News Coverage about Direct-Democratic Campaigns in a Period of Structural Crisis

Abstract

This article examines whether money talks in political campaign coverage. Analyzing news coverage about twenty-nine recent direct-democratic campaigns in Switzerland, it shows that votes involving expensive campaigns and populist proposals—and ideally both—correlate with high media attention. This favors especially Switzerland's largest party with the most resources, the right-wing populist Swiss People's Party (SVP). A case study of one vote on media policy confirms these patterns and shows that news coverage is also shaped by the vested (self-)interests of media organizations. The results imply that news media, affected by the crisis in journalism, fail to cover a truly wide diversity of actors and topics.

Switzerland is seen often times as a paradigmatic case with its long tradition of direct democracy, where citizens are regularly asked to cast their votes on policy issues on several levels (national, cantonal, and municipal) several times a year. When casting their vote, citizens (need to) rely on publicly available information on these issues; among these sources, the news media have become the most important source of information about politics and society and the primary channel of communication between political actors and citizens.¹ Taking into account the crucial role of the news media, the question of media performance becomes highly relevant. Some scholars claim that in Switzerland's “educated democracy,” direct-democratic campaigns are “routine and ritualized business” for the media, which leads to a substantial, differentiated, and balanced news coverage that helps citizens to make informed decisions.² Not all, however, share this rather positive view.

First, scholars point at the fact that direct-democratic campaigns, in comparison to election campaigns, are less predictable for the media, and “short-term campaign strategies and campaign tactics” become more important.³ In this sense, (some) political actors have better chances to “instrumentalize” the media (e.g., by adapting to the media logic) and increase their chances for successful campaigns at the expense of other actors. Recent examples support this view: one can think of successful campaign strategies by the right-wing populist Swiss People's Party (SVP) with its initiatives, which successfully demanded a ban on minarets (2009), an expulsion of “criminal foreigners” (2010), and a curb on immigration (2014)—all of which generated very high media attention. Tellingly, the party managed to generate “free media,” that is, media attention, with its own “paid media,” that is, political advertising. In the clearest example, the party launched a highly provocative poster depicting an aggressive-looking woman wearing a Burka and rocket-like minarets mushrooming on the Swiss flag; it generated controversy, was prohibited to be displayed in some cities, and found its way into one of Switzerland's largest political talk shows, when a whole program was devoted to the SVP campaign. Provocation worked.

Second, the positive view on media coverage about direct-democratic campaigns usually results from findings generated before the most recent period of structural crisis in journalism. This crisis can be observed on the structural level; it includes, among others, sinking advertising and subscription revenue and an erosion of the economic basis of those news outlets that still try to produce quality journalism. In contrast to that, those news outlets that focus on soft news are increasingly rewarded with increasing audience numbers and advertising revenue. Assuming that substantial, relevant journalism can be provided only in a media system with related facilitating structures (financial resources, human resources, time resources, etc.), one would expect sinking resources and growing commercial pressure on news outlets to have an impact on media performance. One result might be the media's ongoing “push to popularize politics,”⁴ which is done to satisfy (alleged) media consumer needs. In this sense, “media logic” on the basis of certain news factors further intensifies, which basically means “the selection, organization, and production of issues according to criteria of competiveness” and thus according to a “commercial logic”—with the goal being the maximization of audience (readers, viewers, and listeners) in order to generate profit.⁵

With this increasing (commercial) “media logic,”⁶ the diversity of issues and actors as presented in the news media is expected to decrease. Two types of actors in particular stand better chances to build the media agenda with “their” issues and “their” messages at the expense of other actors and issues: powerful actors that have the (financial) resources to conduct highly visible campaigns specifically geared to the media logic, and populist actors in general, who with their issues and their provocative communication style serve the media logic—and even better still, populist actors with enough resources. In this sense, the diversity of actors and issues shrinks. In the case of direct-democratic campaigns, there are indications that this recent transformation of media structures indeed affects news coverage: Successful initiatives by the populist radical-right SVP triggered large media attention also because of cost-intensive and highly provocative campaign strategies, which resonate in a media system where news media operate more and more in a commercialized media logic.⁷

Still, there is no systematic research yet as to how the current transformation of media structures affects news coverage about direct-democratic campaigns and whether these populist initiatives are mere exceptions.⁸ Against this background, our article aims to shed light on the current patterns of news coverage about direct-democratic campaigns. With data from an ongoing project, it captures news coverage about twenty-nine different votes (taking place on ten different voting days) between 2013 and 2015. It analyzes which votes trigger media attention and the possible reasons why certain votes are much more in the media spotlight than others. Focusing on the role of campaign actors, apart from more long-term structural factors such as the type of conflict, we analyze whether the intensity of a campaign in the form of political advertising expenditures can explain these differences. In short: Do political actors manage to have the media talk about “their” issue because they have more money? Is it mainly populist actors and populist campaigns that have the upper hand compared to other actors?

While this overall analysis looks at the possible impact of political actors and their campaign strategies, our article includes a case study of one of these twenty-nine votes, which serves to examine more media-centered factors. To this end, we select a case in the field of media policy (a vote on the revision of broadcasting law). This selection was done because, first, this referendum was launched by an organization with close ties to the right-wing populist party SVP and mainly conducted as a populist campaign, and second because news organizations themselves, especially in this period of structural crisis, are affected by this revision. Relying on the literature, one could expect that news outlets fail “to report accurately on issues in which their corporate owners have a vested interest.”⁹ With this case study, we try to show to what extent economic and political interests by news organizations themselves have an impact on campaign news coverage about this media policy vote. Together with the overall analysis, we will be able draw conclusions as to what chances populist actors and their campaign strategies have in a media system in a period of structural crisis.

