THE DELIBERATIVE QUALITY OF REFERENDUM COVERAGE IN DIRECT DEMOCRACY
FINDINGS FROM A LONGITUDINAL ANALYSIS
OF SWISS MEDIA ANDRÉ DONK
FRANK
MARCINKOWSKI
Abstract
The article presents a systematic and standardised content analysis of 4,559 newspaper articles; it covers nine popular votes in Switzerland between 1983 and 2004 and measures the deliberativeness of the mediated public debate. In the last decade, a growing number of studies employ a deliberative framework in analysing mass media contents. However, these studies followed a sceptical perspective and found evidence that mediated deliberation inevitably falls short of the demanding criteria provided by normative theory. Nevertheless, the article demonstrates that there are examples of deliberative journalism in Swiss direct democratic campaigns. We argue that a political system of a mature direct democracy, such as the Swiss democracy is, together with a journalistic culture which is "educated" by initiative and referendum, might provide an appropriate environment for mediated-public deliberation.
Frank Marcinkowski is Professor of Communication in the Institute for Communication Research, University of Münster; e-mail: frank.marcinkowski@ uni-muenster.de.
André Donk is Advisor to the Scientific Director in HIS - Institute for Research on Higher Education, Hannover; e-mail: donk@his.de.
ro c^
si ü
o
rN
i—I
o rN
o~>
£
£ 3
a
0 £
1 *
(R 0 C
! m
Introduction
The public debate over affairs which are of general interest and relevance to the people is a key element of democracy. Hence, public discussions reveal numerous effects on the political attitudes, knowledge and interest of the people, on the rationality of political outcomes, and on the legitimacy of collectively binding decisions. The quality of democracy therefore depends on the quality of public deliberation, which in modern societies, mostly takes place in the mass media (Page 1996, Gastil 2008). In this article, we reconsider the well-established research question: Under which conditions is mediated deliberation possible? Following this line of reasoning, a growing number of studies employ a deliberative framework in analysing mass media content (Ferree et al. 2002; Bennett et al. 2004; Lunt and Stenner 2005; Mutz 2007; Maia 2009). A common finding of this research is that certain intrinsic limitations within the political or media system prevent ideal deliberation conditions from prevailing. Since commercial mainstream media have to refer to news values such as personalisation, negativity, and conflict, they are not likely to provide a top-quality information environment which would enable citizens to decide over complex policy issues on the basis of arguments. Thus, mediated deliberation inevitably falls short of the demanding criteria provided by normative theory for structural reasons (Peters et al. 2008; Habermas 2008a, 158). By contrast, Wessler (2008b) has argued that different types of democracy, together with different types of media systems, should differ significantly in the forms of mediated deliberation they tend to offer. Instead of neglecting the potential offered by mediated deliberation in general, we should therefore investigate the very specific preconditions of a mediated public debate which meets the needs of deliberativeness. Following Wessler's argument, we assume that the political system of a mature direct democracy, together with a journalistic culture, which is "educated" by initiative and referendum, might provide the optimal environment for mediated-public deliberation. Such an environment is supposed to be found in ^ Switzerland. Thus, a case study analysing the deliberative quality of Swiss media coverage in the context of direct democratic votes is presented.
In order to answer our central research question, we first briefly describe the concept of deliberative democracy and relate it to comparative research on media systems, which reveals that the deliberative quality of public communication depends partly on certain elements and constituents of the political system. We use Q these insights to formulate our hypotheses, which are followed by indicators for ^ deliberation as a feature of public communication, and presentation of the results of content analysis.
Conditions of Mediated Public Deliberation
The concept of deliberation refers to a certain type of decision-making process, which is based on free, equal and rational debate on political issues. Thus, deliberation aims for a consensus which is derived from an agreement on the stronger or more valid argument (Cohen 1989; Habermas 1992; Bohman 1996; Schudson 1997; Elster 1998). Basically, as a specific feature of public communication, deliberation is often related to Habermas' notion of an ideal speech situation. In this sense, "deliberation denotes a specific quality of political communication that centers
A
C >
on argumentative exchange in a climate of mutual respect and civility" (Wessler 2008a, 1199). Of course, there are also limitations of the deliberation approach, LO as it conceptualises an ideal situation which cannot be met perfectly in reality. ^ Furthermore, some critics argue that deliberation cannot be adopted in large political entities, because time and resources giving every citizen the chance to participate are lacking (Schmidt 2005, 281). Therefore, in modern societies not all citizens can participate as speakers in public discussions, but everyone can follow these discussions and gather information in the media. Hence, Page (1996) reacts to this criticism and proposes the concept of mediated-public deliberation. He puts forward the argument that the deliberation of political issues is determined by the professional communicators and society's elites who "are responsible for conducting the discussion on major political issues through their contributions" (Page 1996, 4) and use the media as means of transmitting these discussions to the people. As a consequence, serving as arenas of public deliberation, the media can only be as deliberative as the elites themselves. Thus, we argue that the political system is one crucial independent variable that helps to explain the deliberative quality of public deliberation. However, the theoretical literature is largely sceptical as to whether deliberation can be achieved ideally in the context of modern mass media. After defining certain requirements for deliberation within the media, Gastil (2008) arrives at the conclusion that the US media system does not perform very well with respect to these deliberation measures. Many scholars have expressed their scepticism concerning media ability to serve public deliberation. First, only the media, publishers and broadcasters have the resources to reach a mass audience. In this sense, they still own a gate keeping-position. Mass media communication is highly asymmetrical. Second, mass media production does not match the requirements of a discourse, as it (1) promotes only few and prominent speakers, (2) is limited by time and space, (3) is not independent of the spheres of money and power, and (4) speakers in the media generally try to win a majority for their own argument, instead of reflecting on their position in the light of counterarguments or rebutting it (Peters et al. 2008). Indeed, there are only few empirical findings on the level of deliberation in different media systems, but they seem to foster the sceptical perspective. Jenkins and Mendelsohn demonstrate, with the example of a popular vote on the sovereignty of Quebec in 1995, that "media coverage of referendums looks much like that of elections" (2001, 211), because it was similarly focussed on campaigning events, persons, conflicts, and strategy. In addition, actors from the political periphery were underrepresented. Ferree et al. (2002) do not find significant differences in the level of deliberativeness between German and US abortion discourse. Nevertheless, while the German discourse seems to meet most of the criteria of representative liberal theory, the US discourse reveals elements of discursive theory, as the media coverage provides a balance of centre and periphery (320).
