Transregional News Media Coverage in Multilingual Countries: The Impact of Market Size, Source, and Media Type in Switzerland

ABSTRACT In reporting about the regions of a country, news media help define relevant topics and contribute to the integration of society. In multilingual countries, covering news across language boundaries is especially important, yet challenging and costly to perform. To what extent the news media use transregional coverage, that is, look beyond their own language regions, is therefore a particularly relevant question for multilingual societies. This study shows how news media in the three language regions of Switzerland observe each other. A representative sample of news items (N = 25,035) from 47 print and online news outlets was automatically structured based on textual mentions of Swiss place names (n = 189) and linked to manually coded variables. The results show that the degree of transregional coverage differs considerably depending on the size of the language regions, the topic, the source, and the media type. When covering other language regions, news media focus on sports, and news media from the two smaller media markets rely heavily on news agency reports. Both public service media and tabloid media use much transregional coverage, with the former focusing on politics and the latter on sports. Thus, media types contribute to integration differently.


Introduction
Many news media organizations are reducing foreign correspondent networks for financial reasons (Sambrook 2010). Correspondence networks are costly to maintain, as they require specialized journalists with knowledge of the language and political and cultural characteristics of the country or region on which they are reporting (Archetti 2012). While the resulting decline of foreign news coverage or the generally "unequal representation of the world" through unequal country mentions (Segev 2019) has been widely discussed, one should be aware that similar mechanisms could also be at work within countries. Given that societal integration functions via collective identity building (Rauchfleisch and Kovic 2016), any democratic society depends on an integrated public sphere, which not only means integration across strata and political segments but also across geographic regions (Peters 2008). Addressing "place" and "space", media can transcend but also create physical and imaginary boundaries (Usher 2019, 85). In the United States, for instance, the alleged failure of larger news media from the coasts to adequately cover rural regions in the Midwest has been linked to insufficient mutual recognition and, hence, insufficient integration (Cramer 2016).
The question of geographic integration and transregional news coverage is even more pressing in the many multilingual countries of the world, especially when different linguistic regions go hand in hand with segmented media markets-as is the case in countries like Belgium, Canada, Spain, Switzerland, or the United Kingdom (Schönhagen and Trebbe 2009). Journalists are not only required to know the language, but also the political climate in the regions and the relevant actors or regional perspectives on questions of national relevance. Multilingual countries depend on institutions like journalism that contribute to integration not only at the regional and local levels but also at the national level in fostering understanding for others and a common sense of identity (Imhof 2008). Typically, these normative expectations of bridging regions are explicitly spelled out in media regulation, focusing on public broadcasting. In Canada, for instance, the Broadcasting Act of 1991 states the public broadcaster should "reflect Canada and its regions to national and regional audiences" and "contribute to shared national consciousness and identity." In this paper, we analyze how news media in multilingual Switzerland cover regions within Switzerland, targeting whether they bridge language barriers and cover events from other language regions. With four official languages, Switzerland is a clear case of an officially multilingual state segmented along language regions with accordingly segmented media markets (Tresch 2008). It is also a country with strong elements of federalism, that is, the strong autonomy of sub-national units such as cantons (states) and municipalities with direct-democratic votes regularly taking place on all levels (national, cantonal, municipal). Given the political importance of regional units and their characteristics, news media are expected to reflect this regional complexity and adequately cover regions in domestic coverage. Considering the ongoing structural problems of the Swiss news media, which see revenues from advertising and subscriptions dwindle, substantial, cost-intensive, transregional news coverage seems more difficult to conduct than before.
To map the amount of transregional news coverage and find those factors that explain the amount and type of transregional news coverage, a representative sample of Swiss news media articles from 47 news outlets from different media types and three language regions 1 is automatically structured based on the names of Swiss municipalities and linked to manually coded variables. To account for structural variation and explain differences in media content, we distinguish among media types with business models, among media from language regions, and among sources (i.e., news items produced by news media themselves or by news agencies as sources). This analysis allowed us to answer the extent to which news media cover other language regions, which topics enhance or hinder transregional references, and which structural features explain the results.

