Media Systems in the Digital Age: An Empirical Comparison of 30 Countries PDF

Journal of Communication ISSN 0021-9916 ORIGINAL ARTICLE Media Systems in the Digital Age: An Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/jo...

Journal of Communication ISSN 0021-9916 ORIGINAL ARTICLE Media Systems in the Digital Age: An Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 Empirical Comparison of 30 Countries Edda Humprecht 1, Laia Castro Herrero1,2, Sina Blassnig1, Michael Brüggemann 3, and Sven Engesser4 1 Department of Communication and Media Research, University of Zurich, Zurich, Switzerland 2 Universitat Internacional de Catalunya, Spain 3 Department of Journalism and Mass Communication, University of Hamburg, Hamburg, Germany 4 Institute of Media and Communication, Technical University of Dresden, Dresden, Germany Media systems have changed significantly as a result of the development of informa- tion technologies. However, typologies of media systems that incorporate aspects of digitalization are rare. This study fills this gap by identifying, operationalizing, and measuring indicators of media systems in the digital age. We build on previous work, extend it with new indicators that reflect changing conditions (such as online news use), and include media freedom indicators. We include 30 countries in our study and use cluster analysis to identify three clusters of media systems. Two of these clus- ters correspond to the media system models described by Hallin and Mancini, namely the democratic-corporatist and the polarized-pluralist model. However, the liberal model as described by Hallin and Mancini has vanished; instead, we find empirical evidence of a new cluster that we call “hybrid”: it is positioned in between the poles of the media-supportive democratic-corporatist and the polarized-pluralist clusters. Keywords: Media Systems, Typology, Cross-National Comparison, Information And Communication Technologies, Political Communication, Journalism https://doi.org/10.1093/joc/jqab054 Research on media systems has played an important role in comparative com- munication research over the past two decades following the publication of Hallin and Mancini’s (2004) groundbreaking book Comparing Media Systems. This work provided the scholarly community with a theoretical framework to compare Western media systems using a historical-institutional approach. The authors highlighted the differences and similarities between different political and media systems in North America and Western Europe. They further suggested four dimen- sions for comparing those media systems: the inclusiveness of the media market, Corresponding author: Edda Humprecht; e-mail: [email protected] Journal of Communication 00 (2022) 1–20 V C The Author(s) 2022. Published by Oxford University Press on behalf of 1 International Communication Association. All rights reserved. For permissions, please email: [email protected] Media Systems in the Digital Age E. Humprecht et al. journalistic professionalism, political parallelism, and the state’s role. These dimen- sions have been used in comparative research, for example, to explain differences between countries in news coverage or media use (Aalberg et al., 2013; Fletcher Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 et al., 2020; Magin, 2019). In addition, empirical media systems research was stimu- lated by the operationalization of dimensions and empirical validation of the typol- ogy in a study by Brüggemann et al. (2014), as researchers could now explain and measure similarities and differences between countries. More recently, however, researchers have repeatedly pointed out that media sys- tems are increasingly shaped by the rise of information and communication tech- nologies (ICTs) and that this transformation needs to be reflected in the description of media systems and corresponding typologies (Chadwick 2013; Mattoni & Ceccobelli, 2018). The development of ICTs has changed how media content is pro- duced and used. Today, most people consume media content digitally and on their mobile devices. In addition to journalistic websites, social platforms are increasingly being used to consume information. This development has changed the balance of power between traditional media companies and new digital intermediaries and has further intensified competition within the media industry. Moreover, ICTs have also changed media production, as, for example, a feedback channel for media users has emerged, and anyone can publish and disseminate digital information. Liu et al. (2020) argued that while digital technologies and applications—includ- ing mobile phones, social media platforms, and artificial intelligence—are shaping media use and production in many countries, there are few comparative studies on the topic. The authors further lament a lack of research beyond the narrow focus on Western Europe and the United States. One reason for this research gap is the lack of standardized, empirically measurable media systems indicators that reflect the role of digital technologies. Although several authors discuss possible indicators for measuring digitalized media systems and refer to potential data sources (Flensburg & Lai, 2020; Mattoni & Ceccobelli, 2018), few empirical typologies are available. A notable exception is a study by Perusko et al. (2015), who use indicators such as smartphone penetration and broadband Internet to study types of media systems across 22 countries. They find four country clusters, including (a) Eastern European countries, (b) a diverse set of Western European countries (Austria, Belgium, France, Germany, Ireland, Italy, the Netherlands, Spain, and the UK), (c) Nordic countries, and (d) Israel. The authors considered media usage and infrastructure indicators in their study. However, they did not include media system indicators that can be assigned to journalistic professionalism or political parallelism. We aim to close this gap by providing an empirically measurable typology that combines the established theoretical models developed by Hallin and Mancini (2004) and operationalized by Brüggemann et al. (2014) and aspects of digitalized media systems and ICTs. Moreover, we extend Hallin and Mancini’s (2004) frame- work by including press