News Media in a Period of Structural Crisis: Current Transformations of Media Structures

Switzerland's media system, like media systems in many (Western) European countries, is currently in a period of substantial transformation. First, addressing audience shifts, digitization fosters the growth of entertainment media at the expense of information media. Traditional use of news outlets in the press, radio, and television sector has been rapidly sinking in the last years; Switzerland's best and most widely used current affairs program (Echo der Zeit) on public radio, for instance, has lost around 25 percent of its audience in the last five years.¹⁰ In the case of public radio, the growing use of the online platform of the public broadcaster (incl. podcasts, etc.) does not compensate for the losses of traditional use. Also, “legacy” media such as daily subscription papers experience audience losses, while cost-free papers, which offer low-quality news, experience audience gains: 20 Minuten, launched in 1999, is now by far the largest newspaper in Switzerland, and its newspaper Internet site also attracts many more users than those of the “legacy” media. Generally, this trend is aggravated by the rise of social media: Generally, those newspaper Internet sites that are accessed relatively often via Facebook are those with audience gains, and in terms of media content, it is exactly the same news outlets that also display a high degree of “soft news” articles, which are much more likely to spread virally and be linked to Facebook, Twitter, and the like than are “hard news” articles.¹¹ In sum, this new media use is dominated by a use of episodic, moralistic–emotional soft news, and the audience of high-quality journalism is shrinking. Thus, even more so than before, a type of journalism is rewarded with audience attention that offers comparatively low quality.

Second, with this ongoing digitization of the media system and the growing convergence online, which also means increasing competition, information media are hurting financially. Not only has revenue from advertising and subscription sunk, especially for daily papers, but online does not (yet) generate substantial income. One reason is the dominance of international “tech giants” such as Google and Facebook, which siphon off most of the advertising revenue that can be generated online. Another reason is that users are not willing enough to pay for news online, and advertising revenue in the online sector that flows to Swiss news organizations is still very small. In the Swiss case, there is more evidence supporting the findings by Merja Myllylahti, who does not see paywalls, for instance, as a viable business model,¹² than the claims by Robert Picard, who sees enough opportunities for a mixed-financing model in the online sector for media organizations.¹³

These structural problems enhance the trends of media concentration, of news organizations merging their outlets and programs, of news organizations placing more and more emphasis on business units not directly connected to news (e.g., buying online ad-platforms). It also means that news outlets are more willing than before to accept money (and thus control) by political actors. In the last five to ten years, financial difficulties have opened new doors for powerful political actors, shifting a part of the Swiss media system toward “re-politicization.” In the Swiss case, this trend is mainly led by Switzerland's largest party, the right-wing populist SVP, which happens to be the party with the largest campaign budgets. One can point at the case of the daily paper Basler Zeitung, which since 2014 has been officially cofinanced by Christoph Blocher, SVP vice president and former federal councilor (federal government). A second case is the weekly paper Weltwoche, which is owned by Roger Köppel, who entered politics in 2015 as a candidate for the SVP and was elected into national Parliament with most votes out of all candidates in Switzerland's largest voting district, the canton of Zurich. A third case is the near success at Switzerland's most prestigious newspaper NZZ when the editor-in-chief of the Basler Zeitung (and thus an editor supporting the SVP) almost became editor-in-chief of the NZZ in late 2014 after the current editor-in-chief had been fired.¹⁴ This appointment was called off only after massive protests by journalistic staff of the NZZ against the board of directors.

This structural transformation described earlier arguably has an impact on the actual content that is produced. Sinking resources and growing commercial pressure on news outlets could lead to a faster news production cycle in which journalists have less time to invest in their work and put news into perspective.¹⁵ Data from large-scale content analyses show that this is indeed what happens: in most media types, media provide less context in 2014 than they used to in comparison with the period 2010 to 2013.¹⁶ Furthermore, the structural crisis could lead the media to rely more on news factors and to a “push to popularize politics” in order to satisfy (alleged) media consumer needs—a trend which was already observed looking at the time period from the mid-1960s to the mid-2000s.¹⁷ Some of the means to package politics this way are scandalization, personalization, or emotionalization, all of which are beneficial for those actors that adapt best to this new “media logic.”¹⁸ This way, powerful actors that have the resources to conduct highly visible campaigns, and generally populist actors, which with their issues and their communication style serve the media logic, find better chances to build the media agenda with “their” issues and “their” messages at the expense of other actors and issues. In this sense, the diversity of actors and issues shrinks. In addition to this, one can also argue that the recent closeness of two important papers to the SVP (Weltwoche, Basler Zeitung) and the pressure that the SVP is able to apply even on traditional papers like the NZZ increases the party's chances of finding (favorable) media attention in the overall media arena. Again, this structural change is expected to decrease the diversity of actors and issues. Before we test which direct-democratic votes actually find media attention, we give some reasons why media attention itself is important to look at and the possible factors to explain the amount of media attention.

Why Is Media Attention Important and What Drives It?

Assuming that politics is “mediated” and that citizens receive most of their information about politics from the mass media, it is clear that news outlets should devote attention to direct-democratic votes in order to help citizens take informed decisions. While no one would probably object to this statement, the problem becomes a little bit trickier given the fact that in Switzerland, typically several votes take place at the same time. This raises the question whether certain votes generate a much higher degree of attention than others (and why) or whether all votes receive more or less the same amount of news coverage. The amount of media attention, and especially the differences in media attention among votes, is important for a number of reasons. First, it affects vote choice.¹⁹ As Hanspeter Kriesi shows, “high-intensity campaigns” that are clearly visible to the public increase mobilization and thus affect the voting result (Kriesi takes the number of political ads as a yardstick to gauge the visibility and intensity of the campaign). High-intensity campaigns also lead to “argumentation-based opinions” that are more in line with actual voting behavior, whereas in low-intensity votes, people might be more likely to hold opinions on an issue but vote the opposite simply because they use other heuristics such as party preferences or general trust in government. Empirically, the government position tends to benefit from low-intensity campaigns whereas challengers tend to benefit from high-intensity campaigns, as their arguments become more widely visible and accessible.²⁰

Second, the amount of media attention is important for political actors involved in a vote, as it affects “issue ownership” in a competitive party space.²¹ This is especially true for challengers, that is, those responsible for a vote (e.g., a party that launches its own or that successfully manages to fight a government proposal with a referendum), since these are the actors that, from an institutional logic, have actively decided to intervene in the political process. Hence, challengers are highly interested in generating or maintaining “issue ownership” with the issue of the vote at stake; they are therefore interested in high media attention. Proponents of the status quo (i.e., opponents to the challengers) have ambivalent interests in media attention: On one hand, they have an interest that the vote be widely discussed in the media so their own message (against the challengers) reaches the voters, but on the other hand they have an interest in de-emphasizing the issue of the challengers and thus lowering media attention, lest the challenger is given the opportunity to create or maintain a visible profile with the issue.