Diversity and Integration
It is a commonsense argument the media play an important role in integrating any democratic society through diverse reporting. In public sphere theories the media's integrative function is clearly highlighted (e.g., Ferree et al. 2002;Peters 2008). Through the social contract between democracy and journalism (Strömbäck 2005), news media "function as an arena for perception, identification, and definition of problems related to society as a whole" (Marcinkowski 2008). By providing "diverse audience groups with a common basis of issues and knowledge, the public sphere of the mass media establishes the basis for the self-perception of a heterogeneous society as one social community" (Kösters and Jandura 2019, 33) and for a "communicatively integrated community" (Friedland 2001) on the macro level. "[T]he clear consciousness of a communal life, in all its implications, constitutes the idea of democracy", John Dewey (2012Dewey ( [1927, 122) concluded almost century ago. Integration takes many forms, as any public sphere is segmented and stratified according to various criteria, such as age, gender, religion, and political affiliation (Imhof 2008). One crucial dimension of integration, however, is the "where" of journalism (Usher 2019): place and space. This spatial dimension has two aspects: the diversity of spatial units and the correspondence of space and the public sphere.
Public sphere theories underscore that, to fulfill their integrative function, the media should contribute to diversity. Diversity, of course, is multi-faceted (Loecherbach et al. 2020) with various sub-aspects, such as topic diversity and diverse speakers or viewpoints. Considering the importance of spatial units in public spheres and political communities (e.g., regional parliaments), however, geographic diversity has become an important element of overall diversity (McQuail 1992;Magin and Stark 2011). Diversity in media coverage should account for the diversity of levels and geographic spaces where events happen, and which geographic spaces are affected by political decisions. This is also because political interests can be stratified geographically, for instance, "with elite interests concentrated at the national level and non-elite interests at the local, regional, or even global levels" (Humprecht andEsser 2018, 1827).
From public sphere theories, clearly, any public sphere denotes a certain place; ideally, for the integration of a political community, space and the public sphere should correspond (Peters 2008). Not only a democratic nation−state relies on public sphere integration, but any democratic political community needs a public sphere (Imhof 2008). This is illustrated by various debates: whether there is a necessary "European" public sphere accompanying the political integration of nation−states in the supra-national European Union (Koopmans, Erbe, and Meyer 2010), whether there is enough actual political integration (e.g., transnational decision making) following the occurrence of a globally networked sphere (cf. Reese 2016), and whether the decline of local news (Hayes and Lawless 2018) covering political units such as local parliaments weakens regional public spheres (Kübler and Goodman 2019). Hence, an integrative public sphere is multilayered, considering polity levels. Against this normative background, media coverage is often evaluated and explained.

Geographic Diversity and Geographic Integration in Multilingual States
Even with obvious processes of globalization and the need for international cooperation (Brenner 1999), the nation−state and its regional sub-units remain vital, if not the most important, institutions for political communities' decision making-which became apparent in nation−states' reactions to the coronavirus pandemic. Strikingly, most research in the last 10 or 20 years on geographic integration and geographic diversity in public spheres has focused on transnational communication (e.g., the European public sphere) and foreign news reporting (in the tradition of news geography or international newsflow theories). Hence, scholars have provided concrete indicators to measure elements of geographic diversity, for instance, by measuring country mentions in foreign news reporting (Segev 2019;Wilke, Heimprecht, and Cohen 2012) and examining elements of integration in media coverage regarding the correspondence of the public sphere and space regarding supra-national institutions (e.g., Kleinen-von Königslöw 2012).
Compared to the focus of the literature on the supply of transnational (or global) journalism, both in legacy media (e.g., Hellmueller 2017) and on social media (Rauchfleisch, Vogler, and Eisenegger 2020), very little attention has been paid to geographic diversity and integration within nation-states. Of course, there is research on local journalism, for instance, on regional news deserts (Napoli et al. 2017), but it focuses on local news outlets, not nationwide geographic diversity within the news coverage of an outlet. In one of few notable exceptions, Kleinen-von Königslöw (2010) applied concepts from transnational public sphere research to an analysis of a national public sphere (Germany), allegedly segmented into a Western and an Eastern public sphere. Concerning integration, she indicated news media should focus on national political institutions tied to the public sphere (e.g., national government)-what could be called the vertical dimension of integration (Brüggemann and Klein-von Königslöw 2009). In the horizontal dimension, news media should monitor developments in all regions of this national public sphere, for instance, reporting on what happens in the regions and integrating opinions and arguments from regional speakers. This type of discursive exchange and, thus, connection, is more than just the expectation that debates or represented actors in each region should be similar (Marquis, Schaub, and Gerber 2011;Tresch 2012). This is because, when people are familiar with the issues and peculiarities of other regions, they can also develop an understanding of political decisions and processes in these regions, with a sense of common identity and solidarity, which allows the acceptance of majority/minority decisions within the national polity (Kleinen-von Königslöw 2010). Hence, a lack of mediated dialogue between regions would lead to the emergence of increasingly isolated media systems that rarely look beyond their borders.
It is exactly this horizontal dimension of integration and the geographic diversity of regions that are important, particularly in segmented multilingual nation-states. There, language regions often constitute separate cultural entities with their own media markets and arenas. In Belgium, for instance, newspapers are primarily read within a language region, and even public broadcasting is organized regionally, operating with two autonomous units (Jacobs and Tobback 2013). Whenever there are no transregional or national media outlets across the language regions, one necessary condition for national integration is the transregional character of the media coverage of news outlets in a language region (cf. Beier, Fiechtner, and Trebbe 2020a). Therefore, whether the news media devote attention not just to their own language region but also to a diversity of spaces beyond it is an important normative and empirical question to ask.
Overall, the few available studies provide a slightly negative picture of empirical results. Findings from multilingual India suggest news media display "significant regional inequalities in political knowledge supply" with a lack of diverse reporting (Mody 2015). Findings from the bilingual state of Canada show that the three selected national news outlets focus heavily on the region where their headquarters is and, if they cover other areas, coverage is heavily skewed toward very few places. With this coverage, the news media do not "fully exploit even the national news space of Canada.
[…] This means […] they are circumscribing Canada and making broad claims about who is part of, and matters to, the national community" (Gasher 2007, 316). In Switzerland, transregional coverage of Switzerland's public service broadcaster seems rather limited (Beier, Fiechtner, and Trebbe 2020a), even though the national broadcaster, which operates with one rather autonomous unit in each language region, is required by law to cover other language regions, that is, engage in transregional reporting to strengthen national cohesion (Signer, Puppis, and Piga 2011). Regular reports to Switzerland's regulating authority, which examines the news content of public radio or public television, confirm these findings (e.g., Grossenbacher, Hüppin, and Forsberg 2012;Beier et al. 2020b).
As such studies also demonstrate, content analyses on transregional coverage rely on rather small media samples and tend to be descriptive, i.e., do not always link outcomes (media content) to structural factors, such as the character of news (e.g., hard vs. soft) and organizational characteristics reflected in media types (e.g., quality vs. tabloid papers). These studies contrast with research on transnational communication and foreign news coverage, which often uses larger media samples and more explanatory designs. For instance, the literature using gatekeeping theories and news factors, such as relevance, status, or proximity, has found repeatedly that large, economically, and politically powerful countries gain media attention at the expense of many countries or that neighboring countries find more attention than countries that are geographically and culturally distant (e.g., Wilke, Heimprecht, and Cohen 2012). This strand of research also claims that "contradicting the flat world hypothesis, global journalism remains an exception, and provincialism is the rule" (Grasland 2020). Regarding media-internal factors, the literature has found that certain organizational characteristics enhance diversity and integration in transnational reporting. Brüggemann and Kleinen-von Königslöw (2013), for instance, have found that correspondence networks and a cosmopolitan editorial mission lead to more intense and more diverse news coverage beyond the media's origin. This accords with findings that stress that transnational coverage is conducted by quality or elite media types rather than tabloids (e.g., Downey, Mihelj, and König 2012) or that foreign news coverage is more substantial and diverse in public service media than in entirely market-based media (Cushion 2021). Similarly, the diversity and geographic integration of supra-national public spheres tends to be lower in commercial broadcasting than in public broadcasting, as commercial imperatives lead to an audience orientation that favors proximity and, hence, a focus on a narrower audience concerning geographic space (Chan and Lee 2013). This research also suggests that topic or event characteristics determine the chances of specific regions to be covered. Again, commercialized news media tend to engage in foreign or transnational news coverage, mainly in the form of soft news, such as sports and human interest (Aalberg et al. 2013).
Overall, while we know quite a lot about the patterns and factors present in transnational and foreign news coverage, we do not know enough how much and under which conditions media types conduct transregional reporting within one country and thus contribute to the diversity and geographic integration of regions. With this study, we attempt to at least partially close that gap.