freedom, an important indicator that has been neglected in past studies that focused exclusively on Western Europe and the United States. We also include Southern and Eastern European countries in the analysis. Describing 2 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age and mapping media systems in this way can help us understand patterns of media content, use, and effects in the broader context of the structures of a media system. Furthermore, research on differences between media systems can uncover impor- Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 tant issues of freedom of expression, political participation, and platform gover- nance that have enormous implications for the future of digital communication. Based on the current literature, we develop Hallin and Mancini’s (2004) dimen- sions further regarding changes in the use and production of media and political and social developments. Building on the work of Brüggemann et al. (2014), we use aggregated data to operationalize existing dimensions and new indicators character- izing media systems. Furthermore, our study aims to answer two research questions. The first one addresses the relation between existing and newly developed indicators and asks how indicators linked to the digitalization of media systems can be com- bined with dimensions developed by Hallin and Mancini (2004) into an integrated framework (RQ1). The second question deals with typology building: Which media system typology results from the expanded dimensions and the enlarged country sample (RQ2)? Dimensions of Media Systems Hallin and Mancini’s (2004) original theoretical framework consists of four dimen- sions: (the inclusiveness of) the media market, journalistic professionalism, political parallelism, and the state’s role. We revisit those dimensions and build on recent re- search in the field to identify new indicators. In the following, we present a detailed discussion of these changes and the resulting revised media systems dimensions. Inclusiveness of the Media Market The dimension of the media market has been conceptualized regarding inclusive- ness, that is, as the reach of news media among different groups of society (Brüggemann et al., 2014). Existing operationalizations include (a) the overall daily news reach, (b) news reach among the working class, and (c) news reach among women (Brüggemann et al., 2014; Hallin & Mancini, 2004). Previous operationaliza- tions of this dimension focused on the printed press, examining its circulation and readership (Brüggemann et al., 2014; Büchel et al., 2016). However, several research- ers, including Hallin (2020) and Mancini (2020), argue that although leading news- paper brands continue to play an important role in many media systems, the printed press has become less important for news consumption than digital news outlets. Citizens increasingly rely on multiple media platforms and devices to access news content (Newman et al., 2021). In addition, it has been argued that the digitalization of media markets has led to a fragmentation of the audience, resulting in increasing information inequality and an increasing number of citizens who hardly use any news (Van Aelst et al., 2017; Prior, 2005). This inequality is also reflected in the fact that there are media markets Journal of Communication 00 (2022) 1–20 3 Media Systems in the Digital Age E. Humprecht et al. where leading newspaper brands still have a strong position and reach a large part of the population, whereas in other markets, such media organizations only reach a niche audience (Shehata & Strömbäck, 2011). Hence, to reflect these differences, it Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 is necessary to consider the reach of online news and measure the reach of newspa- per brands in their offline and online versions and among different groups of society. Political Parallelism The dimension of political parallelism refers to six indicators that have been opera- tionalized by Brüggemann et al. (2014) as follows: (a) the lack of separation of news and commentary, (b) partisan influence and policy advocacy, (c) political orienta- tion of journalists, (d) media-party parallelism, (e) political bias, and (f) (in)depen- dence of public broadcasting. Most of these indicators are not directly related to changes fostered by the digitalization of media systems and the rise of ICTs. Mattoni and Ceccobelli (2018) argue that although new media outlets, such as online-only news outlets and digital versions of established brands, have gained im- portance in recent years, the way political parallelism manifests itself has not been affected by these changes. According to Mattoni and Ceccobelli (2018), the digitali- zation of media systems does not lead to a complete realignment of journalistic norms and practices. However, social and political developments that lead to changes in traditional and new media and politicians’ communication can be ob- served. For example, indicators such as the political orientations of journalists con- cern offline and online media and established media brands and online-only outlets. In contrast, other indicators, such as the separation of news and commentary, may have become less important since, in the online sphere, news items are published and consumed individually, not in specially marked sections as in newspapers (Boukes et al., 2014). That said, it can be assumed that the rapid technological development in the me- dia sector and the respective adaptation to these changes have led to different out- comes in various countries. The US, for example, have had a comparatively low level of political parallelism (Hallin & Mancini, 2004), and previous studies have empirically validated this assumption (Brüggemann et al., 2014). However, recent research points to a transformation of the U.S. media system, leading to increased political parallelism (Nechushtai, 2018). Moreover, several scholars have argued that increased media fragmentation (Van Aelst et al., 2017) and audience polarization (Fletcher et al., 2020) are linked to stronger political orientations by journalists, which were evident in the United States during recent electoral campaigns (Patterson 2016). Thus, it can be assumed that the digitalization of media systems has, in some cases, created conditions that have led to an increase in political paral- lelism. Therefore, we argue that the existing indicators of the dimension should be compared against current data to reflect recent political and social developments. 