The crucial question therefore is: What drives media attention and what can explain differences in media attention? What makes one vote generate the interest of the media, while another vote triggers hardly any media attention? In the following paragraphs, we specify possible factors (our explanatory variables) and discuss possible hypotheses.

Method: What Drives Media Attention and How Do We Measure This?

We rest on the assumption that various news factors have an impact on the amount of media attention;²² these news factors in turn can be related to “characteristics” of the vote and more general structural factors (e.g., the type of challenger). But news factors are not totally inherent to an event or an issue; their strength has to be perceived and evaluated by journalists, which is why not all news factors are equally important to all media, and media are not equally dependent on news factors.²³ Generally, we argue that the more media attention depends on these news factors, the more we can consider media coverage to be shaped by commercial “media logic.” This is because a strong orientation toward these news factors is at odds with the normative principles by which the media should be guided. Especially in a direct democracy, a wide variety of actors should have (more or less) equal chances to find media attention with their proposals and media coverage should not be really skewed toward those votes fulfilling as many news factors as possible. Thus, media less shaped by commercial imperatives will devote their attention more evenly, and they will not focus only on those votes that are launched by the most prominent actors, with the most provocative and most visible campaign, or the votes that are most attractive to cover simply because poll results anticipate a “close call.”

Against this background, let us now turn to possible explanatory factors (our independent variables) for media attention (our dependent variable). When discussing these explanatory factors, it is important to bear in mind that we assigned each vote a code for each factor, thus classifying the votes according to these factors. In the following paragraphs, we list the basic idea behind these measurements, and details on how we operationalize these factors exactly can be found in the  appendix.

To begin with, one could argue that high continuity as a news factor increases media attention. Continuity is high when an issue is already “familiar” and fulfills expectations on the basis of existing knowledge. Of course, knowing that journalists tend to constantly monitor which content they themselves and others have been producing, it is reasonable to assume that an issue is familiar for journalists when previous media attention on the very same issue has been extensive recently. In this sense, media take into account how intense media attention to an issue had been previously. For our analysis, we therefore examine the amount of media attention to an issue in a previous phase (six months preceding the “hot phase” of the upcoming vote) and expect that the higher media attention in a previous phase, the higher media attention in the hot phase of a vote.

Second, another obvious news factor is the importance or “political relevance” of powerful “elite” political actors and institutions.²⁴ Their decisions have a higher reach, which makes them newsworthy.²⁵ In the case of direct-democratic votes, which are not primarily about actors and institutions but primarily about issues, one could still argue that votes initiated by powerful actors are considered more relevant than votes initiated by small minorities and nonestablished actors (e.g., a small association). Thus, the higher the status of the challenger (i.e., the actor that initiates the vote), the higher the amount of media attention. In a similar line of reasoning, one could argue that the amount of political advertising could be used by the media as a yardstick of how “relevant” a vote is. In short: if political actors—proponents and opponents—invest a lot of money in a campaign, this could signal to journalists that the issue at stake is a “relevant” issue especially worth fighting about. Ceteris paribus, expensive campaigns are expected to lead to more media attention than inexpensive campaigns.

Another important news factor is conflict.²⁶ The more contested and the more controversial a vote is, the more the media believe they have reasons to cover it. Media attention is expected to intensify if the number of (powerful) political actors involved increases, which increases the scope of this problem, the stakes of the actors involved, and the uncertainty of who will actually win. Conflict as news factor is multifaceted. It is connected both to the news factor of relevance (cf. earlier text) and also to the news factor of unexpectedness, which is more in line with a factor that is salient especially in media following a commercialized media logic. Also relating to a more commercialized media logic, conflict as a news factor often stands in tandem with news factors such as damage or aggression.²⁷ In this sense, conflicts are newsworthy because commercialized media can use them to portray politics as an exciting drama.

To operationalize these various facets of conflict, we look at the actors and the content behind the conflicts as well as the degree of uncertainty. First, as regards actors, we look at the “coalitional configuration,”²⁸ meaning which actors actually fight each other on a vote. We hypothesize that a high consensus among the important actors is expected to decrease media attention, whereas a low consensus and a high conflict among important actors, especially between the right-wing populist SVP on the one hand and other large parties on the other hand, to increase media attention. In addition to conflict factors addressing conflict among parties, the degree of party-internal division, thus the conflict within a party, is another important factor. We expect that the more parties are internally divided on a vote, the higher media attention will be.

Looking more at the content of conflicts and the proposals at stake, we first examine the conflict type, distinguishing among the basic structural conflicts, that is, conflicts in the “cultural” dimension (comprising issues such as migration, law and order, minority rights, etc.), in the “economic” dimension (comprising issues such as budget, health care, tariffs, etc.), and those (few) conflicts that cannot be automatically assigned to either category (e.g., environment, traffic).²⁹ We expect conflicts in the “cultural” dimension to generate more media attention than conflicts in the “economic” dimension. This is because conflicts in the “cultural” dimension have gained in importance in several Western European countries. In Switzerland, as in the Netherlands, or in Great Britain, but not in Germany and not in France, the cultural cleavage has become even the dominant one. Along with the growing importance of this conflict axis, a party system has been taking shape where right-wing populist parties acquire a distinct profile in opposition to mainstream parties from the left and the right.³⁰

Focusing on a second indicator of the content side of conflicts, we examine the degree of populism reflected in a proposal because populism with its antagonistic relationship of “the people” and “the elite” inherently expresses a fundamental conflict and thus is expected to enhance media attention. Populism, as defined by the literature, considers the (homogeneous) “people” to be good and pure, while at the same time the “elite” is denigrated and blamed for affecting the people negatively.³¹ Populism also includes a scandalization of democratic procedures of liberal democracy with its checks and balances: typically, popular sovereignty expressed in popular votes is considered the main (or even the only) legitimate channel of democratic politics.³² In terms of rhetoric, the style of populism usually entails the use of “highly emotional, slogan-based, tabloidstyle language,”³³ which aims to “tap feelings of ressentiment and exploit them politically.”³⁴ All this makes populist proposals highly newsworthy and attractive, especially to news media following a commercialized media logic. Thus, we expect populist proposals to trigger more media attention than popular votes not or only partially shaped by a populist proposal.

