two research questions address the behavior of newspapers in Bangladesh and the last two questions address user preferences.
Method
This study examined Facebook posts and user engagement with those posts of five highest circulated English-language dailies in Bangladesh. The pages are (a) The Daily Star (https://www.facebook.com/dailystarnews), (b) Dhaka Tribune (https://www.facebook.com/DhakaTribune), (c) New Age (https://www.facebook.com/newageonline/), (d) The Financial Express (https://www.facebook.com/febdonline), and (e) Daily Sun (https://www.facebook.com/dailysun.newspaper). By using the Facebook Graph API and a system based on Unix system commands, the authors scraped all posts—subject to Facebook graph API limit (https://developers.facebook.com/docs/graph-api/reference/v2.12/page/feed)—and user engagement data (e.g., number of reactions, number of comments, and number of shares) from these five Facebook pages at the beginning of 2018. The data set contains a total of 17,681 posts (The Daily Star: n = 4,601; Dhaka Tribune: n = 3,449; New Age: n = 4,032; The Financial Express: n = 2,816; Daily Sun: n = 2,783). The data set came with information about medium used (e.g., link, photo, video, text) in each status message, date and time of each post, headline of each story, and URL to the news and Facebook URL for each post. To answer the questions relating to engagement techniques, descriptive statistical analyses (e.g., frequency count, average) were conducted on the entire data set. To answer questions relating to the types of news content these Bangladeshi dailies post on Facebook, the authors took a sample of data set (n = 793) from year 2017 and manually coded them to identify types of news content and engagement techniques used. Content analysis is a suitable method for this study as Stempel (1985) suggested that manifest content gives information about antecedent choices as well as the conditions and processes followed to generate content.
Operationalization of Key Concepts
Three major concepts relating to engagement were used in the research questions—news content, engaging techniques, and user engagement. News content was measured by three variables: (a) types of news stories (e.g., hard news, in-depth analysis, soft news/feature, and opinion), (b) topics (e.g., government and politics, health, sports, entertainment), and (c) locality (i.e., national, international). A similar list of content categories was used by van der Wurff (2011) in his study on news media niche. Engaging technique was measured by three variables: (a) headline type (e.g., traditional news headlines, and persuasive headlines), (b) media type in status (e.g., link, photo, text, and video), and (c) type of status message (e.g., summary of news, same as headline, promotional, and none). A traditional news headline “provides an overview of the main idea in an inverted pyramid-structured news story” (Scacco & Muddiman, 2016, p. 1). Persuasive headlines are enticing headlines designed to capture user attention by employing persuasive textual attributes (e.g., question marks, exclamation point) or emphasizing unknown information (Breaux, 2015; Scacco & Muddiman, 2016). User engagement was measured by the number of reactions, comments, and shares on each post. The literature shows that these measures were useful indicators of user preferences (e.g., Carey, 2014; Trilling et al., 2017).
Coding of Status Messages
The data collection system used to scrape Facebook data had precoded 15 variables and relevant categories programmatically (major ones were identified at the beginning of “Method” section). However, the authors manually coded status messages, headlines, and the first paragraph of news for five variables to identify the engaging techniques used by the newspapers on their Facebook pages. The authors separated the 2017 data (n = 4,325) for manual analysis to identify the most recent trends as practices relating to social media are relatively new and continuously changing. Also, social networking sites continuously update their algorithms that affect practices and user engagement on these networks (Bromwich & Haag, 2018). For instance, one Oklahoma news portal, Oklahoma Watch, posted a story about how users can find its story online as Facebook decided to downgrade its news feed (Oklahoma Watch, 2018). From the posts initially selected, posts tagged as link by Facebook were separated and 25% of them were chosen (n = 793) randomly for manual coding. Link refers to status message that contains a link to a webpage.
Coding Protocol and Data Analysis
Much data presented in this study came precoded. However, for the data analyzed manually, a coding protocol was developed. Five variables were identified for this analysis—news type, headline type, status message type, topics, and locality of content. The authors then randomly selected a subset of data and read headlines, status messages, and full stories to identify relevant categories for each variable. One author wrote down description of each category in a codebook highlighting the “focus, a parameter or boundary, for discussing a particular event” (Altheide & Schneider, 2013, p. 53). This same author, who did not code any data, was also responsible for calculating intercoder agreement level. Two training sessions were conducted, each spanning about an hour and a half, for two coders who have more than 20 years of combined newsroom experience in Bangladesh. After the first session, each coder coded a small amount of data, but they could not reach an acceptable agreement level for two of five variables. A second training session was then conducted to identify and discuss the areas of disagreement. Both coders then coded 7% of data separately and agreement was over 85% on each variable. The Krippendorff’s alpha test (Hayes & Krippendorff, 2007) was conducted to test intercoder reliability on each variable. Results are as follows: news type (α = .78), headline type (α = 1.00), status message type (α = .90), topics (α = .85), and locality (α = 1.00). The results show high intercoder reliability for all variables except for news type that is still within acceptable range.
