Every day around the world, reporters choose not only what stories to write but also what sources to include. Gauging the diversity and range of sources that appear in the news is central to a fuller understanding of the democratic role of the media. Of particular importance is the balance of sources representing different societal interests. Are the interests of businesses more commonly heard than those of employees? To what extent are citizen groups representing, for example, social or environmental causes used as sources in news stories? Does the balance between using different interest groups as sources depend on the policy area being reported on? And to what extent does media coverage of interest groups vary across news outlets?
Our current knowledge regarding how and why interest groups get media coverage is scant.Footnote 2 Most analyses are single-country studies, and the few existing comparative studies deal with interest groups as an aggregate category compared to, for example, political parties and bureaucrats.Footnote 3 This article provides the first cross-country study of interest group appearances in the news media. Drawing on the existing literature, we argue that interest group access to the media largely depends on media preferences.Footnote 4 Media attention is limited, and reporters rely on journalistic norms when choosing which sources to include in news stories. In effect, attention is heavily concentrated among a small number of groups, and economic groups receive a higher proportion of attention than citizen groups.
The range of voices expressed in the news is an indicator of the democratic quality of media systems. Democracy functions best when its citizens and decision makers are informed about different viewpoints on policy problems.Footnote 5 Hence, a main concern of media and agenda-setting scholars is assessing whether the media accomplishes its tasks in a way that ensures the inclusion of a wide range of viewpoints and actors in news stories.Footnote 6 The interest group literature shares most of these concerns. In general, a prominent media profile may be seen as a key indicator of interest group success, especially when (1) access is linked with topics interest groups want to politicize and (2) the reporting adopts a positive tone.Footnote 7 Although interest groups also exert influence in other arenas such as parliament and the bureaucracy,Footnote 8 diversity in media appearances indicates diversity in the range of voices that becomes part of the public conversation.
In this study, we take a first step towards furthering our insight into patterns of group appearance in the media by conducting empirical work across three different media and political systems. Our focus is twofold. First, we are concerned with overall patterns of media appearance by groups, and types of groups. Secondly, we concern ourselves with variation in the set of groups that appears in the news across different news outlets and different policy areas. Here, we are able to move closer to some of the factors affecting diversity among group appearances. We are then able to test propositions about parallelism in national media systems, and about the effects of group policy specialization on the distribution of attention within different policy areas.
To explore these questions, the study compares three European countries. Each represents a different model of media and politics: the UK represents the liberal model, Spain the polarized pluralist model and Denmark the democratic-corporatist model.Footnote 9 Across these cases we expect similar patterns of systematic bias in the attention provided to interest groups – detrimentally affecting the diversity in group appearances. The empirical analysis in this article focuses on newspapers. Although electronic media – including new social media – are increasingly important in political communication, newspapers are still central in news production, and many stories appearing in other media have their origin in newspaper reportage. In each country we selected two of the most important newspapers, ensuring we have one right-leaning and one left-leaning title. We recorded all interest group appearances in the newspapers across twenty-six weeks distributed over a full year and included front pages in the remaining weeks.
The article proceeds in four parts. The first substantive section reviews the group literature and develops expectations that we subsequently test. The data utilized in the article are outlined in the subsequent section, followed by a section detailing our empirical results. The final section outlines our conclusions and assesses implications for future work.
GETTING IN THE NEWS: SOME THEORETICAL EXPECTATIONS
As even the most cursory reader of the print media in democratic nations will be aware, interest groups are often in the news. There are a multitude of good reasons for this to be the case. From a policy perspective, getting into the news might help groups reframe an issue, escalate its salience or mobilize the public as to its importance. The media is a key political venue through which interest groups compete with one another, try to alter the status quo, foster new ideas and policy proposals, and promote policy change.Footnote 10 Several studies have thus demonstrated that groups view media strategies as a central tool for affecting politics.Footnote 11
From a membership viewpoint, media appearances may serve the purpose of reassuring potential and actual group members and supporters that the group is actively pursuing its agenda. Group leaders are constantly aware of the need to maintain and build support for the group; therefore appearing in the media is not only targeted at political decision makers, but also at potential and current group members.Footnote 12 This concern may be particularly pressing for public interest groups that find their supporters among the public at large rather than in specific groups related to, for example, the labor market.Footnote 13
While for interest groups media appearance may generally be seen as a valued asset, there are obvious exceptions. In some instances, groups prefer to stay out of the headlines. For instance, a group may wish to avoid public attention regarding closed-door negotiations with authorities or scandals within groups. However, in most cases there is reason to expect groups to prefer more media attention to less. Groups’ capacity to attract media attention is likely to vary significantly. A range of mechanisms is likely to affect their ability to gain media attention, and thus influence the level of diversity in the groups appearing in the media. In this section we review the expectations around variations in group media appearances, and outline the mechanisms at work.