Finally, addressing conflict in relation to unexpectedness, we argue that the (anticipated) closeness of the vote is reflected in media attention. We look at results from opinion polls published around six weeks before the vote takes place and posit that close results overall correlate with high media attention. We also look at the actual official poll results, and again we posit that the closer a vote has turned out to be, the higher the media attention during the “hot phase.”

After discussing these possible explanatory factors, we now turn to media attention as our main dependent variable. To examine which factors drive media attention to direct-democratic votes, we look at the latest twenty-nine votes in Switzerland on the national level (March 2013–June 2015), which took place on ten different days. We captured all news articles dealing with these votes in a period of eleven weeks each, thus capturing the “hot phase” of a campaign. We started twelve weeks before voting day and captured the following eleven weeks. (In the last week before voting day, there is usually very little media coverage, also because a high number of citizens votes by mail, which they send out some weeks before voting day.) We analyzed media coverage in eight newspapers. To give justice to the segmented press market in Switzerland, we incorporated different press types. This is why we included three daily subscription papers, among which two are considered Switzerland's “best” or high-quality newspapers and one Switzerland's largest “mid-market” paper, the two largest daily tabloids, the largest free commuter paper (which is also the largest newspaper in Switzerland in terms of circulation), and two Sunday papers, among which one is a “high-quality” Sunday newspaper and the other is the largest one.

We coded news articles with a whole set of variables such as tone of coverage, and so on, which we use for our case study (cf. the following sections) but for the purposes of our overall comparative analysis, we focus on the intensity of news coverage. As a proxy for how much media attention a vote generates, we use the number of articles. We cross-checked the number of articles for each vote with other possible indicators of media attention such as prominence (front-page articles vs. others) and length (number of characters). The distribution of media attention among the twenty-nine votes is basically the same when we look at the number of front-page articles (11 percent of the total sample) and at the number of articles not on the front page (Pearson's r = 0.97). Similarly, the number of articles for each vote highly correlates with the number of characters, a proxy for article length (Pearson's r = 0.99). Thus, we are safe to assume that the number of articles is a simple but valid indicator of media attention.

Results: Media Coverage about Direct-Democratic Votes in a Comparative Perspective

A first analysis of media attention (cf. Figure 1) shows that media attention to direct-democratic votes is distributed unevenly across the votes. The initiative against “mass immigration” by the right-wing populist SVP generates around twelve times more attention than the referendum initiated by left-wing parties against extended opening hours for shops in gas stations or eleven times more attention than the student initiative for increasing stipends. This already is an important finding, as the actively promoted initiative by the right-wing populist SVP manages to trigger most media attention by far.

To explain the differences in media attention more systematically, we used the factors listed earlier (more details on the operationalization can be found in the  appendix) and used them to build regression models.³⁵ We started with “previous media attention” in a first model because the news factor “continuity” is expected to play an important role and because with this factor, we take into account an autocorrelation effect. This first model shows that previous media attention has a clear effect on media attention in the “hot phase” (B = 0.716, S.E. = 0.161; p = 0.000) and shows relatively strong explanatory power (adjusted R² = 0.401). After having discussed this first simple model, we then began again with all ten variables (cf.  Appendix) in a full model (adjusted R² = 0.760) and used a backward model selection procedure. Step-wise, we excluded those variables whose effect was not significant. This results in the following model, which is displayed in Figure 2.

In the final model, apart from the effect of previous media attention, another three out of ten variables show an effect (cf. Figure 2). Compared to the first model described earlier (consisting only of “previous media attention”), the final model has a significantly higher explanatory power (adjusted R² = 0.775 vs. 0.401). Above all, political advertising expenditures go hand in hand with media attention. Second, conflicts in the cultural dimension receive more media attention than do conflicts not belonging to any of the two conflict dimensions (e.g., traffic) and socio-economic conflicts. Third, proposals with a higher level of populism are those with higher media attention.

Before we discuss these relevant factors in more detail and try to find causal mechanisms, we show which factors have turned out not to be significant and which have therefore been excluded from the final model.

The status of the challengers according to their formal status in the political system is not significant. This means that neither powerful actors per se nor peripheral challengers per se (the “Davids” against the “Goliaths”) can generate especially high or especially low media attention with “their” vote.³⁶ To illustrate this, we can point at the two initiatives with largest media attention: The initiative against “mass immigration” is launched by the largest party (high status), but the initiative against “excessive executive pay” is launched by a single member of Parliament who does not even belong to a party (lower status). Low media attention is awarded not only to an initiative launched by a small student organization (low status) (“increasing stipends”), confirming the hypothesis, but also to an initiative launched by an established government party, the Christian Democratic People's Party (CVP) (“family initiative”), against the hypothesis.

Regarding conflict factors, which we operationalized with seven variables, most variables do not explain much. It does not matter how close or contested the vote was in Parliament according to voting behavior by members of Parliament (the lower chamber usually votes on whether it officially supports a direct-democratic proposal or not). Furthermore, the “coalitional configuration” as such and the according “official” voting recommendation by parties does not play a role. Thus, it is not important whether and how the large parties in government are divided on a vote, for instance whether the vote mirrors a “center-left” conflict (Social Democratic Party [SP], Radical Democratic Party [FDP], and CVP against SVP), a totally divided coalition (e.g., SP and CVP vs. FDP and SVP), a “center-right” conflict (CVP, FDP, and SVP against SP) or a “grand coalition” where all these four parties are either against or for a proposal. Not even the amount of party-internal division has an effect on media attention.