As mentioned above, in this article, we suggest instead that mediated deliberation is enabled by specific structural features and external conditions provided by political institutions and the national media system in Switzerland. Accordingly, the first condition under which mediated public deliberation can perform well is an institutional one, a political context which can be described as "deliberation-friendly." The concept of direct democracy is frequently linked to notions of public
deliberation because popular votes are an invitation for all citizens to engage in public debates. The initiative and referendum process is a form of direct democracy by which ordinary citizens can either submit potential legislation to the voters or challenge a government's decision (Haussler 2006, 304). Switzerland has a long direct democratic tradition. It is the only European country which has constantly held popular votes at national as well as regional and local levels for about 120 years. At least three key features of direct democracy in Switzerland are associated with deliberative communication. Firstly, direct legislation is, by definition, a lay procedure. Every citizen is a decision-maker. Therefore, everyone has both the right and an incentive to engage in deliberating the issue at stake, which means that besides elites from the political centre, the views and positions of actors from the political periphery become important for collectively binding decisions. Accordingly, openness of the public arena and a plurality of speakers can be regarded as the first feature. The second key feature is arguably understandability and refers to the intellectual level of debates. As the decision is open to every citizen, public discussion does not concentrate on elites and therefore should provide information that is comprehensible to all citizens. Thirdly, as there are no decisions on individual politicians direct democracy focuses on political issues. This feature can be described as an orientation towards substantive policy. Communication must concentrate on the issue itself, rather than on the campaign or on personalities. To assure the voters of their position, a campaign needs to put forward relevant, issue-centred arguments. Thus, decision making in direct democracies can be linked with elements of deliberation (Frey and Kirchgassner 1993; Bohnet and Frey 1994, Bohnet 1997; Scheyli 2000), meaning that political communication in direct democracy should meet some key requirements of deliberation.
The second condition, under which public deliberation may or may not originate, concerns the media themselves. In a direct democratic setting like in Switzerland, the public is likely to participate, at least as an active audience. We assume the media to be an instrument for conveying elite discourses about politics, i.e. the y deliberative quality of mass media output depends on the quality of discussions within societal elites (Bennett 1990; Wolfsfeld 1997; Bennett, Lawrence and Livingstone 2007; Kriesi 2005). If direct legislation actually facilitates political debate among the elites and within the general public, the main function of mass media is to reflect and further stipulate the argumentation of pros and cons (Xenos 2008, 486). In other words, mediated deliberation is based on a journalistic culture that supports norms such as proportionality, elite domination, detachment, civility, openness and non-interventionism (Hanitzsch and Seethaler 2009, 473-475). Following the notion that journalism always reflects the norms and values of the political culture in which it is embedded, we argue that this kind of journalistic culture is optimally fostered in political system with a longstanding direct democratic tradition. This argument conforms to the notion of an educational function of direct democracy, as proposed by Smith and Tolbert (2004). They not only demonstrate how the referendum teaches regular citizens to make use of the possibilities offered by popular legislation, but also push organisations and institutions to adapt
c >
to the potential and imperatives of direct democratic institutions. We assume that this argument holds true for journalism, leading to a news culture which strongly ^ supports the needs and values of direct democracy. The Swiss media system re-
veals a strong public-service broadcaster and widely used quality and regional daily newspapers (Meier 2009). These two antecedents, a strong quality-oriented r^ media and long-term familiarity with the instruments of direct legislation, enable ^ a political communication culture in which mediated deliberation is likely. Journalist surveys indicate that Swiss journalists strongly support direct democracy and the underlying values of the political system. The most common is the profile of the neutral journalist who reports the news exactly as it happens, and indeed, this approach meets the approval of over 90 percent of Swiss journalists (Marr et al. 2008). Swiss media assume the role of a more or less neutral disseminator and widely accept the leading role of politicians in political communication by restricting themselves to a non-interventionist style of reporting of elite discourses (Jarren et al. 2010). This self image is mirrored in the media content: An analysis of a 2008 campaign on the naturalisation of immigrants showed that "most of the coverage has to be considered factual and neutral" (Gerth and Siegert 2012, 287). Consequently, most studies on Swiss referendum coverage suggest that, over the past decades, the media in Switzerland have come to approach referendums as a routine part of their reporting, and that a specific set of institutional norms and professional practices are now associated with this type of event, which helps prevent a reliance on standard patterns of electoral reporting (Marcinkowski 2006a; Gerth et al. 2009; Gerth, Dahinden and Siegert 2012). This distinguishes the Swiss case from others, where direct democratic practices are not a routine part of the political process, and where (rare) direct democratic events receive treatment from the media similar to what is normal for electoral campaigns (Robinson 1998; Jenkins and Mendelson 2001; Vreese and Semetko 2004; Schneider 2005; Hoglinger 2008; Tresch 2008). Quite intriguing regarding our example, Schneider (2003) raised the question of the deliberative quality of public communication in the context of a popular vote on genetics in Switzerland, compared to public communication surrounding a legislative process on the same issue in Austria. She arrived at the conclusion that the Swiss media (and statements made in the Swiss media) demonstrated a higher degree of rationality and reason giving than in Austria. Another current study from Switzerland, which fosters Schneider's thesis, indicates that the media perform well in direct democratic campaigns and "offer a considerable amount of coverage that allows citizens to participate in the arguments of different kinds of political actors" (Gerth and Siegert 2012, 296). To sum it up, we argue that the deliberativeness of public communication increases, if the political system has direct democratic elements, which are widely known and used by the electorate and the elites. The longer these direct democratic procedures exist, the more they are reflected in the political and communication culture. That means, a mature direct democracy like Switzerland proves mediated public deliberation as politicians, journalists and electorate are educated by the specific needs and forms of this kind of decision making process.