4 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age Journalistic Professionalism Although the rise of digital media has changed many aspects of the journalistic pro- fession, journalists still play an important role in news production and dissemina- Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 tion (Nielsen & Selva, 2018). Thus, the indicators that Hallin and Mancini (2004) employ in their theoretical framework to evaluate the dimension of journalistic pro- fessionalism in different countries remain fundamental. The dimension has been operationalized by Brüggemann et al. (2014) using the following indicators: (a) the internal and external autonomy of journalists, (b) the presence of professional guidelines and codes of ethics, (c) the degree to which news media enjoy credibility, and (d) the degree of journalists’ orientation toward public service. These indicators are arguably still adequate to examine how the journalistic profession has developed in recent years (Mattoni & Ceccobelli, 2018). For example, media credibility is con- sidered in terms of “fake news” allegations by politicians attacking the news media (Egelhofer & Lecheler, 2019). Researchers have argued that such allegations could cause lasting damage to trust in journalism and lead audiences to turn to alternative news sources (Fawzi & Mothes, 2020; Figenschou & Ihlebæk, 2019). However, me- dia systems differ in how professional and credible journalists are perceived to be. Therefore, such attacks may damage news media in countries with low levels of me- dia credibility more than in countries where established media brands enjoy high levels of trust. In this context, researchers (Phillips, 2010; Singer, 2010) have also highlighted the need for professional guidelines and codes of professional ethics tai- lored to online media, for instance, concerning uncivil user comments or hate speech (Frischlich et al., 2019; Quandt, 2018). Such guidelines combine traditional journalistic values and norms with the requirements of digital publications and user interactions. Addressing these challenges may be of greater necessity in some coun- tries than in others. Finally, journalistic professionalism in the digital age is also characterized by the quality benchmarks of the entire media industry. Many researchers have argued that the orientation toward sensationalism has led to an erosion of journalistic standards, especially on social media (Tandoc et al., 2018; Waisbord, 2018), and that journalists increasingly focus on click numbers instead of the quality of the content (e.g., Christin, 2018). Thus, we argue that the existence of quality criteria and norms that signal journalistic excellence for most journalists can still be regarded as an im- portant dimension of journalistic professionalism. Digitalization has deeply transformed journalistic routines and practices (Lecheler & Kruikemeier, 2016). One important change is related to the more active role of the audience, as follow-up communication is more visible and important than it was in the pre-internet era (Lawrence et al., 2018). Therefore, we argue that journalistic pro- fessionalism in the digital age is also characterized by how journalists deal with audi- ence participation and, in particular, the responsiveness of journalists to the public. For example, neglecting user comments can lead to incivility or misinformation being spread through the comment sections. Reading such comments can also influence how users perceive the content of news and the media (Anderson et al., 2014, 2018). Journal of Communication 00 (2022) 1–20 5 Media Systems in the Digital Age E. Humprecht et al. However, if journalists get involved in user discussions in the comment columns, the quality of such discussions can be increased (Masullo et al., 2020). In summary, in addition to internal and external autonomy, professional guide- Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 lines, media credibility, and public service orientation, we include online audience responsiveness as a new indicator for measuring journalistic professionalism (see Table III in the Supporting Information Appendix). Role of the State The dimension of the state’s role was described by Hallin and Mancini (2004) in terms of the actions a state takes to support and regulate the media sector. Brüggemann et al. (2014) argued conceptually and empirically that the “role of the state” is, in fact, not one dimension but a multi-dimensional construct, which they divided into three subdimensions referring to different types of actions that states can take to engage in the media sector: (a) funding and promotion of public broad- casters, (b) direct and/or indirect subsidization of news media organizations, and (c) the regulation of media concentration, ownership, and competition. However, sev- eral authors have argued that although state interventions still play an important role in shaping media systems, the type of interventions states undertake to govern the media sector have changed (Flensburg & Lai, 2020; Mancini, 2020). The regulation of media markets (understood as traditional mass media) has taken a back seat, as the big tech companies have become the most powerful and least regulated actors in the media system (Haggart & Keller, 2021). Social media platforms are also increasingly used for news consumption and dissemination. However, these companies are so far only regulated to a very limited extent by nation-states because they are mainly based in the United States and do not fall un- der the more traditional media regulation of the media sector, which has been slow to respond to technological change. More recently, there have been more and more proposals to promote greater le- gitimacy for platform regulation, primarily based on the rule of law principles and human rights laws and norms (Suzor et al., 2018; Kaye, 2019). Although questions about the appropriate balance of responsibilities remain, policymakers, such as the European Commission, have started making companies liable for illegal speech propagated via their services. Since platform regulations have been imposed mainly at the transnational (e.g., European) level, this indicator is less suitable for compar- ing national media systems, particularly since our study focuses on Europe. In contrast, the allocation of subsidies for private media plays an important role, especially in times of digitization. New online-only startups have to contend with fi- nancial problems because they first have to establish themselves on the market (Sirkkunen & Cook, 2012). Partial financing from state funds can therefore help to promote media diversity in a country. Some countries have done this by extending subsidy programs for newspapers to online media (Nielsen, 2013). 