Mapping and Explaining Transregional Coverage in Segmented Media Markets
Switzerland's media system is clearly segmented among the language regions. Although some larger media companies own outlets in several language regions, news outlets popular in one language region are hardly used in other language regions . News media face similar challenges when reporting on the language regions in Switzerland as they do when reporting on foreign countries or supranational organizations. Although physical distance may not be a problem in a small country such as Switzerland, the languages as well as the regions' (political) culture and peculiarities require skills from journalists covering the regions. Each region as well as the corresponding media market is also heavily shaped by a "giant next-door neighbor" (i.e., Germany, France, and Italy; Puppis 2009), which means that people in each language region also consume news from a neighboring country and that media in each region tend to give more attention to the larger neighboring country (Vogler and Hauser 2015).
In transregional reporting, the news media in Switzerland traditionally employed regional correspondents who knew the particularities of the regions and could contextualize regional political events. Due to the economic crisis of journalism in Switzerland's (very) small, segmented media markets, most of these networks have been suspended, with few outlets maintaining a network of correspondents (Lob 2019). Essentially, many now leave this task to the public service broadcaster, which operates with four organizations (one for each language region) but is legally obliged to contribute to regional diversity across language regions as part of its public service contract. Thus, reporting on the other language regions might be beneficial for society but not a promising strategy from an economic perspective, as the readership might not be interested in such news, especially if it is about smaller regions. How much the media actually observe other regions and how diverse transregional reporting is will thus depend on the structural conditions and editorial strategies of news media-but both the degree of and diversity within transregional coverage and underlying factors remain open questions.
In this section, we outline possible explanatory factors for the amount and type of transregional news coverage in Switzerland and formulate according hypotheses. We focus both on structural factors on the level of the regions (e.g., size) and on the level of news (characteristics such as the topic) and on media-internal, organizational factors such as legal mandates, commercial imperatives, and resources.