Similarly, conflict factors relating even more to unexpectedness, uncertainty, and the “horse race” aspect of journalism fail to explain differences in media attention. Here, one could possibly expect that, if the media anticipate a close race, this would increase the likelihood of reporting. However, the actual closeness of the vote (measured as actual vote results) alone does not play a role, and neither does the closeness according to opinion polls, which are published six weeks before the vote.

Thus, our final model suggests that continuity is important, that is, high previous media attention leads to high media attention in the “hot phase,” and so is “relevance” according to how much money campaign actors invest (whether pro or con). Conflict, which in several respects does not show an effect, is still important in two respects: first, populist proposals, that is, those proposals that express rather fundamental conflicts of “either-or” instead of conflicts of “more-or-less,”³⁷ stand better chances of finding media attention. Second, direct-democratic campaigns dealing with a “cultural” conflict (e.g., migration, law and order, media policy) are more readily picked up by the media than other conflicts.

On the basis of which factors empirically cluster typically and which factors correlate with each other, we see, for instance, that populist proposals (six votes) are rather “cultural” conflicts than economic conflicts (four vs. two votes), and actors behind these votes are either clear outsiders (three votes), the more radical Youth Section of the Social Democrats (one vote), or the established right-wing populist party SVP (two votes). For political actors, this implies that launching populist proposals is a successful strategy if the main goal of these actors is to trigger media attention.

In our statistical model comprising twenty-nine cases, it is not possible to finally solve the problem of reciprocity and test further causal mechanisms. As regards the effect of previous media attention, we see that it not only correlates with media attention in the hot phase, but also with advertising expenditures (Pearson's r = .462, p = 0.012). In terms of causality, there is some support for an interpretation in which previous media attention influences political advertising and not the other way round. After all, advertising campaigns have to be planned beforehand, and if media attention in the “hot phase” within eleven weeks (unexpectedly) develops in a way unfavorable to a campaign actor, this actor cannot suddenly increase the campaign budget substantially.

As we are left with the effect of previous media attention, we would have to wonder which cues the media got in that earlier phase to gauge the relevance of an issue. Again, given the fact that most “structural factors” such as status of challengers or the “coalitional configuration” do not explain much, while the level of populism, cultural conflicts, and advertising expenditures do, it is more reasonable to assume that both the media and campaign actors together in a sort of an ongoing “feedback loop” constitute the salience and thus the “relevance” of an issue both over a longer period of time and then again during the “hot phase” of the campaign.

Given this likely ongoing “coproduction” between media and actors with large campaign resources, this finding becomes even more relevant considering that advertising expenditures are highly skewed in three ways. First, in each vote, usually one camp invests much more in political ads than the other camp. In the case of the twenty-nine votes, one camp on average spends 83 percent of all money invested in a vote, while the other camp spends only 17 percent. Second, advertising expenditures are highly skewed toward political (and economic) actors from the right. Third, focusing only on political parties, the bulk of advertising expenditures is spent by the SVP. We can substantiate these claims by giving only estimates, since the data does not allow us to assign all ads to a clearly identifiable actor. In many cases, there is only data indicating which ad hoc coalition invests money in an ad (e.g., Committee for the Revision of the Broadcasting Law), making it impossible to say how much money a certain party within an ad hoc coalition invests. In the data, 56 percent of advertising expenditures come from political parties from the right and economic organizations (we know this also because this sum comes from campaigns directed at proposals launched by left-wing actors), while 6 percent can be clearly assigned to actors from the left (e.g., Social Democratic Party, unions). Thirty-eight percent are unclear (e.g., opposition from both left-wing and right-wing parties, with the exact share being unknown). In those cases where political ads clearly can be attributed to a specific party, again we see that the SVP spends most money: While it spends around 3 million Swiss Francs overall, the Liberal Democrats spend around 2.5 million, the Christian Democrats around 1 million, the Social Democrats around 250,000, and the Green Party around 125,000 Swiss Francs.

These results are perfectly in line with a study by Michael Hermann, who analyzed advertising expenditures between 2005 and 2011, including direct-democratic campaign periods, election periods, and periods outside of these concrete campaign periods. Hermann also found that left-wing actors are heavily outspent by political actors from the right, especially by the SVP and the large economic associations such as Economiesuisse.³⁸ If media attention depends partially on the amount of campaign resources, this in turn means that the SVP and economic organizations have better chances to increase visibility to a vote they are interested in (whether in promoting a proposal or fighting it) than other actors. Thus, political communication in Switzerland faces a double challenge: Not only is the financing of political campaigns largely nontransparent—Switzerland is regularly criticized by the Council of Europe's Group of States Against Corruption for this—but available resources are distributed unevenly among political actors.

In terms of party politics, the findings of the model are highly relevant as conflicts in the cultural dimension are those where empirically the right-wing populist party SVP has acquired a clear profile and positioned itself against most other parties, further legitimizing its use of antiestablishment rhetoric against all the other “mainstream” parties.³⁹ In this sense, whether the SVP itself leads cost-intensive and (more or less) populist campaigns or other actors lead intensive countercampaigns against the SVP, it means that the issues that the SVP has come to “own”⁴⁰ generate especially high visibility.

All in all, then, the data show that the amount of media attention to a substantial degree centers around a populist party, the SVP. In this sense, one must conclude that news coverage about direct-democratic campaigns shows a reduced level of diversity concerning issues and actors. In their coverage of direct-democratic votes, the media do not break this cycle and do not manage to offer a more balanced reporting about these various issues that are up for the vote.