6 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age Media freedom is a central variable that shapes a media system (Norris, 2009). This is obvious for the study of media systems across the globe: in some states, criti- cal journalists risk being persecuted and imprisoned or killed, whereas in other Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 countries, they enjoy security and legal protection (Anagnostou et al., 2012; Brants & Voltmer, 2012). The study of media systems focusing on Western systems in the tradition of Hallin and Mancini (2004) was less concerned with this variable, as press freedom was taken for granted and not much variance was to be expected when studying Western Europe, the United States, and Canada. However, when in- tegrating Central and Eastern Europe into the analysis, it is an important variable that helps to cluster similar media systems (Castro Herrero et al., 2017). Also, after four years of populist rule by Donald Trump with constant (albeit so far mostly rhe- torical) attacks on the free press, full freedom of the press is no longer completely self-evident even in Western countries such as the United States. The argument that media freedom should be integrated into the study of media systems is also valid against the backdrop of digitalization, as online content is in- creasingly censored in some countries and independent journalists and bloggers are imprisoned (Voltmer & Wasserman, 2014). Media freedom is measured by the non-profit and non-governmental organiza- tion Freedom House and includes aspects of the legal, political, and economic envi- ronment in which media outlets operate. For example, the legal environment category encompasses examining laws and regulations that could influence media content or restrict the media’s ability to operate. The political environment category evaluates the degree of political influence in the content of news media, and the eco- nomic environment includes, among others, the structure of media ownership, pro- duction and distribution costs, and the extent to which the economic situation in a country affects the development and sustainability of the media (Freedom House, 2017). In addition, these indicators include aspects of the “role of the state” dimen- sion described by Hallin and Mancini (2004) and Brüggemann et al. (2014), political parallelism, and journalistic professionalism. However, Freedom House (2017) states that their criteria are selected and evaluated by country experts with a particu- lar focus on the role of different states in ensuring the right of freedom of expres- sion. Moreover, the methodology has recently been modified to capture changes in the news and information environment and incorporate digital media’s role. Therefore, we argue that while press freedom in the digital age is a multi- dimensional concept spanning various media system dimensions, it is primarily shaped by the state’s role in supporting or limiting these freedoms. Overall, we propose to include the following new indicators for operationalizing and measuring the dimensions of media systems in the digital age. First, for the di- mension of inclusiveness of the media market, we include the reach of online news as well as the reach of offline and online versions of established news media. Second, regarding the dimension of journalistic professionalism, we argue for adding jour- nalist’s online audience responsiveness. Third, we understand the dimension of the role of the state as state support for strong and free media, which entails considering Journal of Communication 00 (2022) 1–20 7 Media Systems in the Digital Age E. Humprecht et al. some variables that have already been included in prior studies, such as support of public service broadcasting and subsidies for online and offline news outlets. In ad- dition, we also consider media freedom as an additional indicator. Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 In the following sections, we will operationalize the indicators discussed above, drawing on many primary sources that provide valid and current data to create an empirical typology that will allow us to compare an enlarged sample of countries along these expanded dimensions. Method and Data Our operationalization of media system dimensions builds on the work of Brüggemann et al. (2014). To do so, we collected updated data for the existing indi- cators following the authors’ data collection strategy. We also included the new indi- cators discussed in the previous sections of this article (see Table 1, where old indicators are signaled with an asterisk). To create a more meaningful typology, we collected data for the countries initially described by Hallin and Mancini (2004) and expanded the sample to include Eastern and Southern European countries, resulting in 30 countries under study. We used several data sources to build new dimension indices. We drew on survey data from the Eurobarometer (Eurobarometer, 2017), the Digital News Report (Newman et al., 2018), expert interviews from the European Media Systems Survey (Popescu et al., 2013), V-Dem (Coppedge et al. 2021), and the Worlds of Journalism Project (Worlds of Journalism Association, 2016), and data from the European Audiovisual Observatory yearbook (European Audiovisual Observatory, 2017). We also employed aggregated data from a policy report by the Center for Media Pluralism and Media Freedom (Brogi et al., 2018). Our data analysis strategy also builds on Brüggemann