Characteristics of the Regions
Based on theories of news values and news factors, research on international news flows and international news geography repeatedly suggests that large-size effects are at play (Grasland 2020). For transregional coverage, this effect has hardly been examined.
According to data from regular reports on public radio and public TV news in Switzerland, however, similar mechanisms can be assumed (e.g., Beier et al. 2020b;Grossenbacher, Hüppin, and Forsberg 2012). In multilingual countries, one could expect smaller language regions to be considered less newsworthy. Moreover, media in smaller language regions are expected to focus especially on the largest language region, which is considered more relevant and newsworthy. In the Swiss case, this would mean the highest share of transregional coverage in the smallest language region, Italian-speaking Switzerland, and the lowest share in the largest language region, German-speaking Switzerland, with Frenchspeaking Switzerland falling in between. H1: The smaller the language region and the corresponding media market, the higher the share of transregional coverage.
As for geographic diversity in multilingual countries, there are indications that coverage is generally skewed with a few places finding attention at the expense of smaller places or peripheral regions. However, the few existing studies do not systematically compare whether media in (language) regions offer more or less diversity in their transregional reporting (Gasher 2007;Beier et al. 2020b

The Role of News Agencies in Transregional Coverage
Research on international news flows has also shown the importance of news agencies in the provision of news (Wu 2000). Even though the rise of the Internet makes it easier for journalists to cover regions without being present and some transnational media have evolved (Cazzamatta 2020), news agencies are still considered a "central hub" in the provision and diffusion of global news (Boumans et al. 2018) and news in general (Nicholls 2019). Similarly, news agencies play an important role in domestic news coverage. An automatic content analysis of Swiss newspapers revealed that using news from the one remaining news agency was very common. Indeed, it was a decisive factor in explaining that news media content is increasingly homogenized . Since news outlets do not have enough resources for an extensive network of correspondents (or do not see the value of having one), transregional news coverage, in particular, is expected to often be provided by the (only) national news agency (Keystone SDA), which has correspondents in various regions. News outlets in smaller media markets (French-speaking and Italian-speaking) especially are confronted with a lack of resources, as the possible audience size and advertising revenue is much smaller than in the larger German-speaking market (fög 2020). Faced with these structural obstacles, we expect a higher reliance on news agency materials in smaller media markets. Finally, as for geographic diversity, there is no indication from the literature on whether news agencies differ from news outlets in their provision of diversity.
H2.1: The share of transregional coverage is higher in coverage based on news agency content than in editorial coverage.
H2.2: The smaller the language region, the higher the share of news agency content in transregional coverage.
RQ2: How does geographic diversity in transregional coverage differ between news agency content and editorial reporting?

Topics in Transregional News Coverage
The news that journalists usually select can be grouped into certain topics. Topics are important structuring elements of news (Boydstun 2013), which can also be seen in the fact that news organizations have built specialized teams around topics or topic clusters, which are also reflected in sections in news content (e.g., politics, sports). Obviously, the topic selection process depends on news factors (inherent in the topics) and on the news values that journalists ascribe to these topics, which again are shaped by several factors, including macro-level factors (e.g., political culture) or meso-level factors (e.g., audience orientation of a news outlet; Kepplinger and Ehmig 2006). However, in the literature, limited indications can be found regarding how news media select topics for transregional coverage. As an exception, Gasher (2007) pointed out that sports topics are intensively covered in transregional coverage by Canadian newspapers. The literature on a related field, international news flows, paints an ambivalent picture. The fact that commercialized news media tend to engage in foreign or transnational news coverage mainly in the form of soft news, for instance, sports and human interest (Aalberg et al. 2013), ties to gatekeeping theory, according to which events outside a home region lack a crucial news factor: proximity (Wilke, Heimprecht, and Cohen 2012). Hence, extra-regional events must fulfill other news factors to pass the selection process. Conversely, a large-scale study comparing domestic and foreign news on television revealed hardly any differences concerning topic fields ( Organizational Factors: Editorial Strategies of Media Types Apart from the system level, editorial strategies at the organizational level explain the amount and type of news coverage across borders. News outlets with commercial imperatives are less willing to invest in journalism, which covers seemingly distant places and regions. This is confirmed by findings on foreign news coverage in general (e.g., Aalberg et al. 2013) and coverage of the European Union (e.g., Brüggemann and Kleinen-von Königslöw 2013). If at all, commercialized media tend to cover soft news issues in foreign news coverage (Chan and Lee 2013), and the literature on geographic diversity suggests that commercial imperatives lead to lower geographic diversity (e.g., Magin and Stark 2011). Accordingly, one could also argue that commercialized media rely more on agency copy regarding news beyond their borders. However, results are ambivalent. A recent study on quality, tabloid, and free media in the Netherlands showed that news agency material was frequent in all media types (Boumans et al. 2018).
We also expect these patterns to hold true in the case of transregional reporting and assume that differences in transregional reporting will occur not necessarily among all news outlets, but by groups or types of news outlets that share similar structural features. In the Swiss context, we can distinguish between several media types that differ in business models and the type of journalism they produce Bachmann, Eisenegger, and Ingenhoff 2021). Tabloids and commuter papers are financed more (or exclusively) by advertising, target wide audiences, and typically conduct a soft newsoriented type of journalism. In contrast, the business model of subscription newspapers is based more on a loyal readership, often tied to a regional community. Subscription newspapers target audiences with more interest in hard news and produce higherquality journalism than commercialized media types. Sunday and weekly newspapers fall between these media types in terms of business models, audience orientation, and the type of journalism they produce. The public service broadcaster (PSB), which also operates online news sites, is mainly publicly funded and is required by law to have diverse, substantial reporting, which it does according to content analyses . The public service broadcaster also has a legal mandate to consider the interests of the language regions, that is, adequately covering them and contributing to national cohesion. When looking at transregional coverage, we expect notable differences between media types. H4.1: Public service media and subscription media have a higher share of transregional coverage than tabloid and commuter media, with Sunday papers falling in between.
H4.2: In their transregional coverage, public service media and subscription media show a higher geographic diversity than tabloid and commuter media, with Sunday papers falling in between.
H4.3: In their transregional coverage, public service media and subscription media focus more on hard news topics, such as politics or the economy, than tabloid and commuter media, with Sunday papers falling in between.
RQ4: How do media types differ in how much they rely on news agency material in transregional news coverage?