Case Study: “Vote on the Revision of Broadcasting Law”

In the overall analysis described earlier, we learned two things. Political advertising correlates with media attention and populist proposals and votes in the cultural conflict dimension increase media attention, especially when Switzerland's largest party, the right-wing populist SVP is highly involved, either because it is internally divided (which was the case in two votes) or because it invests even more money on a campaign than it does for other votes. Our case study, the vote on the revision of broadcasting law, fits exactly this pattern. It generates media attention above average, and political advertising expenditures are also above average, albeit slightly. The content of the vote takes the form of a cultural conflict because it deals with media policy, where for decades the right-wing populist party SVP has been positioning itself as a vocal critic of the (allegedly) left-wing public broadcaster, whereas the other three large mainstream (center-right) parties have been clear supporters of public broadcasting. Also, the referendum against this law was initiated by the Swiss Federation of Small and Medium Enterprises (Schweizerischer Gewerbeverband), a powerful economic interest organization with close ties to the SVP (the president of this organization is a member of Parliament of the SVP). Overall, this makes it a highly representative case to study in more details the patterns of news coverage. At the same time, this is a special case where the media are directly affected by the law. This raises the question whether media's own interests are additional factors that explain not only media attention to this vote but the overall tone and style of news coverage.

A few words are necessary to describe the content of the vote. In June 2015, Swiss people voted on a revision of Broadcasting Law. In Switzerland, not only public broadcasting but also private broadcasting is regulated. The regulation concerns not only what is expected from broadcasters in terms of content to be produced but also the financing (who receives and who needs to pay how much?). As concerns the content and the related public service functions broadcasters should fulfill, the revision practically did not change anything. As concerns who receives license fees, the revision meant that the public broadcaster SRG (Swiss Broadcasting Corporation) would still get by far the largest share (more than 90 percent), and private broadcasters, obliged to offer regional news, would receive a slightly larger share of the whole revenue than before (in sum between 4 and 6 percent instead of 3 to 5 percent). Compared to the financing of public broadcasting, this overall growth could be considered non-substantial; however, on the level of individual broadcasters, given that they usually have a relatively small budget, this growth would mean around 10 or 20 percent of a broadcaster's revenue income, which then does not seem non-substantial. Finally, as concerns who needs to pay for the license fee, the revision would bring substantial changes. The overall sum of license fees should stay the same, but the division of who needs to pay would change. In the old version, every household and every company was required to pay for a license fee once a year only provided they had access to a device with which content by public or private broadcasters could be received and used. If a household (or a company) did not use this content—or at least claimed they did not use it—they could opt out and would not have to pay. In the revised law, this opting-out principle was given up; all households and all companies (but, in the revised law, only those with a minimal size, i.e., a minimal business volume of 500,000 CHF) would have to pay for a license fee. In sum, households overall would benefit financially, since the new fee was announced to be around 400 CHF instead of around 460 CHF a year. Also, while companies overall would have to share a higher burden than before, the revised law would shift the burden from smaller companies to bigger companies because of a progressive fee system (plus an exemption from the license fee if the business volume was lower than 500,000 CHF).

As for the media covering this vote, it is clear that this law affects them directly. This is true not only for news organizations that own regional radio and TV stations and which would benefit from the new law. It is also true for news organizations in general in the current crisis of journalism, where private news organizations face increasing commercial pressures due to dwindling audience and advertising revenues. With the audience shifts toward online news media (and social media such as Facebook), a growing convergence and similarly a growing competition emerges in the online sector. News organizations, which formally had worked in different media sectors, now compete in the online sector. This explains, for instance, part of the conflict between the public broadcaster SRG and private media organizations—because all actors attempt to find users offering media content online.

No matter whether private media organizations are threatened more by the public broadcasters' activities online or by large international corporations such as Google or Facebook that have the upper hand in generating advertising revenue online, from a normative perspective, media in their coverage should foster the public good and offer a fair and transparent coverage instead of promoting particular interests.⁴¹ Commercial imperatives, it is argued, increasingly reduce this public good orientation.⁴² In an important strand of the literature, scholars analyze to what extent media foster the public good by critically reporting on how other media are covering important issues or by critically reporting on issues that directly affect them (media policy issues).⁴³ Many scholars observe a decline of “media journalism” or “media critique,” even though the media increasingly tend to observe each other constantly to gauge which topics are relevant and with which stories an outlet could have an edge over competitors.⁴⁴

In this context, scholars observe that a typical news outlet tends not to cover itself and its related media organization critically enough (transparency and self-observation).⁴⁵ Stephan Russ-Mohl, for instance, argues that basic journalistic rules are threatened when interests of a media organization come into play.⁴⁶ This is because media tend to cover its “own” media organization more positively than others, either neglecting competitors or portraying them more critically.⁴⁷ In the still rare empirical studies on how media actually cover themselves and other media, Pointner, for instance, looked at media coverage about media organizations over more than a decade (1992–2006) and concluded that this type of media coverage is increasingly shaped by commercial imperatives.⁴⁸ Press outlets report more critically about broadcasting than on (other) press outlets, and media coverage about outlets belonging to the same media organization is less critical still. Ralph Weiss also observed that press outlets in Germany cover (public) broadcasting quite critically.⁴⁹ Similarly, news outlets shy away from media policy issues that could be harmful to their owners.⁵⁰ In a case study on a takeover in the media sector, Beatrice Kemner showed that the media coverage is shaped by interests of the related media organizations.⁵¹ Not only the intensity, but also the tone of coverage is shaped by the interest of an outlet and its related media organization. In another study, Catie Snow Bailard provided similar insights by analyzing news coverage about the controversial Citizens United ruling by the Supreme Court, which loosened restrictions on campaign spending and led to a remarked increase in expenditures on television advertising. She showed that “corporations with more television stations, which equates to more airtime to sell to political advertisers, were significantly more likely to mention positive consequences associated with the Citizens United ruling compared to companies with fewer or no television stations, and they were also somewhat less likely to mention negative consequences relative to their counterparts who stood to gain less financially from the ruling.”⁵²

Apart from economic self-interests on the level of news organizations, factors on the individual level could also play a role in this process. In a commercialized and highly concentrated media system, journalists would rather hold back with criticism of their own organization, lest they get sanctioned, possibly even losing their job and having few alternative organizations to turn to in order to find another job.⁵³ In a recent study by Manuel Puppis and colleagues, Swiss journalists surveyed indicated it was clear to them that they would (have to) report positively about their own media organization, while critical reports about their own organization were not possible.⁵⁴

To analyze whether a news organization's own interests shape news coverage in the case when news organizations are themselves directly affected, we classify media organizations along their supposed interests in the revision of the broadcasting law, including the discussion about the public broadcaster SRG. We then set this classification in relation to the actual tone of coverage we can identify in several news outlets belonging to these media organizations. For this case study, we extended the media sample to twenty-six news outlets and slightly extended the time period (cf.  Appendix).