et al.’s (2014) approach. However, in addition to the changes following the extended conceptualization of dimensions outlined above, the following operationalization had to draw on data sources that have been updated since the earlier measurement, leading to some changes in the indicators used. We z-standardized all variables since they were measured on different scales and combined them into the four dimensions: the inclusiveness of the media market, po- litical parallelism, journalistic professionalism, and state support for free media. In addition to the indicators operationalized by Brüggemann et al. (2014), we included the new indicators discussed above to account for media systems’ digitalization and consider media freedom. To address current changes regarding journalistic profes- sionalism, we included new data from the Worlds of Journalism Project on profes- sional guidelines for offline and online media, media credibility, public orientation, and online audience responsiveness (Worlds of Journalism Association, 2016). We also used data from the European Media Systems Survey on excellence criteria in journalism, such as generally agreed-on criteria for judging excellence in the jour- nalistic profession and perceived media credibility (Popescu et al., 2013). 8 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age Table 1 Dimensions, Indicators, and Data Sources Dimension Indicators Data source Cronbach’s a Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 Media Market Daily newspaper  Eurobarometer.86 reach 2017 Working class news-  DNR 2018 paper reach Women newspaper reach Online news reach Political Partisan influence &  EMSS 2013.86 Parallelism policy  WJS 2012-2016 Advocacy  V-Dem 2021 Political orientation  Media Pluralism of journalists Monitor 2017 Media party parallelism Political bias PSB dependence Journalistic External autonomy  EMSS 2013.67 professionalism and internal  WJS 2012-2016 autonomy Professional guidelines Criteria for excellence Media credibility Public service orientation Online audience responsiveness State support Market share of  EAO 2017.82 public TV  Media Pluralism Support to the me- Monitor 2017 dia sector  Freedom House Media freedom 2017 Note. Indicators apply to online and offline media unless otherwise indicated. To account for state support for the media, we relied on data from the Media Pluralism Monitor (Brogi, 2018), which considers three sub-indicators: spectrum al- location, government subsidies, and the state’s advertising rules. The first sub- indicator on spectrum allocation assesses the existence and implementation of a le- gal framework that enacts the general regulatory principles and policy objectives of the European Commission’s “Radio Spectrum Policy Program,” including rules for the authorization for digital terrestrial TV. The second sub-indicators on state subsi- dies include direct and indirect media subsidies, such as cash grants and interest- Journal of Communication 00 (2022) 1–20 9 Media Systems in the Digital Age E. Humprecht et al. free loans, tax exemptions, reduced postal service, and telephone rates. The third in- dicator on advertising rules refers to any advertising paid by governments (national, regional, and local) and state-owned institutions and companies to the media. This Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 operationalization of state support differs primarily from previous measurements in that we considered more aspects of state influence, including regulation of digital TV. In contrast, previous studies have focused primarily on ownership regulation of TV and the press. In addition, we used data from the European Audio Observatory (EAO, 2017) to measure the market share of public TV. Finally, media freedom was measured using data from Freedom House (2017). The Freedom House index includes indicators related to the legal environment of the media (e.g., laws and reg- ulations influencing media content), the political environment (e.g., the degree of political influence in the content of news media), and the economic environment (e.g., economic influence on the media sector). The exact wording and operationali- zation of all indicators used, and data sources are summarized in Tables I–IV in the Supporting Information Appendix. Next, we conducted a hierarchical cluster analysis to validate previous typologies of media systems. To do so, we followed the procedure described by Brüggemann et al. (2014). To identify the clusters, we used Ward’s method and applied squared Euclidian distances to measure similarity (for similar approaches, see Castro Herrero et al., 2017; Perusko et al., 2015). Based on an analysis of the scree plot and a dendrogram, we iden- tified a three-cluster solution. To optimize the cluster membership of the countries and validate the analysis, we also used the centroid-based k-means method (Milligan, & Sokal, 1980), which widely confirmed our cluster solution. Results This study aims to answer two research questions: whether indicators linked to the digitalization of media systems can be combined with Hallin and Mancini’s (2004) original dimensions into an integrated framework (RQ1) and which media system typology can be built based on the integrated framework and the enlarged country sample (RQ2). To answer RQ1, we examined the internal consistency of the newly formed dimensions, which yielded satisfactory values of Cronbach’s a, that is, a ¼.86 for the dimension of the media market, a ¼.86 for political parallelism, a ¼.66 for journalistic professionalism, and a ¼.82 for state support. These comparatively high reliability values (DeVellis, 2003; Hair et al., 1998) confirm that our theoretical considerations regarding the correlation between original and new indicators are empirically supported. Next, we conducted correlation analyses, which confirmed the results of previ- ous studies regarding the relationship between the dimensions. Political parallelism correlates negatively with all other dimensions, thus: the inclusiveness of the media market (.59, p <.01), journalistic professionalism (.47, p <.01), and state sup- port (.59, p <.01). 