Methods
This study examines the reporting of Swiss media in the three major language regions of Switzerland: the German-, French-, and Italian-speaking parts of the country. The reporting on the cantons and municipalities was determined automatically and linked to manually collected variables. This procedure makes it possible to determine the amount of reporting on Swiss cantons and municipalities, as well as the thematic focus and significance of news agency reports.
The study is based on a sample of the 2016, 2017, and 2018 coverage of 47 print and online news outlets, with the highest reach in the three language regions. For daily publications, a randomly selected artificial week was analyzed. For Sunday newspapers and magazines, which appear once a week and contain more articles, four editions in each calendar year were randomly selected. The news articles were accessed through the Swiss Media Database (SMD). The study considered 27 (57%) outlets from the German-speaking region (produced in seven cantons), 14 (32%) from the French-speaking region (produced in four cantons), and six (11%) from the Italian-speaking region (produced in one canton, Ticino, which makes up most of the entire Italian-speaking language region). The media sample reflected the sizes of the three media markets and provided justice to regions within language regions. The sample also represents the types of outlets in the Swiss press and online news markets (see supplementary file for a sample of news outlets).
To determine the relevant reporting, the most important municipalities, as well as the names of the cantons, were automatically sought in the news articles. Where available, the names of the towns and villages in German, French, and Italian were considered. The selection of municipalities was based on the number of inhabitants (Swiss Federal Statistical Office 2019). The sample included all municipalities with more than 10,000 inhabitants as well as all municipalities that made up more than 10% of the inhabitants of a canton. This selection procedure ensured that the most important municipalities were represented in the sample for both densely and sparsely populated cantons. Accordingly, 189 place names (cantons and municipalities) were analyzed for this study. Of these, 141 were located in the German-speaking region (74%), 43 in the French-speaking region (23%), and five in the Italian-speaking region (3%). In our sample, the distribution of the municipalities over the regions is thus almost identical to the distribution of the population over the regions (German: 72%; French: 24%; Italian: 4%).
Overall, 48,861 unique mentions of place names in 25,035 media articles were found (see supplementary file for a list of place names and mentions). This included 31,524 unique mentions (65%) of place names in the German-speaking part of Switzerland, 12,316 unique mentions (25%) for the French-speaking region, and 5,021 unique mentions (10%) for the Italian-speaking region. We chose place names as the unit of analysis to account for reporting that mentioned more than one region in the same article. This focus allows us to capture transregional references more comprehensively, for instance, when the place names of the French-and Italian-speaking regions are mentioned in the same article from an outlet from the German-speaking media market.
To ensure the validity of the automated analysis, the resulting lists of articles with ambiguous place names were adjusted after manual checks. This included, among others, the cities of Zug (which also means "train" in German) or Bulle (a municipality in the French-speaking region, which also means "bull" in German). The media articles for place names were then grouped according to cantonal affiliation to language regions. For multilingual cantons, place names were assigned to the language regions according to their main language. For example, for the bilingual canton of Valais/Wallis the municipality of Brig was assigned to the German-speaking region, whereas Monthey was assigned to the French-speaking region.
Based on this information, we could identify transregional coverage. For each media market, we determined media articles with mentions of place names in the other two regions (transregional references). To compare the diversity of transregional coverage in regions with unequal population sizes (i.e., number of place names), we used Shannon's diversity index in its standardized version (the so-called evenness; Mcdonald and Dimmick 2003). Shannon's Diversity Index is frequently used by journalism scholars to assess the content diversity of news (e.g., Humprecht and Esser 2018;Masini et al. 2018). The evenness indicates how evenly the elements of a population, in our case, place names, are distributed. It ranges from a hypothetical value of zero (no evenness at all; only one place name is mentioned) to one (perfect evenness; all place names are equally often mentioned). We calculate the index for the distribution of place names in transregional coverage for each media market and for topic, source, and media type.
The automated structuring of the data was combined with manual content analysis. We coded the 25,035 articles with mentions of place names and, thus, gathered additional information on transregional coverage. First, we coded the main topic of the news article (Krippendorff's Alpha = 0.90), distinguishing between news stories on politics, economy, arts, and culture (which includes media, science, technology, religion), sports, and human interest (fög 2020). The category of human interest includes news on celebrities, sex, and crime, and other entertainment-centered stories. In Curran et al.'s (2009) definition, these topics are subsumed under the term soft news together with sports, which we analyze separately in this study. Second, as regards the source, we coded whether a news article was fully or partially based on news agency content (Krippendorff's Alpha = 0.85). The identification of news agency content relied on self-declaration by the journalist in the news article, which usually happens by adding the acronym of the agency within parentheses at the end or beginning of the text. The analysis was conducted by a team of trained coders with sufficient knowledge of German as well as either French or Italian. Intercoder reliability was assessed with a random sample of unique articles (n = 525), which were all coded by three members of the coding team.
To test our hypotheses, we calculated a binary logistic regression model with transregional coverage as the dependent variable and source, topic, media markets, and media types as predictors (see Table 2). We included the interaction effects between media market and source, as well as media type and source, as we expected the likelihood of transregional references in agency-based coverage to depend on the size of the media market and the media type. We also added the year as a continuous variable to see if the likelihood of transregional references changes. In addition, we ran a separate model in which we controlled for mentions of Berne, which is the seat of the national parliament and government, and which might be mentioned especially often in transregional coverage. Indeed, Berne was mentioned more often in transregional coverage compared to other place names. However, as the overall results remained very stable, we did not include the variable in the main model (see the supplementary file). To analyze the transregional coverage of topics, we calculated a binary logistic regression model with the same predictors for each of the five topics, which was possible due to our sample size. We did not include interaction effects in the models for the topics.

