Comparative international studies of election campaign communication: What should happen next?

Setting the scene

Research into election communication is getting more international. However, it is still in a process of development and maturation, although the situation has improved since de Vreese (2003) stated that ‘evidence from cross-national comparisons of national elections is virtually non-existent’ (p. 184). The situation is improving only slowly because the group of scholars engaging in systematically comparative communication research is still relatively small. Comparative research is favored when researchers have access to or can establish an international network of experienced partners; when they have resources available for meetings and data collection in different countries; and when they have a small team at their home institution that can secure, coordinate and repeatedly revitalize the transnational collaboration. Such an ideal scenario is not easy and, above all, not continuously achievable for most researchers, and this partly explains the gradual progress in this field. On the other hand, large does not always have to mean better, and small, original, in-depth comparative studies also make a significant contribution to progress. There are a number of instructive reports from Cohen (2012), Esser and Hanitzsch (2012), Hanusch and Hanitzsch (2017) as well as Livingstone (2003) on how to organize small and large comparative research projects.

This article highlights the great potential of comparative international approaches for a deeper understanding of current developments in election campaign communication. Thus, comparative analysis can offer a counterweight to the strong Anglo-American character of scientific debates and assumptions around campaigns; it can provide a vaccine against US centrism and naive overgeneralizations, so to speak. To underline this point, Vaccari (2013) even speaks of online communication in the United States as a ‘deviant case’ that cannot be generalized to the conditions in other Western democracies he examined – such as Australia, France, Germany, Italy, Spain, or the United Kingdom (p. 4). The lack of comparative international studies of election campaign communication is a regrettable weakness and this article calls for increased efforts to remedy it. This is particularly the case in election reporting, which is the main focus of this Special Issue. It arose out of a preconference of the International Communication Association entitled ‘Making sense of election reporting – new directions, new challenges?’ This preconference took place 2018 in Prague and I presented this article as the keynote.

However, such a call for more comparative research is easier to make than to put into practice. The situation has become complicated: a multitude of political and media actors (parties, candidates, news journalists, and social media producers) are nowadays involved in the creation of election campaign messages; and in spreading these messages, old and new channels (offline and online) work together and reinforce each other. This heightened complexity has greatly increased the requirements for studies that also take into account the conditions in different countries. In the following, I would like to highlight the achievements of comparative international studies of election campaign communication to date, while also acknowledging their challenges. I conclude my observations with some suggestions on the direction future comparative studies could or should take.

The value of comparative research for election communication research

Why is the lack of comparative research regrettable? The answer is because we know from the comparative studies available that both institutional and cultural context factors have a great influence on election campaign communication – and that these context factors cannot be systematically identified with individual case studies. Esser and Strömbäck (2012a) already argued a few years back that country-specific conditions determine which of the ‘ideal types’ of campaign communication (e.g. the modern campaign, the postmodern campaign, the marketing-oriented campaign, or the digital campaign) is ‘realized’ in a concrete setting. In theoretical terms, these contextual conditions constitute characteristic patterns of constraints and opportunities, which help explain how and why campaigns differ from country to country or why they diverge from a theoretical baseline model.

For empirical research, this knowledge is essential for forming hypotheses and making predictions about the context-dependency of campaign styles. Comparativists have developed comprehensive theoretical models that identify relevant contextual conditions (as explanatory variables) and have also argued for the need to measure campaigns styles identically across countries to learn something about their meaningfulness and determinants (for details, see Esser and Strömbäck, 2012a)

In addition, comparative studies are important for researching international diffusion effects between certain groups of countries, which play a major role in election communication. A recent study by Kiyohara et al. (2018) entitled ‘Internet Election Campaigns in the United States, Japan, South Korea, and Taiwan’ has shown that although there are exchange relations between these countries, the idea of simple Americanization in the spread of digital campaign technologies is unjustified. Particularly in the case of Japan, several contextual factors work against Americanization – including draconian campaign regulations, low Facebook penetration and an unstable party system in Japan.