10 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age Furthermore, journalistic professionalism is positively correlated with the dimensions of inclusiveness of the media market (.50, p <.01) and state support (.60, p <.01). The dimensions of inclusiveness of the media market and state sup- Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 port also show a positive correlation (.69, p <.01). To answer RQ2, we conducted a hierarchical cluster analysis, which resulted in three clusters. We examined the three clusters regarding country membership and the characteristics of their assigned media systems to label them appropriately (Figure 1). The first cluster consists of Nordic countries (Finland, Denmark, Norway, and Sweden), German-speaking countries (Austria, Germany, and Switzerland), and the Netherlands. This country sample corresponds largely to Hallin and Mancini’s (2004) democratic-corporatist model and combines Brüggemann et al.’s (2014) Central and Northern models. As Table 2 shows, the cluster is characterized by a high degree of journalistic professionalism, an inclusive media market, a strong role of the state, and low degrees of political parallelism. Due to the similarity of the clus- ter’s characteristics to Hallin and Mancini’s (2004) democratic-corporatist model, we adopt this label for our first cluster. The next cluster includes Eastern and Southern European countries, such as Bulgaria, Cyprus, Croatia, Greece, Hungary, Latvia, Montenegro, Poland, Romania, Slovakia, Slovenia, and Spain. This cluster differs from the democratic-corporatist cluster in having a weaker media market, lower levels of journalistic professionalism, lower levels of state support, and higher levels of political parallelism (Table 2). These characteristics largely correspond to Hallin and Mancini’s (2004) polarized- pluralist media system model, although the cluster in our analysis additionally includes Eastern European countries. Therefore, we adopt this label and name it the polarized-pluralist model to highlight this cluster’s similarity with the original model. The last cluster is situated between the two extremes of the democratic- corporatist and the polarized-pluralist cluster. The countries in this cluster show Journal of Communication 00 (2022) 1–20 11 Media Systems in the Digital Age E. Humprecht et al. Table 2. Cluster Profiles Media market Journalistic professionalism Political parallelism State support Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 Democratic- 1.11 High 0.46 High 0.62 Low 0.93 High Corporatist Hybrid 0.09 Medium 0.06 Medium 0.44 Medium 0.27 Medium Polarized-Pluralist 0.62 Low 0.28 Low 0.68 High 0.81 Low Note. Values are means of z-standardized dimension indices. lower state support and journalistic professionalism than countries in the democratic-corporatists cluster and slightly lower levels of political parallelism. Moreover, the inclusiveness of their media market is significantly weaker compared to countries in the democratic-corporatist cluster. The cluster consists of countries that have been assigned to the liberal media system model (the United States, the UK, and Ireland), the polarized-pluralist model (France, Italy, and Portugal), and the democratic-corporatist model (Belgium) by Hallin and Mancini (2004). In addi- tion, the cluster includes three Eastern European countries (the Czech Republic, Estonia, and Lithuania). The cluster’s composition is similar to the Western model of Brüggemann et al. (2014), with France and Italy as exceptions. In summary, our analysis showed that original and new indicators of media sys- tem dimensions can be combined into a meaningful framework. Moreover, compar- ing a broader country sample using the extended framework resulted in three media system clusters. Two of the clusters identified in our analysis largely correspond to models identified by Hallin and Mancini (2004), namely the polarized-pluralist and democratic-corporative media system models. However, the hybrid cluster includes a range of countries that have been assigned to other system types before but were described in the literature as borderline cases. Discussion The results of our study provide a multifaceted picture of how media systems vary in the digital age. To better understand how different media systems may have evolved in recent years, we discuss our clusters and compare them with the typology by Brüggemann et al. (2014). Although our analysis does not allow for a direct com- parison due to divergent measurements, we can draw conclusions about which countries show greater similarities and formulate assumptions on which countries are likely to have converged due to recent developments. That said, we emphasize that quantitative approaches such as cluster analysis do not allow us to describe me- dia systems in their entirety. Therefore, the choice of indicators and countries strongly affects the cluster solution. The first important observation is that the countries in the so-called Northern model of Brüggemann et al. (2014), which comprises Denmark, Finland, Norway, 12 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age and Sweden, seem to have stable media systems with strong media organizations. They all fall into the democratic-corporatist model in our analysis. Second, from the so-called Central model of Brüggemann et al. (2014), the UK is split from Austria, Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 Germany, and Switzerland in our extended and more current comparison between countries. As our analysis shows, the German-speaking countries show greater simi- larities to the Nordic countries and fall into the democratic-corporatist cluster. The UK, in contrast, shares more characteristics of countries such as Ireland, Belgium, Portugal, and the United States, which Brüggemann et al. (2014) refer to as the Western model. The media system of the UK differs from the media systems of the German-speaking and Nordic countries in particular because of the lower inclusive- ness of the media market. Third, the media systems of Southern European countries seem to vary with regard to state support and political parallelism, splitting the so- called Southern model of Brüggemann et al. (2014), which comprises France, Spain, Greece, and Italy. Italy, for example, falls into the hybrid cluster but is ranked rather low regarding journalistic professionalism and