Characteristics of the Regions
As the first step, we analyzed the share of transregional reporting in the three media markets. The results show that the share of transregional references is lowest in the German-speaking media market (17.8%) and highest in the Italian-speaking media market (50.8%). The share in the French-speaking media market lies in between at 37.4% (see Table 1). The transregional references in the German-speaking media market are unequally divided between the French (13.1%) and Italian region (4.7%). In the French-speaking media market, clearly more references to the German (33.4%) than to the Italian (4.0%) region occur. Additionally, outlets in the Italian-speaking media markets focus more often on the German (36.0%) than the French region (14.8%). Thus, in all three cases, the larger region attracted more attention than the smaller region.
Our regression model shows that, compared to the German-speaking media market as the reference category, the likelihood of transregional references is clearly higher in the French-speaking (OR = 2.31; p < .001) and Italian-speaking media markets (OR = 3.64; p < .001). The effect sizes (odds ratios) indicate that transregional references are more likely to occur in the Italian-speaking market than in its French-speaking counterpart (see Table 2). Hypothesis 1, which assumed that transregional coverage is higher in smaller media markets, is therefore supported.
We also asked how the geographic diversity of transregional reporting differs between language regions (RQ1). According to Evenness, transregional reporting in German-speaking Switzerland is the most diverse (E = .68) and lowest in the French-speaking media market (E = .64). The Italian-speaking market fell in between (E = .65). The disparities are, though, low.

The Role of News Agencies in Transregional Coverage
Transregional references are used more frequently in agency-based coverage (39.6%) than in editorial coverage (22.3%). Our binary logistic regression model also showed that the likelihood of transregional references is higher in coverage based on news agencies (OR = 1.65; p < .001). Hypothesis 2.1 is, therefore, supported. The results further show that German-speaking media use news agency content less frequently when reporting about municipalities in other language regions (27%), while transregional coverage in the French-speaking (43%) and Italian-speaking (50%) media market relies more often on news agencies as sources. The interaction terms in our regression model show that, compared to the German-speaking media market, the likelihood of transregional references appearing in agency-based coverage is higher in the French-speaking (OR = 1.75; p < .001) and Italian-speaking media market (OR = 1.62; p < .001). Hypothesis 2.2, which assumes that, in transregional coverage, news agency content will be more important in smaller media markets, is supported.
Interesting results occurred for the geographic diversity of transregional reporting in the dependency of the source (RQ2). When looking at the German-speaking media market, editorial coverage (E = 0.68) and news agency content (E = 0.68) did not differ respecting geographic diversity. However, in the French-speaking media market, the diversity of transregional reporting was slightly lower in editorial coverage (E = .62) than in news agency content (E = .64). The difference between editorial coverage (E = .60) and news agency content (E = .67) was most pronounced in the Italian-speaking media market. This shows the importance of news agencies for geographically diverse transregional reporting in smaller media markets.