The fourth era of political campaign communication

Campaign communication is a dynamic object of investigation. In this article, I will focus on research on election reporting, but I will also deal again and again with studies on the messages of other communicators in election campaigns (candidates and parties) in order to illustrate the diversity of comparative content analyses in the context of campaigns. In most countries, candidates, parties, and the media communicated very differently 30 years ago than they do today. While single-country case studies like to concentrate on the latest developments and innovations, multicountry comparisons often lag somewhat behind the most up-to-date trends because they focus more on fundamental structural factors. This becomes clear when we look at the various ‘eras’ of political communication or ‘periods’ of election campaigning proposed in the literature. Schmitt-Beck and Farrell (2002) suggested that we ought to think about eras broadly corresponding to the newspaper age, the television age, and the digital age. A similar argument was made by Norris (2002), who identified premodern, modern, and postmodern eras of campaigning and later redefined these as people-intensive, broadcast-based, and Internet-based eras (Norris, 2005).

The current ‘fourth era of political communication’ (Blumler, 2016) involves new campaigning tools, techniques and capabilities opened up by the rise of big data technology – and the impact it has on computational propaganda and election reporting. Semetko and Tworzecki (2018) argue that two further characteristics – in addition to advances in data mining, data analytics, microtargeting, automated bots, and algorithms – warrant the claim of a new fourth era: first, the rising importance of social media platforms (Facebook, Twitter, and their counterparts in other countries) as content delivery mechanisms for political news and information and second, the border-transgressing diffusion of campaign technologies and strategies thanks to a global network of media, consultants, and other hidden channels that span democratic and authoritarian regimes alike.

However, these alleged innovations are all gradual in nature. Transnational diffusion and adaptation processes have always existed, and the growing importance of social media and algorithmically distributed messages based on personalized big data does not replace earlier forms of mass communication but rather complements them. Therefore, in the context of the fourth era of political communication, Chadwick (2017) aptly speaks of ‘hybrid’ communication relationships, where older and newer media operate simultaneously, interact with each other, and create relationships of adaptation and interdependence – despite the inevitable competition that also exists between them.

I have made the observation – and this is one of the central messages of this article – that comparative international content analyses of election campaign communication have not yet even begun to do justice to the hybrid realities of the fourth phase of political campaign communication. Very few studies deal with the parallel existence and cross-fertilization of traditional news media and new social media platforms. Researchers have yet to rise to the challenge of developing research designs that can comparatively and systematically analyze the fluidity of election reporting across new online and social media platforms.

There are isolated comparative studies that investigate the use of social media and traditional media (in terms of their relevance for political participation) in different countries (e.g. Saldaña, McGregor and Gil De Zúñiga, 2015), but these studies are not linked to election campaigns. There are also surveys of campaign managers asked to assess the importance of classic and new communication channels in order to obtain clues about cross-national differences in levels of professionalization (e.g. Lilleker et al., 2015; Tenscher et al., 2016). However, in these survey studies, we learn nothing about differences in the content of campaign messages being communicated.

Back to the current state of comparative international campaign communication research

Many examples illustrate that research into election campaign content – especially election reporting in the news media – is becoming increasingly international. Thus, in their book ‘Reporting elections’, Cushion and Thomas (2018) exploit the growth of comparative studies in journalism and political communication literature by drawing on well over 100 empirical studies to examine the different types of logics shaping campaign coverage across different countries. While the book focuses on UK and US election reporting rather than developing a large N study of many countries, the analytical approach is consistent with a series of conceptual recommendations that Jesper Strömbäck and I had formulated some time ago for cross-national content analyses of election news coverage.