high in political parallelism, thus rep- resenting a borderline case to the polarized-pluralist cluster. Fourth, Eastern European media systems are not homogenous but divided between the polarized- pluralist and the hybrid cluster. The media systems of the Southeastern European countries—Hungary, Romania, Bulgaria, Montenegro, Croatia, Slovenia, and Slovakia—seem to be similar to the media systems of Southwestern European coun- tries, such as Spain, Greece, and Cyprus. In contrast, the media systems of the northeastern European countries Estonia and Lithuania share more similarities with the United States or the UK. Finally, we find a cluster of hybrid countries that share features of Brüggemann et al.’s (2014) Western model, including Ireland, Belgium, Portugal, and the United States. This cluster is situated between the extremes of the democratic-corporatist and the polarized-pluralist clusters and is characterized by medium values on all four media system dimensions. Several authors have argued that countries such as Belgium, France, Italy, Portugal, the UK, and the United States do not correspond to the ideal types defined by Hallin and Mancini and that they are “border” or “mixed” cases regarding their media systems (Büchel et al., 2016; Hallin & Mancini, 2004). This is the rationale for labeling this cluster hybrid. For example, the United States show similarities to certain European countries in several aspects (Nechushtai, 2018) regarding its level of political parallelism. Previous studies also showed that France, Italy, and Portugal are less similar to other Southern European countries, such as Spain and Greece, and more similar to, for example, Belgium or the UK, particularly regarding state support for the media sector (Büchel et al., 2016; Hallin & Mancini, 2004). The hybrid cluster includes, among others, countries whose media systems have been described in the literature as contrasting, for exam- ple, France and the United States (e.g., Benson et al., 2012). Indeed, our analysis shows differences between these two countries, especially regarding the state’s role and political parallelism. The analysis also reveals that the United States remains quite unique (grouped only with Lithuania in the first step of clustering). The three Journal of Communication 00 (2022) 1–20 13 Media Systems in the Digital Age E. Humprecht et al. Eastern European countries in this cluster also seem to show lower levels of political parallelism and higher levels of press freedom than many of their neighbors in the region (Castro Herrero et al., 2017) that fall into the polarized-pluralist category. Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 Moreover, the fact that the countries in the hybrid cluster form one group may be due in part to the greater variance caused by the inclusion of additional countries from Central and Eastern Europe in the equation. In earlier work on Eastern European media systems, the term “hybrid” was used to describe a mixture of the liberal model and the polarized-pluralist model, especially with regard to the com- mercialization of journalism and the instrumentalization of media by political actors (Dobek-Ostrowska & Glowacki, 2016; Perusko, 2021). With regard to media sys- tems in Central and Eastern Europe, Perusko (2021; p. 42) argues that “hybridity can be seen as a sensitizing concept that brings attention to the shifting environment of media systems and the new institutional arrangements coming about with the in- tegration of values, norms and practices, and procedures and genres (logics) from different media eras.” Our analysis shows that the combination of lower journalistic professionalization, non-inclusive media market, and little state support unites sev- eral Eastern as well as Western European countries and the United States. For exam- ple, the United States and France as different as those countries may be, have more in common with the Czech Republic or Lithuania than they have with many coun- tries in Northern or Southern Europe. This comparison also underscores how cru- cial the country selection is for the study of media systems. Our study constitutes a small step forward beyond the study of Western Europe and the United States. Which broader implication can be drawn from our study? First, it can be as- sumed that the clustering of the United States with Southern and Eastern European countries reflects the outcome of an increasing polarization that has been observed in some of these countries in recent years (Mancini, 2020). Second, our findings also seem to confirm a trend that was already visible in Brüggemann et al. (2014) com- pared to Hallin and Mancini (2004): the demise of the liberal media system. Following previous studies, we do not find empirical evidence for a configuration of low state support for free media and high journalistic professionalism, broad news reach, and low parallelism as envisaged by the ideal type of the liberal system. This may be due to the increased need for state support for strong free media in times of digital media change, where elite newspapers, as the traditional backbone of profes- sionalism, suffer from an enduring revenue crisis. Journalistic professionalism and broader news reach are, on the contrary, fairly stable in the Nordic and German- speaking countries. In this regard, the findings from previous typologies, such as those of Brüggemann et al. (2014) and Büchel et al. (2016), seem robust. Overall, our findings confirm many of the assumptions and observations that have been expressed in recent research in the field (Hallin, 2020; Liu et al., 2020; Mancini, 2020). However, the current lack of up-to-date typologies, including aspects of digital media technologies, demonstrates the need to assess how digital communication is structured within different societal settings. By providing an operationalized framework and an empirical analysis of an expanded country sample, we hope to stimulate future 14 Journal of Communication 00 (2022) 1–20 E. Humprecht et al. Media Systems in the Digital Age research in comparative communication science and inspire theoretical work on the current transformation of various media systems. Naturally, this study has limitations that are related to data availability and