Based on our theoretical considerations on the interplay of direct democratic institutions, "educated" journalism, and mediated deliberation we propose the following hypotheses:
H1: Swiss referendum coverage focuses on substantive policy, rather than on individuals and the campaign.
H2: Swiss referendum coverage features a wide variety of speakers, especially from the political periphery, instead of focusing on government elites.
H3: Swiss referendum coverage is shaped by the civility (mutual respect) of the political discourse.
H4: Swiss referendum coverage provides arguments for or against propositions, instead of just referring to statements and claims.
The empirical test of these hypotheses deals with mediated deliberation in Switzerland, a country which matches both of the above mentioned conditions. Elements of mediated public-deliberation will be measured against certain criteria by analysing the media content of direct democratic campaigns in Switzerland.
Research Design and Data
What specific indicators within media coverage have been proposed in current research? The deliberative quality of public communication runs on a continuum between the extremes of no and ideal deliberation (Steiner et al. 2005, 55). A complete set of indicators is still a desideratum. Drawing on our findings from above as well as on Habermas (2005) and Wessler (2008b), who recently summed up the issues, we can arrive at the following media indicators measuring deliberation:
•	Solid information base: Media should provide fair and impartial reporting (Gastil 2008, 52) and shown an orientation towards substantive policy: Communication must concentrate on the issue itself, rather than on the campaign or on specific people.
•	Openness and Inclusion: The arenas of public deliberation have to be accessible for all actors even for speakers from the political periphery, i.e. the deliberative quality of public communication depends both on the inclusion of many and varied actors and their balancing during the debate (Schudson 1992, 147; Ferree et al. 2002, 301). The plurality of speakers will lead to a plurality of arguments (Zhou et al. 2008).
•	Argumentative exchange can be identified as a central value in the deliberative decision-making process (Wessler 2008a, 1199); media content must be analysed with respect to the structure of the arguments presented.
(1)	Reciprocity (Kratochwill 2009, 5); Responsiveness: Actors should refer to each other's arguments.
(2)	Justification: The arguments presented should be based on a transparent and understandable justification.
(3)	Rationality; Complexity: The complexity of arguments depends on the way Q in which counter-arguments are integrated into a speaker's argumentation.
•	Civility; Ideal Role-Taking: Actors should respect each other, which also implies the avoidance of inflammatory speech and personal attacks (Wessler 2008b, 4).
^	The abovementioned indicators describe the positive occurrence of certain
speeches, contents or coverage attributes, e.g. the more speakers or the more arguments presented in the media, the higher the expected level of deliberativeness. We
maacnrnn fVirliiil i]->nro f ix i^cc r~\f fViiii rv-»iiirl i o frl i K1 -i n rlof\ArifK o	t~\f i rl ir^n fr-ir-o
u
S
c >
measured the deliberativeness of the mediated public debate with a set of indicators as described below in Table 1.
The methodological basis of the study is a systematic and standardised content analysis of 4,559 newspaper articles, which cover nine popular votes in Switzerland between 1983 and 2004. The analysis was conducted from April 2007 to June ^ 2008; there were five coders, which were trained on the codebook twice. Reliability
Table 1: Set of Indicators
Concepts	Indicators	Codes
Information	substantive policy information	type of reporting (giving information vs. presenting opinions); perspective (focus of the article on policy, politics or campaign); absence of personalisation and horse-race-journalism
	non-interventionist reporting	journalistic style (descriptive vs. interpretative)
Openness	spectrum of speakers	number, function and background (e.g. member of parliament or individual citizen) of speakers in each article
Argumentation	arguments	number of arguments in each article
	exchange	number of references to other speakers and their arguments in each article
Civility	statements about other speakers	positive vs. negative utterances
	scandalisation	presentation of the issue in form of a scandal
	negativity	promote negative sides of an issue
CT> CT>
was tested with a sample of 100 articles in December 2007. Intercoder-Reliability is between k = 0.7 and k = 1.0 over all variables. The analysis includes all articles covering the respective referendum published up to three months in the run-up to the vote in one of the following newspapers1: Blick (dt); Tages-Anzeiger (dq); Neue Zürcher Zeitung (dq); Mitteland Zeitung (dr); Berner Zeitung (dr); Die Südostschweiz (dr); Neue Luzerner Zeitung (dr); Basler Zeitung (dr); St. Galler Tagblatt (dr); Sonntagsblick (wt), Sonntagszeitung (wq). The sample consists of two daily quality and one daily tabloid newspapers, six regional daily newspapers and two weekly newspapers (one tabloid and one quality newspaper) and represents the leading media (in terms of leading the elites and leading the publics) in the German speaking part of Switzerland. As we needed to ensure that the selected media actually represent the informational basis of the electorate, different media types (regional vs. national; quality vs. tabloid) with different political orientation were chosen.2
Below, we present a short overview of the topics of the analysed popular votes. The case selection focuses on three broader policy issues (international relations, road traffic, immigration) with three popular votes in each policy field. The selection is based on two criteria: First, for a longitudinal analysis only such issues were chosen which came repeatedly to the vote in the last three decades. Second, the policy fields should vary to control issue effects. In Switzerland, popular votes can be differentiated into three distinct categories: 1) Volksinitiative (people's initiative) for changes in the constitution, initiated by more than 100,000 people; 2) Obligatorisches Referendum (obligatory referendum) for parliamentary decisions on international treaties, changes in the constitution or federal law; 3) Fakultatives Referendum (facultative referendum), initiated by the electorate or the cantons, for decisions on certain federal law, treaties etc. To represent all types of direct democratic decision making processes in Switzerland our study covers initiatives as well as obligatory/ facultative referendums.