In a programmatic article (see Esser and Strömbäck, 2012b), we had recommended that further comparative research should agree on a core set of dependent variables and that the relevance of this core set of reporting features should be derived from democracy theory. We had suggested as variables media depoliticization (with policy coverage being marginalized, sensationalized, or strategy-framed), personalization (with stories being leader-centered instead of institution-centered), interventionism (with an interpretive style being used to enhance journalistic voice and diminish political voice), and negativity (characterized by negative topics and tonality, confrontation, and conflict) – potentially supplemented with media bias and media diversity. We further argued that the pervasiveness of these reporting features in election news is dependent on the transnational spread of a certain kind of ‘news logic’, linking this research to the mediatization of politics literature. Comparative research, we argued, is interested in describing and explaining transnational similarities and nation-specific differences in news logic. It is further interested in classifying national reporting styles according to relevant dimensions and assessing news performance according to democratic norms (Esser and Strömbäck, 2012b).

To make this possible, we need more studies comparing more (than two) countries and investigating an agreed-upon canon of variables with comparable operationalizations. Comparative studies should further recognize the multilevel structure of election news production and investigate explanatory relationships between contextual conditions (independent variables) and reporting features (dependent variables). The studies should be hypothesis-led and explanatory. We designed a multipage catalog of sample hypotheses as a suggestion. We closed our chapter with the following statement:

Only through a more thorough and explicit conceptualization of key concepts will it be possible to develop valid and reliable operationalizations, and only through more standardized operationalizations will it be possible to increase the comparability of findings and research cumulativity. (Esser and Strömbäck, 2012b: 323)

Large comparative studies have sometimes had problems in achieving high intercoder reliabilities for more demanding, complex content-analytical concepts, and they can, at times, provide more breadth than depth when interpreting their findings. Hence, smaller scale empirical studies as well as critical reviews of the available comparative literature can also contribute to a deeper understanding of cross-national interrelationships, continuities, and characteristic variations, even if they cannot reach the scientific precision of an all-encompassing, large-N comparative analysis.

Since our chapter appeared, Jay Blumler and Denis McQuail (2016) have also reviewed the comparative international literature on election campaign communication. Blumler’s study ‘Communicating to Voters: Television in the First European Parliamentary Elections’ (1983) provided the initial impetus for this line of research. In a retrospective on this pioneering study, Blumler and McQuail (2016) also emphasized the importance of hypothesis-driven and explanatory research designs. They further stressed that comparative research must address issues of transnational relevance (such as populism, immigration, and security) and that cross-national content analyses must take into account new forms of online news and social media. In light of these various recommendations, I would like to take a closer look at the recent comparative international research on the content of election campaign communication.

A closer look at recent studies on election campaign reporting

Fortunately, recent research is indeed increasingly focused on a core set of relevant dependent variables; moreover, the growing number of high-quality studies on these variables ensures that we constantly know more about characteristics of international election campaign communication. Many of the studies discussed in more detail below establish a connection to the mediatization of politics scholarship; however, the link to democratic norms of news performance – something that Jay Blumler has always demanded from comparative political communication research (see, Blumler and Coleman, 2015; Esser, 2015) – usually remains weak. Classification of country findings by group is also rare, as most studies compare only two or three countries. Another striking feature is that the research designs of most current campaign content analyses are rather descriptive – without systematically including explanatory variables in the data analysis. It should be acknowledged that most of these descriptive studies strive for interpretative post hoc rationalizations in their conclusions; however, these causal assumptions were not systematically examined (quantitatively or qualitatively) using clearly identified explanatory variables in the analysis. Granted, rigorous statistical tests make little sense in two- and three-country studies, and their scarcity is a logical consequence of small country samples. Multicountry studies with sophisticated causal data analysis remain the exception in the field of comparative campaign news content analysis. As we will see below, even the most recent publications content-analyze exclusively traditional media (press outlets or television news broadcasts). This is perhaps understandable given their long held dominance, particularly in television news programming, which still, in many countries, reaches the most amount of people during an election campaign. But since voters increasingly rely on online and social media platforms for news, many studies have not kept up with the reality of political communication in a ‘hybrid’ campaign environment codetermined by digital media.