On a five-point scale, we tried to position news organizations (and their related news outlets) and their expected stance toward the vote (and thus, the public broadcaster SRG), from “very negative” to “slightly negative,” “balanced,” “slightly positive,” and “very positive.” For this scale, we consider political interests—the cases of Weltwoche and Basler Zeitung with close ties to the SVP, a vehement critic of public broadcasting—and economic interests, which we operationalize according to two factors. First, we ask whether a media organization would benefit from the revised law financially (by owning private broadcasting stations), and second, we ask to what extent a media organization considers the public broadcaster to be a (threatening) competitor, especially regarding SRG's online activities that allegedly undermine the possibility of private media organizations to generate enough revenue with newspaper Internet sites—important as readership numbers of printed versions dwindle and advertising revenue in the press sector is shrinking.⁵⁵ Brief explanations for this categorization of the media organizations are given in the  appendix. For instance, we expect a negative tone of coverage, albeit only a “slightly negative” coverage, in outlets of Switzerland's private media organization NZZ Mediengruppe. On one hand, it shares economic interests similar to Tamedia and Ringier in the online sector (SRG as a competitor) but on the other hand, it owns regional radio and TV stations that would benefit financially from the law. However, those TV stations are not very important in the organization's business strategy. Thus, we would expect a criticism of the SRG but a promotion of the law itself—in sum a “slightly negative” coverage.

The tone of coverage—our dependent variable—was measured on the article level. We coded either a “positive,” “negative,” or “controversial” tone on the basis what the main message of the article was in relation to the revised law and/or in relation to the public broadcaster. For each news outlet, we calculated a “tone of coverage index,” where we subtracted the number of articles with “negative” tone from the number of articles with “positive” tone, then divided this number by the number of all articles by a news outlet (thus, also including the articles with “controversial” tone) and finally multiplied this value by one hundred. In Table 1, we display the average values of the “tone of coverage index” in those cases where we have more than one news outlet in one group (e.g., the group of news outlets and related news organizations that are expected to show a “slightly negative” tone of coverage).

Table 1 shows that political and economic interests indeed correlate with the tone of coverage. News outlets with a higher interest in a rejection of the proposal show a substantially more negative coverage than those news outlets with a higher interest in the proposal being accepted (p = 0.0008). There is some variance within media organizations, and it seems that in the case of NZZ Mediengruppe, AZ Medien, and Tamedia, the “flagship” paper tends to be more balanced compared with other news outlets (NZZ, Aargauer Zeitung, and Tages-Anzeiger) (not displayed). In this sense, quality orientation and an orientation toward the public good in a news outlet could be a buffer toward interests on the level of a media organization. All in all, however, the results are fairly clear, and differences are significant. Also, if we group together those outlets with expected negative coverage (“very negative” and “slightly negative”) and those with balanced and (slightly) positive coverage, again we see significant differences (t-test: p = 0.033). The tone of coverage is significantly more positive in the “positive” news outlets (avg. 7.1) than in “negative” news outlets (avg. –18.2).⁵⁶ This correlation between own interests and the tone of coverage is quite substantial, with Pearson's r being 0.394 (p = 0.046) and Spearman's ρ being 0.390 (p = 0.049).

In more qualitative terms, we looked at arguments exchanged in the media and we saw a clear pattern. The overall focus of media coverage was not on the core revision itself but on the role of the public broadcaster. Typically, this discussion about public broadcasting took the form of a problematization of the SRG. Statements criticizing the budget of the SRG, the “excessive” pay awarded to the top management, or the “dominance” of the SRG compared to private media were more common than statements pointing at the positive role the SRG plays in guaranteeing a public service. Also, we have to bear in mind the vote overall generated media attention that is above average (cf. Figure 1). Given this finding, one can conclude already that opponents of the law, including the media, managed to shift the focus of the vote away from a discussion on who needs to pay how much to a broad discussion on whether the public broadcaster SRG should actually receive (so) much.

Whenever the new model itself came into focus, it was rather the proponents' arguments that find media attention than the opponents' arguments. This is in line with the discursive strategies of the pro-camp, which tried to stress the practicability of the new financing system and the advantages for households. For instance, the proponents advanced the argument that with the new model, one household would have to pay around 60 CHF less a year than now; this was broadly reflected in the media. The reason why the pro-camp all in all still used more arguments focusing on the SRG than on the model itself is simply because the pro-camp had to react to the public service debate that the con-camp had successfully launched.

One case in point is the reaction to the campaign of the challengers, who, one has to remember, invested more than twice as much than the proponents of the law (1.3 million vs. around 500,000 Swiss Francs). In a provocative leaflet sent to all Swiss households, the right-wing Swiss Federation of Small and Medium Enterprises (responsible for the referendum) depicted the director of the SRG in a cartoon as a cross-eyed, greedy figure that fills his pockets with money. In the text, the director is explicitly accused of “stealing from the people.” In a gruesome-looking photograph, it showed bloody and distorted fingers caught in a mouse trap, which was meant to illustrate that the alleged reduced license fee for households is a trap, eventually leading to higher “coercive fees” or “taxes” for households because the SRG would use a yes vote as a legitimization to extend its programming and then demand even higher license fees. Especially when the media learned that this campaign was designed by the same advertising agency that for years has been supporting the campaigns of the right-wing populist SVP, criticism amounted quickly. For example, even the president of Switzerland explicitly addressed the campaign and condemned it for its style as “disquieting, even dangerous” for Switzerland's political culture.⁵⁷ On one hand, scandalizing this campaign supported the pro-side but on the other hand, as with every truly populist campaign, it generated exactly this kind of media attention that populist campaigners actually want. Not only did it increase the salience of the issue as such, but it also allowed populists to play the successful game of being “unfairly” attacked by the (allegedly left-wing) media, whose motives were only to divert attention away from the “real” issue. This provocative leaflet then is a perfect example of how “paid media” successfully generate “free media” and how provocative, populist campaign strategies manage to find attention in a commercialized media system.⁵⁸

To conclude, our analysis suggests how successfully the Swiss Federation of Small and Medium Enterprises in tandem with the right-wing populist party and with the help of private media and their non-explicit political and economic self-interests manage to shape news coverage.