There were three popular votes on the question of naturalising immigrants during the period of our investigation. The government initiated the three investigated
u
s 5
0
£ *
1 *
(ß 0 c
o
referendums concerning the naturalisation of young immigrants in Switzerland, but the electorate accepted none of them. The first referendum took place in December 1983. The proposal suggested easing the naturalisation process for young migrants who had been raised in Switzerland. The electorate rejected such a measure with 55.2 percent of the votes. A second attempt was again unsuccessful: The referendum on that issue in June 1994 was indeed accepted with 52.8 percent of the votes, but rejected by a majority of the cantons. In September 2004, two proposals on the eased naturalisation of young migrants were again initiated. Both proposals were rejected - the first with 56.8 percent of the votes, the second with 51.6 percent.
The second investigated issue was the sphere of international politics, from which again three proposals have been selected. In March 1986, the Swiss electorate had to vote for or against a full membership of their country in the United Nations. Despite the government's endorsement of this full membership, the Swiss voted overwhelmingly against it with 75.7 percent. An initiative which was supported by government and parliament started a second attempt in 2002 - then, the UN full membership was accepted with 54.6 percent of the votes. The third popular vote with respect to international politics was held in 1994, on the question of foreign peacekeeping missions. The proposal was rejected with 57.2 percent of the votes.
A third investigated issue related to road traffic. There were two votes on the introduction (1984) and renewal (1994) of tolls on national roads, and both were accepted. In 2001, a proposal on speed limits in Swiss cities came to the polls, but the initiative was rejected with a majority of 79.7 percent of the votes.
Results
Mediated-public deliberation requires media space to unfold arguments and to cover the debate. Hence, we first look at the general media attention. If we compare the media coverage over time and differentiate between the issues, we can show that there are different levels of media attention to the subjects of the popular votes. While the topic of international politics yields large number of articles, immigration and traffic issues yield less coverage. However, the total numbers within each subject have remained almost constant over the past thirty years. This means that the media do pay different levels of attention to the subjects of the popular votes, but relatively similar levels of attention over time. One reason for the different level of attention might be the different news values of the subjects or the affectedness or controversy in Swiss society with each particular subject. While international politics are highly controversial in Switzerland, as there is long tradition of neutrality, there is substantial consensus on the question of tolls on national roads. Furthermore, it is possible that eased naturalisation does not at first glance affect the majority of the electorate.
Table 2: Media's Attention (number of articles; N=4,559)
	1980s	1990s	2000s
Immigration	238	346	429
International Politics	932	779	997
Traffic	293	260	285
Our first hypothesis deals with the media focus on substantive policy. That is, do journalists report impartially and fairly? Do they concentrate on substantive policy rather than political strategy or aspects of the campaign? A first indicator of the journalistic treatment of the issues pertaining to Swiss popular votes is the reporting type (Tab. 3).
Table 3: Type of Reporting (N=4,559)
	Share
Information	88 %
Opinion	10 %
Other	2 %
A total of 3,997 of 4,559 articles belong to reporting types can be described as information giving in orientation. These types can be described as neutral in tone and based on fact. For comparison, only 460 articles present subjective material such as commentaries or guest opinions. Thus, the coverage of popular votes concentrates predominantly on informing the citizens and meets (direct) democratic expectations which demand a solid information base. This effect does also reflect the nature of referendum campaigns which are usually more issue focused that election campaigns.
These figures correlate with another finding; the journalistic style is mostly descriptive: 78.3 percent of all articles reveal such a style. Interpretative style functions as an indicator of the degree of journalistic intervention. The journalistic tendency towards intervention in campaigns is normally high, when journalists report the campaign in their own words, scenarios, and assessments — and when they give politicians only limited opportunities to present themselves in the news (Semetko et al. 1991; Blumler and Gurevitch 1995, 2001; Hanitzsch 2007). However, our findings show that journalism in Switzerland is far from intervening. The share of articles using an interpretative style is low and stabile over the decades (Tab. 4). The highest share of interpretative style can be found in coverage of the popular vote on international politics in the 2000s with 16.2 percent.
Table 4: Journalistic Style (in % per issue and decade; N=4,559)
	Ambivalent	Descriptive	Interpretative
Immigration 1980s	15.1	63.9	21
Immigration 1990s	8.7	83.2	8.1
Immigration 2000s	11.9	76.9	11.2
Int. Politics 1980s	7.6	77.5	14.9
Int. Politics 1990s	8.1	80	11.9
Int. Politics 2000s	13.8	69.9	16.2
Traffic 1980s	1	90.1	8.9
Traffic 1990s	2.7	88.1	4.9
Traffic 2000s	7	78.3	12.5
For a detailed analysis of the deliberative quality of Swiss media coverage of popular votes, we now exclude all articles which are too short to present arguments or a certain perspectives, i.e. short news or the mere documentation of paroles. These article types just present brief news or tables, but do not have enough space for subjective deliberation. In the following discussion, the presented results are based on the consolidated sample of 2,307 press articles.
Taking a closer look at the diverging perspectives (Tab. 5) of the media coverage of the nine popular votes, it is evident that aspects of the campaign or the legal procedures itself - like news stories about the latest polls or the financial support of each party - dominate the reporting (43 percent). If we further subtract those percentages of articles which do not present a typical policy perspective (e.g. a political perspective which focuses on actor strategies or personality), and of those articles without any specific perspective, 623 articles (27 percent) remain which directly discusses the issues of the popular votes from a social, cultural, economic, legal or ethical perspective. Only articles from such a distinct policy perspective can be regarded as focussed on substantive policy. In a comparison of the media coverage of national elections in 2003 and on a popular vote in 2005, Marcinkowski (2006b) demonstrated that the share of articles focussing on policy issues in the case of elections is nearly 20 percent, in case of the popular vote, nearly 30 percent. Thus, the share of policy perspectives seems to be higher in the reporting on popular votes, which again is to some extend a feature of this kind of decision making. However, elites and journalists seem to have adapted the "nature" of direct democracy and therefore do actually cover more policy issues.