It is to be welcomed that most recent comparative campaign news content analyses investigate a similar core set of reporting features, which at least in principle enables comparability of findings and cumulativity of knowledge. This core set of often-used coding categories includes (1) media visibility: prominence of political actors in the news; (2) depoliticization: ratio of policy versus non-policy issues; (3) media personalization: focus on top candidates and their personal lives and personality traits; (4) game framing: use of strategy and horserace scenarios in news; (5) tonality: valence of positive, negative, or neutral evaluations; (6) negativity; (7) conflict; (8) interventionism; and (9) response to populism (see Table 1 for details).

OPEN IN VIEWER

Table 1.

Variables used in recent comparative international content analysis of election news.

Concepts	Studies
Media visibility of political actors	Holtz-Bacha et al. (2014); Magin (2015); Nord et al. (2017); Schaerdel and Koenig (2017); Vliegenthart et al. (2010)
Depoliticization: policy versus non-policy frames	Magin (2015); Nord et al. (2017); Schaerdel and Koenig (2017); Vliegenthart et al. (2010); Van Aelst et al. (2019)
Media personalization: focus on top candidates and their personal lives and personality traits	Holtz-Bacha et al. (2014); Magin (2015); Nord et al. (2017); Vliegenthart et al. (2010); Van Aelst et al. (2019)
Strategy and horserace framing	Nord et al. (2017); Schaerdel and Koenig (2017); Vliegenthart et al. (2010); Van Aelst et al. (2019)
Valence: evaluation of political actors	Magin (2015); Nord et al. (2017); Vliegenthart et al. (2010); Van Aelst et al. (2019)
Negativity: actor evaluation and story tone	Magin (2015); Schaerdel and Koenig (2017); Van Aelst et al. (2019)
Conflict	Schuck et al. (2013, 2016)
Interventionism: descriptive-neutral versus critical-aggressive reporting and interviewing; political voice versus journalistic voice	Ekström and Firmstone (2018); Magin (2015)
Response to populist parties: valence and style	Ekström and Firmstone (2018); Schaerdel and Koenig (2017)

However, a basic comparison of the findings is made more difficult by the fact that these reporting features are defined and evaluated differently in the various studies. In addition, they are operationalized and interpreted in different ways. There is still a lack of awareness of consolidated instruments and cumulative knowledge production among researchers focusing on comparative international campaign coverage. However, this criticism must be put into perspective because the individual comparative studies in themselves provide valuable insights, as the following overview illustrates.

Holtz-Bacha et al. (2014) presented a binational analysis of press coverage of the 2009 and 2010 election campaigns in Germany and the United Kingdom, analyzing the extent of media personalization in 680 stories from three German and five British newspapers. They find both differences and similarities in terms of ‘media attention for top candidates’ and ‘attention for their personal lives’. The study works without explicit hypotheses or research questions, but it does offer some qualitative explanations for the observed country differences, referring mainly to the factors of the political system and press market. For a comprehensive understanding of the differences, personality traits of the candidates must also be taken into account.

Another binational content analysis of press coverage comes from Magin (2015). She compared four German and two Austrian newspapers from 1949 to 2009, analyzing visibility, tonality, personalization, depoliticization, and interpretation in 8000 election stories. She found great similarities between the two countries, but the indicators of news logic she examines fluctuate erratically without recognizable patterns over the decades. Based on these findings, Magin questions the idea of mediatization as a historically unilinear and transnationally uniform process. However, she has difficulty offering clear explanations for her cross-temporal and cross-national findings. Nevertheless, her long-term observations remain illuminating and hint at the importance of contextual factors in understanding national reporting trends.