Topics in Transregional News Coverage
We also investigated variations in transregional coverage across topics (RQ3.1). First, our results show that sports is the topic with the highest share of transregional references (36.2%), followed by news on politics (26.6%), the economy (26.3%), human-interest stories (24.0%), and arts and culture (17.5%; see Table 1). Our model shows that, compared to politics as the reference category, news on sports (OR = 1.70; p < .001) is likelier to contain transregional references. In reporting on arts and culture (OR = .61; p < .001) and on human interest topics (OR = .83; p = .001), such references are less likely to appear. No difference was observed in the comparison with the news on the economy (OR = .94; p = .116).

Organizational Factors
As the last step, we analyzed transregional reporting in media types overall and in dependency on topic and source. As shown in Table 1, the share of transregional references was highest on the websites of the public service broadcaster (PSB) (42.7%), followed by commuter media (36.6%), tabloids (30.3%), the Sunday press (26.3%), and subscription media (23.6%). Consequently, our regression model shows that, compared to subscription media as the reference category, transregional references are likelier to appear in all other media types (see Table 2). This only partially supports Hypothesis 4.1. Surprisingly, tabloids (OR = 1.51, p < .001) and commuter media (OR = 1.28, p < .001) weigh transregional coverage rather highly. Geographic diversity within transregional coverage does not diverge substantially among media types. In average, PSB (E = 0.61), tabloids (E = .62), commuter media (E = .63), the Sunday press (E = .64), and subscription media (E = .65) all had similar diversity scores. The results tend to support some of the assumptions of Hypothesis 4.2, especially with subscription media being the most diverse type. Against our expectations, however, PSB shows the least diverse reporting.
To analyze distinctions between topics, we calculated an individual regression model for each topic with the same predictors (see Table 3). Concerning topics, stark differences can be observed compared to subscription media as the reference category. Commuter media feature more transregional references only in their coverage of sports (OR = 1.35; p < .001) and human interest (OR = 1.60; p < .001). PSB media are likelier to use such references in all topics except sports reporting (OR = .88; p = .714). When looking at effect sizes, we see that transregional references in PSB are especially likely to appear in coverage of culture and arts (OR = 2.82; OR < .001). In the Sunday press, the likelihood of transregional references is higher than in subscription media, except for reporting on the economy (OR 13514.8 Note: Effect sizes as odds ratio (OR). a German-speaking media market is the reference category; b subscription media is the reference category. * p < 0.05 ** p < 0.01 *** p < 0.001. = 1.09; p = .347). In tabloids, transregional references are likelier to appear in coverage of all topics except for politics (OR = 1.09; p = .344). The results, although very insightful, do not support the overall assumptions of Hypothesis 4.3.
To answer RQ4 we look at the interaction effects in our initial model (see Table 1). Compared to subscription media, the likelihood of transregional references appearing in agency-based coverage is lower in tabloids (OR = .78; p < .01) and the Sunday press (OR = .49; p < .001) but does not vary from coverage in PSB (OR = 1.01; p = .893) or commuter media (OR = .92; p = .216).