oper- Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 ationalization. For example, due to availability constraints, we could not collect updated data for some of the original indicators used by Brüggemann et al. (2014). This applies particularly to the “separation of news and commentary” indicator of the political parallelism dimension. As argued before, however, the separation be- tween news and commentary seems to be declining in online information environ- ments; hence, this indicator may have lost some of its significance. Furthermore, for the dimension of the state’s role, we had to rely on a data source other than the one used by Brüggemann et al. (2014) because the World Press Trends report did not cover all countries in our sample. Therefore, we had to rely on a recent policy report by the Center for Media Pluralism and Media Freedom, European University Institute (Brogi et al., 2018). Moreover, in three cases, missing values had to be replaced with other data sour- ces, and we had to use a nominal scale that resulted in lower variance. However, we are confident that the applied data sources represent our theoretical assumptions. We also find that the data from various sources are highly correlated. Another limitation of our study is that our data sources represent only one or a few points in time. However, the underlying theoretical considerations are related to processes of change, such as digitalization. Therefore, comparing our typology against previous studies provides a first approximation of the extent of such changes (e.g., countries closer to models other than those Hallin and Mancini originally asso- ciated them with). Nevertheless, to examine these processes more accurately, long- term studies are necessary. Finally, we want to point out that the data and results for individual countries should not be over-interpreted, as they only allow for approximations. As we have already mentioned, the results of quantitative typology building de- pend on the countries and indicators chosen. If we were to perform the same analy- sis with a larger, more diverse sample of countries, including, for example, countries under authoritarian rule, other differences or commonalities (e.g., the existence of independent media) might be more pronounced. Thus, qualitative analyses are needed to make more in-depth statements about individual countries. Therefore, we agree with Hallin (2020) that digitalization and recent social and political systems changes should prompt researchers to “conceptualize both change and complexity in media systems.” That said, with our expanded sample and the new indicators that take into ac- count the digitalization of media environments, our study extends and complements Hallin and Mancini’s (2004) framework. Our study also provides several points of departure for further research. For instance, our newly developed indicators and dimensions can be used in future studies to explain differences between media sys- tems regarding news content or media use. Such cross-national comparisons may be of particular interest regarding the media systems we have assigned to the hybrid Journal of Communication 00 (2022) 1–20 15 Media Systems in the Digital Age E. Humprecht et al. cluster. In addition, future longitudinal research should seek to understand why po- litical parallelism has increased in some countries and what role media organiza- tions and state institutions play in this regard. We hope that the framework will be Downloaded from https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqab054/6509144 by University of Zurich user on 18 January 2022 extended to other countries. Such an expansion would allow researchers to identify additional dimensions that may be relevant for comparing media systems. Also, even though we have already used an expanded sample, the underlying theoretical assumptions are largely based on European and North American literature. Therefore, further studies are needed to extend this knowledge, especially in Asia, Africa, and Latin America. Finally, challenges remain in identifying the relevant factors that shape different media systems in recent years. The development of ICTs has not only changed me- dia production and media use in many ways. It also shapes the way politicians com- municate, for example, by addressing citizens directly through social media and enabling the dissemination of sometimes misleading information through alterna- tive news providers. Therefore, comparative research is fundamental to understand- ing how well different media systems are equipped to tackle these challenges. Supporting Information Additional Supporting Information may be found in the online version of this arti- cle. Please note: Oxford University Press is not responsible for the content or func- tionality any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. References Aalberg, T., Blekesaune, A., & Elvestad, E. (2013). Media choice and informed democracy: Toward increasing news consumption gaps in Europe? The International Journal of Press/Politics, 18(3), 281–303. https://doi.org/10.1177/1940161213485990 Anagnostou, D., Avadani, I., Craufurd Smith, R., Docquir, P.-F, Gusy, C, Harro-Loit, H,... Van Besien, B. (2012). Media freedom and independence in 14 European countries: A comparative perspective. https://www.rcmediafreedom.eu/Publications/Reports/Media- freedom-and-independence-in-14-European-countries-A-comparative-perspective Anderson, A., Brossard, D., Scheufele, D., Xenos, M. A., & Ladwig, P. (2014). The “nasty ef- fect:” Online incivility and risk perceptions of emerging technologies. Journal of Computer-Mediated Communication, 19(3), 373–387. https://doi.org/10.1111/jcc4.12009 Anderson, A., Yeo, S. K., Brossard, D., Scheufele, D., & Xenos, M. A. (2018). Toxic talk: How online incivility can undermine perceptions of media. International Journal of Public Opinion Research, 30(1), 156–113. https://doi.org/10.1093/ijpor/edw022 Benson, R., Blach-Ørsten, M., Powers, M., Willig, I., & Zambrano, S. V. (2012). Media sys- tems online and off: Comparing the form of news in the United States, Denmark, and France. Journal of Communication, 62(1), 21–38. https://doi.org/10.1111/j.1460-2466. 2011.01625.x 16 Journal of Communication 00 (2022) 1–20