The study by Nord et al. (2017) also belongs to the group of descriptively oriented content analyses of election campaign reporting by traditional news media. The authors examined 933 articles from four British, four Italian, and four Swedish daily newspapers on their coverage of the 2016 US presidential campaign. They found differences in the degree of visibility that the US election received in the three countries’ newspapers. Apart from that, however, they found mostly similarities. The newspapers across all countries were very similar in the high level of attention they granted Trump, in their negative evaluation of Trump, and in their heavy focuses on scandal and horserace/game frames in their campaign stories.

Despite its timeliness, the study by Nord et al. (2017) does not reach the analytical level of a previous study on the same topic. A few years earlier, Vliegenthart et al. (2010) had examined the election coverage of 17 newspapers from eight European countries about the 2008 US presidential campaign. In the 1700 stories coded, the authors found clear differences in visibility, tone, and strategy/horserace framing; they explained those differences by considering country characteristics, media features, and temporal aspects. The results demonstrated that all three explanatory factors determined campaign coverage to some extent; however, temporal aspects, particularly the stage of the campaign and polling trends, were of highest influence. Interestingly, the same research team repeated this study using the 2016 US presidential election as an example (Van Aelst et al., 2019). Eight years after the initial study, many country differences had disappeared. Campaign coverage in American and European news media hardly differed in tone (negativity) and substance (issue vs strategy coverage); there was convergence due to shared news values, intermedia agenda setting, and event-driven campaign specifics (Van Aelst et al., 2019). Using different types of regressions, both studies of this research team are fine examples of explanatory comparative analysis.

Schaerdel and Koenig (2017) published another descriptively oriented content analysis of the campaign coverage of traditional news media in three countries. They studied how six quality newspapers from Austria, Germany, and Italy covered three populist Eurosceptic parties – the Team Stronach, the Alternative for Germany, and the Five Star Movement – during national elections in 2013. Their analysis of approximately 1500 articles found substantial cross-national commonalities: campaign coverage in all newspapers showed a high level of negativity and game/horserace aspects, combined with a de-emphasis on policy issues. The six newspapers converged further in their depiction of the three parties as dangerous, unreasonable, and incompetent; the parties were presented as irresponsibly populist and challenging the status quo. However, there were also substantial cross-national differences. The reporting on the Alternative for Germany was less personalized and more issue-oriented (mostly by pointing out its Eurosceptic stance), whereas coverage of Team Stronach and the Five Star Movement was more personalized, game-centered, and negative. While the Alternative for Germany was portrayed as a Eurosceptical single-issue party, Team Stronach and the Five Star Movement were presented as more broadly populist and as parties that use Euroscepticism as just one of several facets of their orientation.

The topic of populism was included in another descriptively oriented content analysis of election campaign reporting. Ekström and Firmstone (2018) analyzed 20 hours of TV news bulletins plus 80 hours of election specials and current affairs programs in the 3 weeks before the 2014 European Parliamentary (EP) elections in France, Greece, Italy, Sweden, and the United Kingdom. They did not take a quantitative approach like all other studies discussed so far but instead undertook a qualitative discourse analysis. The authors found that television stations in the five countries reported the 2014 EP elections primarily through a national lens and tied their coverage to national issues. In addition, there were clear cross-national differences in reporting political news. For example, the study shows how journalists in the United Kingdom have developed more interpretive and skeptical styles of reporting that differ from the more informative approach in French news. The reporting styles were constructed not only through the framing and editing of stories but also through the mode of interview questions, which were most aggressive on British TV.

A noteworthy cross-national difference concerned the coverage of populist political leaders. Here, Ekström and Firmstone (2018) identified three strategies with which journalists responded to populists in their coverage of the 2014 EP election:

However, as a limitation of the study, it must be noted that the central theoretical concepts ‘reporting style’ and ‘journalistic responses to populist discourse’ were not clearly operationalized; their examination and interpretation was performed qualitatively hermeneutically. Finally, Ekström and Firmstone (2018) stated explicitly that it was not their intention to explain the country differences found in campaign reporting; therefore, they did not work with previously explicated expectations. It perhaps also points to the difficulty in operationalizing reporting style or journalistic responses to populist discourse for content analysis studies. There may be methodological limits to what can be achieved quantitatively when developing comparative assessments of election reporting. Put another way, the aim to systematically compare campaign coverage between nations should not result in researchers providing more breadth than depth in their research designs.