Discussion and Conclusion
This study analyzed how the media report on regions within a multilingual country, focusing on the case of Switzerland. We propose the term transregional coverage for media coverage, which focuses on other regions in a defined geographic area-in this case, a multilingual nation−state. However, we argue that the concept of transregional coverage should not be limited to multilingual countries segmented into language regions. Instead, to what extent parts of society observe each other and develop an understanding of the peculiarities and needs of other regions or groups is relevant beyond linguistically constructed borders. News media, in general, play an important role in the integration of societies. Our results suggest the concept of transregional news coverage could be an underestimated part of the puzzle of media performances for society and democracy and could stimulate ongoing conversation about the role of "place" in journalism (Usher 2019).
From an empirical stance, the case of transregional coverage within a country revealed similar results as international and transnational reporting. Thus, our study complements the strand of research on international and transnational news flows and finds four important explanatory factors: size (of the language region and the media market), media type, source type, and topic type. The size of a region affects how much it is covered by other regions. We found lower media attention to the smaller language regions, whose fewer and smaller cities produce fewer newsworthy events. The focus on a few huge countries is well known in the literature on news values and international news flow (Wilke, Heimprecht, and Cohen 2012). The size of a region's media market also affects how much its media cover other regions. According to our results, transregional coverage was the least important in the German-speaking media market, the largest of the three markets. Again, this is like findings from transnational communication, which shows news coverage in larger countries is more nationally than transnationally oriented (Brüggemann and Kleinen-von Königslöw 2013).
From a normative perspective, it is an open question whether the very high focus on its own language region in the German-speaking media market is justified. The German language region is substantially larger in size and population and in the number of cantons than its two counterparts; hence, news media have more territory, institutions, and events to cover. However, this is only one possible normative perspective stemming from a representative liberal model of democracy (cf. Ferree et al. 2002). From other normative perspectives, one could also argue that whenever stable structural differences (such as language regions) exist, it is even more important to highlight or even privilege the perspective of the "other" to foster common understanding. One could even see as a problem the fact that, compared to population size, the smaller media markets still devote too much attention to their own language regions. Media in Italian-speaking Switzerland, for instance, which makes up less than 10% of the Swiss population, includes almost as many references from its own language region as from the other two language regions. Overall, the results show that transregional coverage does take place in multilingual Switzerland, but language regions show a substantial degree of segmentation.
As became clear, the type of topics matters for geographic diversity within transregional coverage. For instance, human-interest stories come from a more diverse set of spaces in other language regions than news items on politics. This suggests that political news coverage is more oriented toward a few large and (seemingly) relevant places and institutions, whereas newsworthy human-interest events can occur anywhere. Topics also matter for the share of transregional coverage. In sports reporting, transregional references are most likely to occur. Sports, thus, proves to be a connecting, integrative element between the language regions and makes a central contribution to the mutual perception of the three parts of the country. This specific finding is noteworthy as it also differs from observations in international news flow research, which did not find large distinctions between domestic and foreign coverage of topics (De Swert et al. 2013). The fact that sports find that much interest in transregional coverage calls for explanations. A possible factor could be the country's small size, which means even relatively small teams and athletes from the three language regions regularly compete, especially in the intensely covered national football and ice hockey leagues.
Our investigation also highlights the importance of news agencies. Especially in transregional coverage, we see agencies make "the (online) news world go round," to quote a recent paper on the use of news agency material in the Netherlands (Boumans et al. 2018). In Switzerland, the vast majority of agency material comes from only one news agency (Keystone-SDA), which is problematic in terms of source diversity (Boumans et al. 2018;Welbers et al. 2018). However, our study also reveals that, in smaller media markets, transregional reporting based on news agency content is more geographically diverse than editorial coverage. In times of increasingly scarce resources for journalism, news agencies might ensure geographic diversity in transregional coverage rather than reducing it.
Beyond the specific variances in the use of news agencies as sources, our study confirms the importance of media types. Generally, not all media types engage in transregional reporting to the same extent. As is so often the case, "media types matter" and should be more systematically considered "antecedents" of news in general (Strömbäck and Van Aelst 2010). Somewhat surprisingly, tabloid media focus relatively often on other language regions, albeit rather seldom on politics. Public service media conduct the most transregional coverage on politics and arts and culture, and thus contribute to the integration function for a political community, but geographic diversity within their transregional coverage is still limited. The study also confirms the importance of media with a national, cross-regional focus-especially compared with more regionally oriented subscription media, which tend to underweight transregional coverage. Although the latter assume other important functions, namely integration at the local level, they cannot replace media that report across language regions. Examining a truly wide sample thus yields results on outlets' and media types' performances for an overall national media system, which can help inform the public and media regulators in media policy debates, which still take place mainly at the national level (Cushion 2021). One implication for media policy, for instance, is to ask which types of media are best suited to provide transregional news coverage and, therefore, could (and should) be financially supported: a variety of individual private media outlets, the public service broadcaster, or one news agency that provides news from regions to all other media?
Our concept of transregional reporting addresses the importance of news media for democracies in multilingual societies. Further research could not only apply the concept to other countries with two or more official languages, such as Belgium or Ireland, or two or more unofficial languages, such as the United States (Darr et al. 2020), but also to geographic areas with other segregating lines, such as the periphery versus the center or countries with minority regions not segregated by language.
Our approach and methods have some limitations. The automated analysis allowed us to investigate a large sample of media articles. In combination with the coding at the article level, we could determine how often the places are mentioned in transregional coverage and in which topical context. However, we did not retrieve any information on how exactly the places were covered. In some cases, we could not distinguish between mentions of cities and cantons with the same name. For example, regarding the special case of Bern, we cannot distinguish whether the mentions refer to Bern as the political center of Switzerland or to the canton or city of Bern. Furthermore, the automated text analysis did not allow us to analyze TV and radio newscasts, but print and online news only. For the case of Switzerland, this is a limitation, as the programs of the public service broadcaster are regarded as especially important for regional integration (Schönhagen and Trebbe 2009) and as the websites of the public service broadcasters with textual news, which we analyzed, somewhat differ from the linear radio and TV news casts. Another aspect we could not investigate was how transregional news was consumed, for instance, compared to global news (Widholm 2019). There is little integrating potential of news when it is produced, but it is not widely received. Questions about how transregional news is received would therefore complement our study, which focused on the content of news. Despite these limitations, our study contributes to journalism scholarship by offering the concept of transregional news coverage to account for a better understanding of media performances within multilingual or segregated societies, where "place" plays a crucial role in various aspects. Note 1. Switzerland officially has four language regions (German, French, Italian, and Romansh). The Romansh language region is by far the smallest (speakers of German: 63%; French: 23%; Italian: 8%; Romansh: less than 1%; not counting speakers of foreign languages). Therefore, we focus on the three larger language regions.

Disclosure Statement
No potential conflict of interest was reported by the author(s).