Table 5: Substantive Policy (N = 2,307)
u
S
i
0
£ *
i *
(R G C
fN
	Share
Policy	27 %
Politics	15 %
Campaign	43 %
Others	16 %
For the present study, we also examined those attributes of coverage which can be regarded as negative signifiers of deliberation. Thus, the presence of the following attributes means that the media do not concentrate on substantive politics. If reporting focuses on certain political stars (personalisation) or tends to describe the political process as similar to a horse race, a loss of substantive policy issues in campaign coverage might be one effect. With respect to our study material, there are some indicators that run contrary to this thesis:
•	Does the story present the private live of politicians? Only a very small number of 118 articles (5.1 percent) does so.
•	Does the story refer to winners or losers? In 131 articles (5.7 percent), the media coverage focuses on the polling results and portrays the political parties as winners or losers. Such a horse-race style would be in contrast to a deliberative style.
Gerth and Siegert stated in their analysis of a 2008 campaign on the naturalisation of immigrants that the campaign was presented as a "contest of arguments
rather than a contest of personalities" (2012, 295), which supports our findings.3 To sum up the results concerning Hypothesis 1: On the one hand, we can state that CO referendum coverage in Switzerland reveals a high level of orientation to informa- 2 tion and impartial reporting. On the other hand, a considerable number articles focus on aspects of the campaign instead of aspects of policies. Thus, Hypothesis 1 can neither be verified nor rejected totally.
Hypothesis 2 states that the spectrum of speakers in the Swiss media coverage of popular votes shows a wide variety of speakers, especially from the political periphery. From the point of view of deliberation theory, the public sphere must be open to all kinds of speakers - from the political centre, as well as from the political periphery. Our data show that there is a tendency in Swiss media to cover more statements of speakers from the political centre (Fig. 5) - Gerth, Dahinden and Siegert found in an analysis of three direct democratic campaigns a quite similar results, as members of the political institutions gathered more media attention in general (2012, 118-121). All in all, 8,506 speakers are cited in the analysed media coverage. The well-established actors of the political centre (members of the Swiss governments or administrations and members of the Swiss parliament - including governmental actors from abroad) together comprise almost 71.2 percent (6,061) of all speakers. Compared with figures from the abortion discourse (Ferree et al. 2001, 90) in Germany, which yield a 73 percent share for speakers from the state or political parties, the Swiss numbers are quite similar. However, compared to the US abortion discourse, they differ remarkably. In the US, the state and political parties only account for 40 percent of all speakers. These differences can be explained in terms of the importance of political parties in the German and Swiss political systems. Other political parties, smaller ones without any seats in parliament, gain a 4.8 percent share in the Swiss media coverage. Associations which cover both powerful actors like the unions and non-profit organisations from civil society, together arrive at a 11.4 percent share, which again is comparable to the German situation (19 percent), but differs widely from the US media coverage numbers (43 percent). 12.6 percent of all speakers in the Swiss media are individuals who are not members of organised parties or any other institutions. There are no great differences over the years or the issues and the shares remain relatively stabile. Another analysis of the speaker spectrum shows that the share of articles with two or more speakers, which indicates an exchange and rebuttal of arguments, differs considerably between the issues at stake. If we take a closer look at these articles, we can find the lowest share of articles with two or more speakers in the 1980s on the issue of naturalisation (26 percent), in the 1990s on naturalisation issues (23.1 percent) and road traffic (31.2 percent), and finally in the 2000s, with the issue of road traffic (35.4 percent) again. All articles about popular votes on questions of international politics yield a higher share of articles with more than two speakers (80s: 56.2 percent; 90s: 58.5 percent; 00s: 56 percent). Apparently, the more controversial the issues, the more conflicts between the political parties, and the more speakers will be cited. As in the cases of naturalisation and road traffic, all major political forces were of the same opinion; the media covered just one of them instead of more party members, which may explain the low figures to some extent. Nevertheless, the overall high numbers of articles with two or more speakers lead to the conclusion that Swiss journalism does provide a mostly inclusive public arena for the deliberation of relevant political issues. Another indicator reflects these findings: In 886 articles
(38.4 percent), Swiss media focus on opposition between the rivalling views. Thus, the reporting presents different positions and opinions on the issue. Based on the evidence provided, Hypothesis 2 can be regarded as verified.
Table 6: Speakers (N=8,506)
	Share
Government/Administration (int)	9 %
Government/Administration (nat)	43 %
Political Parties (Member of the Parliament)	19 %
Other Political Parties	5 %
Associations	11 %
Individuals	13 %
Another criterion for deliberation, which refers to those communicating or being cited in the media, is civility. This can be characterised as the absence of the so-called hot button language and is an indicator of mutual respect. In discourse theory, respect is the basis of a rational and serious exchange of arguments - only when discussants demonstrate respect for each other, can a consensus be achieved. Our study material shows that there is almost no hot button language at all in Switzerland. There are 108 articles with statements about other speakers in the sample (N=2,307), 61 of them reveal negative statements (56.5 percent). Hence, the number of 2,199 articles without any statements about other speakers is, on the one hand, quite positive, as it shows that, in relation to the total number of articles, inflammatory or insulting language is seldom used in the media. Nonetheless, in relation to all statements we find a high share of negative ones, whereas Ferree et al. (2001, 242) only find a 10 percent share of incivility in the US and German abortion discourse. In addition, 32.6 percent of all articles present polemics against one particular opinion. On the other hand, these numbers also mirror the lack y of mutual references - 90.1 percent of all articles do not contain any reference to another speaker. When speakers do not refer to each other, when they do not react to each other's arguments, they therefore cannot make negative comments either. Hence, the mere absence of hot button language is a weak sign of civility in our case. We therefore looked for other attributes of the coverage to determine the civility of the discourse. Only 7.5 percent of all articles (N=2,307) present the respective issue as a form of scandal and only 3.7 percent of all articles (N=2,307) promote the negative sides of an issue. Both attributes do not occur very often, which is an indicator of deliberativeness. Hypothesis 3 claimed that Swiss media show civility and respect in the coverage of popular votes. This hypothesis can be verified. However, the verification is limited. In relation to all statements and all references, the number of negative ones is quite high - although negative does not
c >
necessarily mean disrespectful.