Conclusion

This article started from the assumption that the current structural crisis of the news media in Switzerland has detrimental effects on news coverage. Taking as an example news coverage about twenty-nine recent direct-democratic campaigns, usually praised as a “routine business” for news media, it showed that media attention to these various votes is not distributed equally at all. Of course, no one would expect the news media to devote their attention perfectly equally on issues, but the analysis suggests that the diversity of issues (here: votes) is restricted. With regression analyses, it showed that media attention does not depend on the actual status of the challengers or the level of conflict according to how many parties are divided on a vote. Media attention increases when the media have already reported extensively in a previous phase, when the proposal is at least partially populist in character, when it deals with a “cultural” conflict instead of an “economic” conflict and when political actors increase their advertising expenditures. This is an indication that media and resourceful campaign actors co-orient each other in gauging how “important” a vote is—a finding that is supported in the literature on the agenda-building capacities of powerful political actors in Switzerland.⁵⁹ Since advertising expenditures are very unevenly distributed in each vote and generally among political parties, our findings imply that political (and economic) actors with large campaign resources manage to increase media attention. Also, actors launching populist proposals find better chances for “their” issue to be covered by the media (partially also because other actors react intensively to these proposals). Empirically, only the right-wing populist party SVP fulfills both criteria: it leads populist campaigns and it uses the largest campaign resources. Overall, the importance of campaign resources and the level of populism does not mean yet that actors with these strategies and resources are necessarily successful in actually winning the vote, but one important condition, that is, a high salience of a vote, is fulfilled already.

Our article also showed that votes in the field of “identity politics” addressing the “integration–demarcation” cleavage trigger large media attention, and whenever the right-wing populist SVP is highly involved—either because of party-internal division (conflict about whether to support a nonestablished actor that “stole” the issue of immigration by launching another initiative) or because it invests substantially more in a vote than it does normally—media attention increases. Together, this means that media coverage to a large extent centers around the SVP, stressing the importance the media (and other actors) ascribe to this party and their “own” issues. We believe that the current structural crisis facilitates these trends, as these types of conflict and the provocative campaign strategies by the populist radical right are newsworthy because they fulfill the “media logic.”

In a case study focusing on a vote on the revision of broadcasting law, we showed some of the same mechanisms involved, and we added another factor that explains the amount of media attention to the vote and the arguments exchanged: political and economic self-interests by private media organizations. In addition to being receptive to an expensive and highly provocative campaign conducted by an economic organization (and a campaign organization) with close ties to the SVP, also because these types of campaign are generally newsworthy, the media seem to have followed their own political and economic self-interests in this vote. More so than before, the public broadcaster SRG was attacked by the (private) media and stylized as a problem that threatens private media in this current structural crisis; the law was attacked on the grounds that it would illegitimately further privilege the private broadcaster in an era of media convergence. In contrast to that, media organizations that would actually benefit from the law, as they own private regional broadcasting stations that would also receive some more money, reported substantially more positively. In general, own interests were hardly made transparent in news coverage. In this case, news coverage in the form of general “media logic” was supplemented with an interest-driven coverage, which overall did not meet the criteria of a fair and balanced debate.

Of course, this study has some limitations that need to be addressed. One point concerns the dependent variable of our overall comparative analysis, media attention. In our article, we have given several reasons why we think it is necessary to focus primarily on media attention. But, naturally, the characteristics of media coverage also entail other important elements such as the tone of coverage (as we examined in the case study), actors quoted or journalistic “routine frames” such as personalization, scandalization, and so on. These elements, we believe, cannot be captured by quantitative analysis alone, also because they receive their meaning mainly through the actual dynamics of a campaign. Following this, we plan for more qualitative case studies for each of the votes, which in turn makes it difficult to obtain a large number of votes. Another point is the number of votes for which we have both data on political advertising and on media content; while twenty-nine votes are too many to conduct in-depth qualitative analyses for all the votes, this number at the same time allows only very basic statistical tests. Thus, a further step of the analysis could be adding another method that fits the size of n cases (e.g., qualitative comparative analysis). Finally, in our case study, we have tried to operationalize self-interests of media organizations based on document analyses (e.g., business reports). Ideally, however, expert interviews of media managers would be an important addition in order to validate this operationalization.

Still, we believe that our study is a first step toward a systematic, comparative analysis of direct-democratic votes that comes at a critical time for journalism in Switzerland. All in all, our article shows that not all is well in the state of Switzerland when it comes to how one of its cornerstones, direct democracy, is conveyed in the mass media. Against this background, it seems more necessary than ever to find ways to solve the current structural crisis of journalism and to find ways to deal with the glaring inequalities in the financial resources that political actors have at their disposal to lead political campaigns.

APPENDIX

This  appendix gives more details regarding the media sample and operationalization of the independent and dependent variables used in this article.

Overall analysis

Media sample:

Eleven weeks during “hot phase” (twelve weeks before voting day up until one week before voting day)

Eight news outlets: Subscription papers (NZZ, Tages-Anzeiger, Le Temps [F]), Commuter paper (20 Minuten), Tabloids (Blick, Le Matin [F]), Sunday papers (NZZ am Sonntag, SonntagsZeitung) (an “F” indicates an outlet from French-speaking Switzerland)

Case study

Media sample:

1.3.2015 to 14.6.2015 (voting took place on 14.6.2015)

Twenty-six news outlets and programs from eight media organizations (an “F” indicates an outlet from French-speaking Switzerland)

Independent variable: political and economic interests and their assumed effect on news coverage (5-point scale)

Footnotes

BIBLIOGRAPHY