Statistical tests using several methods were conducted to find out statistically significant differences. Nonparametric tests (chi-square) with post hoc analyses were conducted on the data relating to content types. The analyses were meant to see if newspapers put more emphasis on some topics, formats, headline types, or locality more than others. Post hoc analyses would reveal similarities and differences among newspapers. ANOVA with post hoc analyses and t tests were conducted on the engagement data to find out if variances in user engagement with different content types are statistically significant. Some descriptive analyses were conducted on the whole data set to show different types of status messages posted by the newspapers under investigation.
Results
Content Types on Facebook Pages of English-Language Newspapers in Bangladesh
Note. Frequencies have been reported in percentages. Star = The Daily Star; Sun = Daily Sun; Tribune = Dhaka Tribune; Express = The Financial Express; N Age = New Age; All = Five newspapers combined. For Topics variable, categories that account for less than 3% of all content were removed from the table. Each subscript letter denotes a subset of newspapers that does not differ significantly at .05 level.
A post hoc analysis revealed two subsets of newspapers in terms of news types across most categories. There are no statistically significant differences within newspapers in a subset on most categories. One subset (A), consisting of Daily Sun, Dhaka Tribune, and New Age, post significantly more hard news than the other subset (B) consisting of The Daily Star and The Financial Express. Subset B posts significantly more in-depth stories than Subset A. The same pattern exists in locality variable. Although a chi-square test (χ2 = 29.9, p < .001) shows that the newspapers differed significantly in terms of locality of the news posted on Facebook, post hoc analyses suggest that significant difference exists only between the two subsets. Subset A posted significantly more national news than Subset B. When it comes to news topics, stories on politics appeared to be posted more often on Facebook than others. A chi-square test indicates significant differences among newspapers in terms of topics (χ2 = 272.7, p < .001). However, post hoc analyses suggest that major differences exist in a small number of categories. New Age appeared to be posting significantly more political news than any other newspaper while The Financial Express was posting more business stories than others. Differences among newspapers on most other categories are not statistically significant.
The third variable used to measure engaging technique was the type of status message associated with each link to a story. Quotes or part of a linked story were used as status message in nearly half of all Facebook posts (49.6%) while 43.8% contained no status message. A chi-square test on this variable indicates statistically significant differences among newspapers (χ2 = 821.1, p < .001). Post hoc analyses revealed two subsets of newspapers in terms of how they utilize status message option. The Daily Star, Dhaka Tribune, and Financial Express heavily use quotes or parts of linked stories as status messages while the Daily Sun and New Age rarely use any status message at all.
Distribution of User Engagement With Different Content Types
Note. F and t values without any asterisk symbol are not statistically significant.
p < .05. **p < .01.
The number of comments and shares that the news stories under investigation receive on Facebook is very low. A number of statistical tests found no significant difference among the numbers of comments on news categories relating to any variable. A one-way ANOVA test on sharing of stories found statistically significant difference in only one variable—topics (F = 2.32). The posts that were shared most often were about diplomacy (M = 13.2).
Discussion
Grounded in the theory of the niche, the current study performs several tasks. It provides an understanding about the niche of English-language dailies in Bangladesh—a market rarely studied by academics. By analyzing both news stories and user reactions to them, the study identified some ways these newspapers could strengthen their relationships with users and broaden their niche. The findings also revealed ways likely to attract user attention to the public affairs stories. Significant differences within the industry have also been identified. The study extends the theory of the niche (Dimmick, 2003) by (a) proposing and defining a new resource dimension—user engagement, and (b) by applying the theory and the new concept in a relatively unexplored media market.
Findings indicate that the English-language dailies in Bangladesh have a relatively narrow niche breadth often defined by the utility that users perceive to receive from a medium (Dimmick, 2003; Dimmick et al., 2000). Though these newspapers cover stories in various formats on a wide range of topics, users appeared to engage more with stories on a relatively smaller number of topics (e.g., diplomacy, national security, lifestyle, science and technology, and international stories) than others. Users also prefer some formats (e.g., in-depth analysis and feature) over other formats. The findings about user engagement start to address concerns about how online users want news. The key here is that users engage with serious topics such as diplomacy and national security if they are produced in preferred formats such as in-depth analysis or features. This lends support to the findings of Bialik and Matsa (2017), a descriptive study that found online media users love public affairs stories if they were produced in the way users want.