One of the central elements of deliberation theory is the concept of argumentative exchange which in our case yields a fairly small basis. About a third of all
articles (34.5 percent) does not contain any argument at all. Another share of 13.1 percent of all articles features one argument. Two or more different arguments q can be found in 52.4 percent of all articles. A comparison with the argumentative
structure of the German public debate on drug policy by Wessler and Schultz (2007, 21), who analysed deliberation at the level of individual statements in the media, LH shows that only 1.6 percent of all researched statements express two or more ideas. 2 Our numbers are therefore relatively high, if we take into account that our analysis only measured arguments which were listed in official governmental information ("Abstimmungsbuchlein"). The various pros or cons, which were not mentioned in this information, could not be counted. Hence, our criterion is a very strict indicator. We regard Hypothesis 4 as being verified, because there is a comparatively high share of articles presenting arguments.
If we finally compare our measures of deliberativeness over time, no great shifts can be found, even if some of the indicators decreased slightly over the last thirty years (Tab. 7). Thus, Swiss journalism has evidently developed a long-term, stabile programming to report on popular votes. This routine programming combines elements of media and democratic logic. On the one hand, journalists have to rely on news value and transform the issues of the popular votes into news stories. In doing so, they not only focus on highly substantive political debates, but also on personality, strategy, and the horse race. On the other hand, Swiss journalism serves the cause of neutral and balanced information to a high degree, which is the intended and appropriate basis for the decisions of the electorate.
Table 7: Measures of Deliberativeness (N=4,559 for style and reporting type and N = 2,307 for other categories)
	1980s	1990s	2000s
Descriptive Style	77.8 %	83.2 %	74.7 %
Factual Reporting Type	80.1 %	83.0 %	79.8 %
Reference to other Speakers	6.9 %	8.3 %	13.1 %
More than 1 Argument	64.6 %	47.4 %	52.4 %
More than 1 Speaker	49.9 %	47.9 %	50.4 %
Discussion and Conclusion
To sum up, our central research question investigated the deliberative quality of Swiss media coverage of popular votes over the last three decades. Deliberation was measured against certain indicators like inclusion, argumentative exchange or focussing on substantive politics. A content analysis of regional quality and tabloid newspapers revealed that the Swiss media perform very well in informing the electorate about the issues of popular votes. All hypotheses could be verified, although with a varying degree of support. These findings indicate that the political institutions of direct democracy in Switzerland (at least to some extent) shape public deliberation. The deliberative quality of Swiss media coverage of popular votes thus reflects the deliberation within the political elites.
The coverage is predominantly neutral and descriptive. This meets the expectations of impartiality. However, a high level of neutral information is only a basis for deliberation. Other indicators yield similar findings of a moderate type of deliberation. The public arena is inclusive, as all relevant actors from the political centre and the political periphery have a say, but they are not balanced. The
actors of the political centre are overrepresented. We evaluated the civility of the discourse positively. Nonetheless, an argumentative exchange between speakers does not exist and many articles do not present arguments or justifications at all. One reason for the low level of arguments might lie in the repetition of popular votes on the same issue within the decades in question. As the citizens should be familiar with the issue at hand, the media do not provide as many arguments as they did the first time.
As stated earlier in the paper, as a feature of public communication, medi-ated-public deliberation can be located on a continuum between no deliberation and ideal deliberation. On the deliberation continuum, Swiss media coverage of popular votes remains far from ideal, because of limitations within the media (How many speakers can be cited?) and distinct media routines to attract interest (campaigning-perspective and conflict orientation) and not all indicators are fully met. However, having said this, the Swiss media constantly meet the requirements of a direct democracy of information, mediation and reason giving - a routine journalistic treatment of popular votes can be assumed. As this treatment meets direct democratic requirements over the last thirty years in a quite similar way and the media has been challenged economically to a considerable degree, it seems plausible that the institution of direct democracy entails "educated" journalism. But even if direct democratic logic influences Swiss journalism, it is also affected by media logic. Hence, not all measures of deliberativeness score high.
Our findings contribute to the debate on the mass media's role in democracy, as we show that media logic does not per se dominate political communication. The logics of the political system are still of relevance: More participation, more and mature elements of direct democratic procedures can "educate" elites, journalists and electorate and therefore foster public deliberation. Thus, we found indications of a link of institutional setting and political communication culture as Stomback (2005) suggested. Nevertheless, further research is needed to measure delibera-tiveness in different countries, with different institutional settings and over long y periods as well. Such a comparison with other countries would indicate whether the Swiss media systems scores higher or lower than other media systems with respect to the dimension of deliberation.
Notes:
1.	Daily tabloid (dt); weekly tabloid (wt); daily quality newspaper (dq); weekly quality newspaper Q (wq); daily regional newspaper (dr).
2.	The presented analysis does not focus on differences between tabloid and quality media as it looks at the overall quality of deliberation in Swiss media.
^	3. All (now and following) presented comparisons are in so far limited as the codings and countings
Mare not exactly the same. Nevertheless, the figures from other recent research give a hint at the g	explanatory power of our numbers.