The focused, quantitative-explanatory study by Schuck et al. (2013, 2016) is a sharp contrast to the comprehensive, qualitative-descriptive study by Ekström and Firmstone (2018). Schuck and colleagues focused on a single reporting feature – conflict news framing – and investigated its effects on media consumers with an unusually large-scale research design. They combined a content analysis of the coverage of the 2009 EP election in 21 countries (550,000 print and TV stories) with a two-wave panel survey (530,000 interviews). Consistent with their expectations, the authors found that conflict framing mobilized news consumers to vote. However, the mobilization effect of conflict news was stronger in countries where the European Union (EU) was evaluated more positively. The authors further found that conflict framing in election news is more prominent in proportional political systems with simultaneous national elections and in information environments with strong public service broadcasting and quality newspapers. A wealth of explanatory factors from the political and media sphere is included in the analysis, and the results indicate that context is key to understanding the nature of election news reporting and its effects on citizens cross-nationally.

Even if the last study is the second example of a fine explanatory study, it cannot obscure the fact that the majority of the recently published quantitative comparative content analyses are rather descriptive. By ‘descriptive’, I mean quantitative studies that explore and explain something by providing additional information about the topic; they try to answer What and How questions and look primarily for revealing correlative relationship between variables. ‘Explanatory’ research, by contrast, tries to explain relationships between variables; it asks Why questions and strives to establish, at least to a reasonable extent, causality. This has implications for the research design and the strategies of quantitative data analysis. I am not saying that descriptive research is generally inferior, and I am not talking about qualitative research right now. I am just saying that among the recently published quantitative comparative studies explanatory approaches are less common. Perhaps this situation will improve once researchers have made more headway in developing theoretical models that can guide the development of explanatory research designs for studying election news comparatively (see Esser and Strömbäck, 2012a, 2012b for initial proposals).

However, there is also good news: with regard to the reporting features listed in Table 1, the studies I have just discussed clearly contribute to an increase in knowledge – especially when read as a whole. Furthermore, I welcome the fact that populism (as suggested by Blumler and McQuail, 2016) is attracting more attention among comparatists. However, the overview so far has also made it clear once again that comparative content analyses of election reporting have neglected the offerings of online news media and social media platforms.

So what about digital communication?

It would be wrong to assume that internationally comparative election communication research to date has not yet taken account of the many new digital channels. On the contrary, several comparative studies have long since arrived in the fourth era of election campaign communication. However, they have only analyzed the contents of the campaign messages of candidates and parties on social media, not the digital contributions of the news media.

I would like to comment on two recent Facebook and two recent Twitter studies. They show how multinational and binational studies complement each other. Ceccobelli (2018) content-analyzed how 51 political leaders from 18 countries used 16,000 Facebook posts during national elections (compared with nonelection periods). He found that during election campaign periods, politicians were motivated not only to increase their number of posts and increase the personalization of their posts but also to reduce the share of policy discussions and reduce the level of negativity. The multinational approach finds few differences among countries and instead a general trend of homogenization with regard to campaigning on social media.

A binational study by Magin et al. (2017) looked at the same social media platform and content-analyzed how five German and six Austrian parties used their Facebook pages ‘to inform, interact with, and mobilize voters’ during national elections. They found that the parties of both countries did not make use of Facebook’s interactive and mobilizing potential (as Obama did so famously in 2008) and primarily relied on broadcasting mass-centered information. The authors attributed this conventional digital campaigning approach to regulatory framework conditions that are tighter in Germany and Austria than they are in the United States.