The results indicate that most English-language newspapers in Bangladesh lack a clear understanding of their niches. Except for The Financial Express, all newspapers posted more stories about politics than any other topic while stories on politics received the least amount of reactions from users. In contrast, the newspapers posted a relatively small amount of stories on national security (less than 3%) and diplomacy (8.4%), while these two topics received relatively higher amount of reactions than most other topics. In terms of story format, more than 70% of the stories analyzed were hard news and opinions, while these two were the least popular news formats among the Facebook users. However, the newspapers rightly focus on entertainment stories—one of the most engaging topics. From the perspective of the theory of the niche and other relevant literature, these findings about what the English-language newspapers produce, what their users want, and how they want can be explained in three ways.
First, these newspapers ignore a major proposition of the theory of the niche, which suggests that an industry or organization must differentiate its products from others’ if products are perceived by users to be similar (Dimmick et al., 2004; Dimmick et al., 1992; Li, 2001). A high level of similarity would cause strong competition between industries and lead to differentiation of products or displacement of one by others from the ecosystem. As English is not the native language of most readers of these newspapers, it can be easily assumed that local public affairs topics such as politics and crime are covered widely by newspapers publishing in the native language—Bengali. Thus, English-language newspapers may be in competition with Bengali-language newspapers as far as these topics are concerned. Also, this study’s findings indicate that English-language newspapers may not have niche superiority over Bengali-language newspapers with regard to these topics and, therefore, are in need of differentiating their products in terms of the topics they emphasize on.
Second, the literature suggests that users of English-language newspapers in Bangladesh are relatively more educated, richer, and more aware of complicated public and international affairs than the average people in the country (Banu & Sussex, 2001; Genilo et al., 2016). These users may find English-language newspapers more reliable than Bengali-language newspapers for coverage of national security and diplomacy, which often involves international players and interpretation of events, issues, and statements that come in English. This study’s findings reiterate that these users put a high level of trust into English-language newspapers for these stories, which indicates opportunities for these newspapers to broaden their niche. However, data show that these topics get very little coverage, meaning that these newspapers are ignoring opportunities in their ecosystem.
Third, it has been widely reported that online news users consume news in different ways than traditional mass media users (Castells, 2011; Mitchell & Holcomb, 2016; Tewksbury & Rittenberg, 2012). They are fragmented and specialized. In other words, online media users are topic-driven; they are heavily engaged with topics they like and ignore other topics. This can explain the popularity of story formats such as in-depth analysis and features. These formats are known for providing greater details and insights about any event, issue, or trend. This coincides with points Philip Meyer made in several books (Meyer, 2002, 2004). As mentioned earlier, journalists “have to process that truth in ways that make readers want to go to the effort of receiving it” (Meyer, 2004, p. 229). Focusing on in-depth analysis and feature stories may also be the key to engaging users with stories on other public affairs topics that received relatively less user engagement. In other words, English-language newspapers may broaden their niche to topics such as politics and crimes by focusing on the formats users want.
The findings suggest that the theory of the niche provides a useful framework to identify the functionality of media organizations in a market. A combined analysis of what the industry produces and what users prefer could help organizations design better content strategy, serve the audience better, and enhance financial sustainability. The concept niche breadth appeared to be an effective concept with which to examine functionality of media organizations.
The concept user engagement has also appeared to be a useful resource dimension to explore functionality or niche breadth of media organizations. Knowledge of programming languages—particularly the uses of libraries relating to data scraping or data mining (e.g., Python’s urllib and newspaper3k)—would give researchers access to valuable media use data to build stronger theories. Such data sets contain more nuanced and organization- and product-specific data. The current study also provides a specific definition of user engagement that would help reduce confusion among researchers regarding many overlapping definitions of engagement.
This study makes three major contributions. First, it explored the niche of a media market in a South Asian country, which is rarely studied. Little systematic research has ever been done about this country’s media. Second, the study extends the theory of the niche by applying it to a non-Western media market. Third, this study proposed addition of a new resource dimension for niche studies—user engagement. By using data scraped programmatically from Facebook, the study also reveals new possibilities in the study of news media. Overall, the study contributes to various areas of scholarly literature, including newspapers, online news media, media economics, strategic management, media in South Asia, and media in Bangladesh. It also contributes to Bangladesh’s media industry.
Despite its contributions, the study has several limitations that future studies should avoid. As this is the first such study about the media markets in Bangladesh, analysis has largely been descriptive. Second, analysis of newspapers in the native language would have provided a better understanding of how the niche of English-language newspapers in Bangladesh is different from the niches of other major news industries such as Bengali-language newspapers, broadcast news, and magazines. Future studies could address these issues.