References:
Bennett, W. Lance. 1990. Toward a Theory of Press-state Relations in the U.S. Journal of
Communication 40, 1, 103-125. Bennett, W. Lance, Victor W. Pickard, David P. lozzi, Carl L. Schroeder, Taso Lagos and Evans C. ^q	Caswell. 2004. Managing the Public Sphere: Journalistic Construction of the Great Globalisation
i—>	Debate. Journal of Communication 54, 3, 437-455.
Bennett, Lance W., Regina G. Lawrence and Steven Livingstone. 2007. When the Press Fails: Political Power and the News Media from Iraq to Katrina. Chicago: University of Chicago Press.
Blumler, Jay and Michael Gurevitch. 1995. The Crisis of Public Communication. London: Routledge.	O
Blumler, Jay G. and Dennis Kavanagh. 1999. The Third Age of Political Communication: Influences and Features. Political Communication 16, 3, 209-230.
Bohmann, James. 1996. Public Deliberation: Pluralism, Complexity and Democracy. Cambridge: Cambridge University Press.
Bohnet, Iris. 1997. Diskurs - die Protektion der Kommunikation. Analyse & Kritik 19, 1, 15-32.
Bohnet Iris and Bruno S. Frey. 1994. Direct-Democratic Rules: The Role of Discussion. Kyklos 47, 3, 341-354.
Cohen, Joshua. 1989. Deliberative Democracy and Democratic Legitimacy. In A. Hamlin and P. Pettit (eds.), The Good Polity, 17-34. Oxford: Blackwell.
Elster, Jon. 1998. Introduction. In J. Elster (ed.), Deliberative Democracy, 1-18. Cambridge: Cambridge University Press.
Ferree, Myra Marx, William A. Gamson, Jürgen Gerhards and Dieter Rucht. 2002. Four Models of the Public Sphere in Modern Democracies. Theory and Society 31, 3, 289-324.
Frey, Bruno S. and Gebhart Kirchgässner. 1993. Diskursethik, Politische Ökonomie und Volksabstimmungen. Analyse & Kritik 15, 2, 129-149.
Gastil, John 2008. Political Communication and Deliberation. Los Angeles: Sage.
Gerth, Matthias A., Patrik Rademacher, Karin Pühringer, Urs Dahinden and Gabriele Siegert. 2009. Challenges to Political Campaigns in the Media: Commercialisation, Framing, and Personalisation. Studies in Communication Sciences 9, 1, 149-169.
Gerth, Matthias A. and Gabriele Siegert. 2012. Patterns of Consistence and Constriction: How News Media Frame the Coverage of Direct Democratic Campaigns. American Behavioral Scientist 56, 3, 279-299.
Gerth, Matthias A., Urs Dahinden and Gabriele Siegert. 2012. Coverage of Campaigns in the Media. In H. Kriesi (ed.), Political Communication in Direct Democratic Campaigns. Enlightening or Manipulating?, 108-124. Houndsmill: Palgrace Macmillan.
Habermas, Jürgen. 1992. Faktisität und Geltung. Beiträge zur Diskurstheorie des Rechts und des demokratischen Rechtsstaates. Frankfurt a.M.: Suhrkamp.
Habermas, Jürgen. 2005. Concluding Comments on Empirical Approaches to Deliberative Politics. Acta Politica 40, 3, 384-393.
Habermas, Jürgen. 2006. Political Communication in Media Society: Does Democracy Still Enjoy an Epistemic Dimension? The Impact of Normative Theory on Empirical Research. Communication Theory 16, 4, 411-426.
Habermas, Jürgen 2008a. Hat die Demokratie noch eine epistemische Dimension? Empirische Forschung und normative Theorie. In J. Habermas (ed.), Ach, Europa, 138-191. Frankfurt a.M.: Suhrkamp.
Habermas, Jürgen. 2008b. Medien, Märkte und Konsumenten - Die seriöse Presse als Rückgrat der politischen Öffentlichkeit. In J. Habermas (ed.), Ach, Europa. [Oh, Europe], 131-137. Frankfurt a.M.: Suhrkamp.
Hanitzsch, Thomas. 2007. Deconstructing Journalism Culture: Towards a Universal Theory. Communication Theory 17, 4, 367-385.
Hanitzsch, Thomas and Josef Seethaler. 2009. Journalismuswelten: Ein Vergleich von
Journalismuskulturen in 17 Ländern. Medien and Kommunikationswissenschaft 57, 464-483.
Häussler, Thomas. 2006. Die kritische Masse der Medien: Massenmedien und deliberative
Demokratie. Skizze zu einer analytischen Umsetzung. In K. Imhof, R. Blum, H. Bonfadelli and O. Jarren (eds.), Demokratie in der Mediengesellschaft, 304-318. Wiesbaden: VS.
Hatcher, John A. 2008. A Comparative Analysis of Press Systems in 36 Democracies. Journal of Global Mass Communication 1, 3/4, 142-164.
Höglinger, Dominic 2008. Verschafft die direkte Demokratie den Benachteiligten mehr Gehör? Der Einfluss institutioneller Rahmenbedingungen auf die mediale Präsenz politischer Akteure. Swiss Political Science Review 14, 2, 207-243.
u
s 5
o £
0 c
(6
00
Jenkins, Richard and Matthew Mendelsohn. 2001. The News Media and Referendums. In M. Mendelsohn and A. Parkin (eds.), Referendum Democracy. Citizens, Elites and Deliberation in Referendum Campaigns, 211-230. New York: Palgrave Macmillan.
Jamieson, Kathleen H. and Karolyn K. Campbell. 2001. The Interplay of Influence: News, Advertising, Politics, and the Mass Media. Belmont: Wadsworth.
Jarren, Otfried, Schwab Cammarano, Stephanie and Patrick Donges. 2010. Politische Kommunikationskultur im Wandel. Studies in Communication Sciences 10, 1, 29-43.