Graham et al. (2016) content-analyzed how 454 British candidates and 221 Dutch candidates used 55,000 Twitter messages during national elections – particularly with regard to the function and topic of the tweets. The authors found that Dutch politicians tweeted more often and in a more interactive way. The authors attributed the more conservative ‘old habits’ model of the British candidates to a more party-dominated campaign style, whereas the Dutch candidates pursued with their ‘innovator’ model a more grassroots-oriented approach.

Also looking at Twitter, a multinational study by Nulty et al. (2016) content-analyzed the 900 most popular Twitter hashtags circulating between 3200 members of the EP and their followers during the 2014 EP election. The authors found that while all politicians used hashtags associated with anti-EU and pro-EU-stances roughly equally often, the anti-EU hashtags did a better job in capturing the attention of the citizens. A high use of negative emotions in Twitter messages also seemed to correlate with anti-EU policy preferences (and not so much with left-right positioning in the national issue spectrum). The authors conclude that emotional communication on social media focused more on a politicized debate over the future of the EU than on a debate over national issues. However, they refrained from trying to explain the country differences found.

These international comparative analyses of the content of politicians’ social media messages are very commendable. Unfortunately, however, they do not do justice to the hybrid character of modern election campaign dynamics either, as they do not integrate the role of the news media. Thus, it can be stated that the available comparative content analyses are dedicated either only to the messages of the media (and here, so far, only the reporting of traditional media is taken into consideration) or only to the messages of the candidates (here, social media channels are preferentially studied due to easier data availability). The combination of both, meaning a study that is devoted to the interplay of traditional and new channels in an international comparison, is not easy to find because of the fluidity of today’s complex media ecology and the methodological challenges involved in researching this interplay.

Looking back, looking ahead

At this point, we should be mindful that my literature review was selective in that I only considered comparative studies that were published in the recent past and were accessible to me. Also, my somewhat one-sided emphasis of shortcomings should not distract us from appreciating some very positive developments. Thanks to the studies discussed here, we have learned in recent years a great deal about the extent and conditions of media bias, media tone, media personalization, media interventionism, media populism, depoliticization, game framing, negativity, and conflict in an internationally comparative perspective. We have also learned a lot about the scope conditions of digital campaigning by candidates on Twitter and Facebook. These studies help us better understand differences in political communication systems and help us recognize the context-dependency of our findings. However, it has also become obvious that we need more comparative research with more complex designs.

While acknowledging the difficulty many researchers face trying to fund cross-national studies, I have already introduced this article with an appeal for ‘more comparative research’. However, my central message is that we need ‘more comprehensive designs’, especially in content analyses, that take into account the messages of political actors and media actors in equal measure (citizens’ communicative contributions should also be gathered more comprehensively than previously). In addition to such ‘comprehensive’ studies, ‘small-format’ comparative studies are of course also justified and necessary. By this, I mean studies that deal with a theoretically original research question, a purposefully composed country sample, and a desire both to describe and explain – even if these ‘small-format’ studies can only examine selected segments from the multifaceted campaign communication.

‘Comprehensive’ comparative studies place high demands on researchers. The complexity of a comparative international analysis seems to slow down scholars with regard to also taking the hybrid multichannel nature of modern information environments into account in a content analysis. Additional desirable characteristics of high-level comparative research (such as a large number of countries, a large number of explanatory variables, an agreed-upon set of dependent variables, and causality-oriented data analysis) seem to be feasible only for a few larger and well-resourced research teams to date. However, for scientific progress it would be very gratifying if we had more comparative studies that fully meet the conditions of the fourth era of political campaign communication.

If more of our studies additionally build a bridge between the Western and non-Western world – as Kluver et al. (2019) have tried to do by analyzing the coverage of the 2016 US presidential election in Arab, Chinese, Russian, and Farsi news providers – we would do even better justice to the demands of truly international communication research.