Kratochwill, Markus. 2009. Webbasierte Deliberation? Potenziale und Grenzen des Web bei der Konstitutierung einer deliberativen Öffentlichkeit. In Sociology in Switzerland: Toward Cybersociety and Viral Social Relations. <http://socio.ch/demo/t_mkratochwill.pdf>
Kriesi, Hanspeter. 2005. Direct Democratic Choice. The Swiss Experience. Oxford: Lexington Books.
Lunt, Peter and Paul Stenner. 2005. The Jerry Springer Show as an Emotional Public Sphere. Media, Culture and Society 27, 1, 59-81.
Maia, Rousiley C.M. 2009. Mediated Deliberation. The 2005 Referendum for Banning Fire Arm Sales in Brazil. The International Journal of Press/Politics 14, 3, 313-334.
Marcinkowski, Frank. 2006a. Media System and Political Communication. In U. Klöti et al. (eds.), Handbook of Swiss Politics, 381-402. Zürich: Neue Zürcher Zeitung NZZ Libro.
Marcinkowski, Frank. 2006b. Medienlogik oder institutionelle Logik in der (Wahl- und) Abstimmungskommunikation? Paper presented to the Annual Conference of Swiss Communication Association (SGKM) on "Cultural Differences and Diversity in Communication Processes", Lugano (TI), 7.- 8. April 2006.
Marr, Mirko, Vincent Wyss, Roger Blum and Heinz Bonfadelli. 2008. Journalisten in der Schweiz: Eigenschaften, Einstellungen, Einflüsse. Konstanz: UVK.
Meier, Werner A. 2009. Das Mediensystem der Schweiz. In Hans-Bredow-Institut (ed.), Internationales Handbuch Medien [International Media Handbook], 592-602. Baden-Baden: Nomos.
Mutz, Diana C. 2007. Effects of "In-your-face"Television Discourse on Perception of a Legitimate Opposition. American Political Science Review 101, 4, 621-635.
Page, Benjamin I. 1996. Who Deliberates? Mass Media in Modern Democracy. Chicago: University of Chicago Press.
Patterson, Thomas E. 1993. Out of Order. New York: Vintage Books.
Peters, Bernhard, Tanjev Schultz and Andreas Wimmel. 2008. Contemporary Journalism and Its Contribution to a Discursive Public Sphere. In H. Wessler, (ed.), Public Deliberation and Public Culture. The Writings of Bernhard Peters, 1993-2005, 134-159. Hampshire: Palgrave Macmillan.
Picard, Richard. G. 2004. Commercialism and Newspaper Quality. Newspaper Research Journal 25, 1, 54-65.
Robinson, Getrude J. 1998. Constructing the Quebec Referendum: French and English Media Voices. Toronto: University Of Toronto Press.
Scheyli, Martin. 2000. Politische Öffentlichkeit und deliberative Demokratie nach Habermas:
Institutionelle Gestaltung durch direktdemokratische Beteiligungsformen? Baden-Baden: Nomos.
Schmidt, Manfred G. 2005. Demokratietheorien. Eine Einführung. Wiesbaden: VS Verlag.
Schneider, Marie-Luise. 2003. Zur Rationalität von Volksabstimmungen. Der Gen technikkonflikt im direktdemokratischen Verfahren. Wiesbaden: VS.
Schudson, Michael. 1992. Was There Ever a Public Sphere? If So, When? Reflections on the American Case. In C. Calhoun (ed.): Habermas and the Public Sphere, 143-163. Cambridge: The MIT Press.
Schudson, Michael. 1997. Why Conversation is not the Soul of Democracy. Critical Studies in Mass Communication 14, 4, 297-309.
Semetko, Holli A., Jay. G. Blumler, Michael Gurevitch and David H. Weaver. 1991. The Formation of Campaign Agendas: A Comparative Analysis of Party and Media Roles in Recent American and British Elections. Hillsdale: Erlbaum. Smith, Daniel A. and Caroline J. Tolbert. 2004. Educated by Initiative: The Effects of Direct Democracy on Citizens and Political Organisations in the American States. Ann Arbor: University of Michigan Press. Steiner, Jürg, André Bächtiger, Markus Spörndli and Marco R. Steenbergen. 2005. Deliberative Politics
q	in Action. Analyzing Parliamentary Discourse. Cambridge: Cambridge University Press.
Strömbäck, Jesper. 2005. In Search of a Standard: Four Models of Democracy and Their Normative
Implications for Journalism. Journalism Studies 63, 3, 331-345.	^
Tresch, Anke 2008. Öffentlichkeit und Sprachenvielfalt. Medienvermittelte Kommunikation zur	O
Europapolitik in der deutsch- und Westschweiz. Baden-Baden: Nomos.	1 '
Vreese, Claes H. de and Holli A. Semetko. 2004. Political Campaigning in Referendums: Framing the Referendum Issue. London: Routledge.
Wessler, Hartmut. 2008a. Deliberativeness in Political Communication. In W. Donsbach (ed.), International Encyclopedia of Communication, 1199-1203. Malden and MA: Blackwell.
Wessler, Hartmut. 2008b. Investigating Deliberativeness Comparatively. Political Communication 25, 1, 1-22.
Wessler, Hartmut and Tanjev Schultz. 2007. Can the Mass Media Deliberate? Insights from Print Media and Political Talk Shows. In R. Butsch (ed.), Media and Public Spheres, 15-27. Houndsmill: Palgrave Macmillan.
Wolfsfeld, Gadi. 1997. Media and Political Conflict. News from the Middle East. Cambridge: Cambridge University Press.
Xenos, Michael. 2008. New Mediated Deliberation: Blog and Press Coverage of the Alito Nomination. Journal of Computer-Mediated Communication 13, 2, 485-503.
Zhou, Xiang, Chan, Yuen-Ying and Zhen-Mei Peng. 2008. Deliberativeness of Online Political
Discussion. A Content Analysis of the Guangzhou Daily Webstite. Journalism Studies 9, 5, 759770.