Concentration in Media Appearances
The relationship between sources and reporters may be seen as one of mutual dependence, in which organized interests and other actors compete in a zero-sum game for control of the public agenda.Footnote 14 In this ‘game’, groups’ capacity to supply content is likely to differ significantly. Thus groups with low levels of financial resources – or those without a dedicated or professional press staff – are likely to be able to ‘push’ less material (and less convincing material) to journalists.Footnote 15 Thus from a purely group resource perspective, we would expect media appearances to be highly concentrated. Further, the media exercise a good deal of discretion in terms of both what stories to report and the sources used.Footnote 16 Media selection is based on factors – for example, status, power and relevance – that make some events and actors newsworthy and therefore increase their chances of making it into the news.Footnote 17
The media tends to concentrate its attention on actors perceived to be legitimate and reliable sources of information due to their position in the political system.Footnote 18 This ‘elite status’ might be explained by formal rules governing the political system – especially in the case of closed policy communities, or neo-corporatist arrangements – as well as informal rules related to the seniority of an interest group or tradition.Footnote 19 The tendency of reporters to develop common understandings of who the important sources are reinforces this effect: thus groups that attract some media attention will be likely to attract more attention in the future.Footnote 20 If this argument is correct, then we should expect the same groups that dominate the inside game of politics to also be dominant in the media arena.Footnote 21 Thus we expect media appearances to be heavily concentrated among a small number of high-status, well-resourced and legitimate groups.
‘Bias’ across Group Types
In light of the rising importance of the media as an arena for interest groups, it is of crucial interest to investigate not only the concentration of attention, but also the range of interest group voices heard in the media. Consistent with the interest group literature,Footnote 22 we address diversity by exploring the mix of types of interest groups appearing in news stories. Interest groups are defined as associations of members or other types of supporters working to obtain political influence. Group members can be individuals, firms, governmental institutions or other interest groups. We distinguish between the following categories of groups: (1) trade unions, (2) business groups, (3) institutional groups, (4) professional groups, (5) identity groups, (6) leisure groups and (7) public interest groups.
The first four types are economic groups related to production in the private or public sector. The two first groups are almost indispensable: a classic discussion thus concerns the balance between business and labor, which is of particular interest in the European context because of the tradition of involving labor market groups in corporatist arrangements.Footnote 23 A third group type is producers and providers of public services. Local authorities in most countries have established interest groups, and schools, universities, museums and other institutions are organized into associations. These are categorized as ‘institutional groups’. Finally, professional groups represent the many different professions represented in the labor market. They are distinguished from trade unions because they do not negotiate work-related terms and conditions.
The last three types of groups may all be seen as citizen groups, but we distinguish between those that represent specific groups of citizens (such as leisure groups and identity groups) and those working for broader causes (for example, public interest groups such as Friends of the Earth or Amnesty International). The latter are distinguished from other groups because they seek ends that will not materially benefit their members.Footnote 24 Identity groups encompass, for example, groups representing demographic or minority groups. Leisure groups’ members are united by participating in a common leisure activity.
What might be our expectations here? Economic groups are expected to have more media coverage than other types of groups. These types of organizations have a higher capability to overcome the difficulties of collective action and attracting resources.Footnote 25 In general, economic groups may therefore be expected to have higher levels of resources (be it financial or staff) than citizen groups. Further, they possess resources such as technical information, which gives them a greater chance of attaining an insider position in the political system.Footnote 26 In turn, they may benefit from the media’s tendency to give more coverage to actors with insider positions. According to the above, we expect economic groups to obtain a higher share of media attention than other groups.
There are two counterarguments to this expectation. First, some citizen groups may also have news value because they represent broadly appealing causes such as protecting the environmentFootnote 27 or ending violence against women. Secondly, citizen groups may be more prone to using outsider tactics such as scaling a construction crane to unfurl banners, engaging in a sit-in or using violence, which are often assumed to be more attractive to the media.Footnote 28 However, even though citizen groups may also possess news value and engage in strategies that attract attention, their capacity to get into the news has a more variable character. We therefore expect the resources of economic groups to be more important for overall patterns of media coverage.
Differing Patterns across Policy Areas
Perhaps the most fundamental aspect of a news story is its topic. Most interest groups tend to concentrate their lobbying work on a very small number of policy areas,Footnote 29 and their appearance in the news is therefore highly contingent on the set of stories reported on.Footnote 30 In stories about labor market politics, major trade unions and business groups are among the most relevant sources to include, while reporters drafting stories related to health issues may choose to hear from groups representing doctors, patients or representatives of public authorities responsible for the health system. Thus it is reasonable to expect variation in group appearances across issue areas, and for groups with a broad policy remit to appear more often in the news than policy specialists. In effect, the level of diversity in group appearances will also depend on the policy area in question. Some areas may only attract the attention of a limited number of groups representing the same type of causes, while other areas may attract a much more diverse set of groups.
Previous studies have found highly divergent patterns of group appearances depending on the issue area. Dimitrova and StrömbäckFootnote 31 found different patterns of sources used depending on the issue analyzed, while Tiffen et al.Footnote 32 explain the high number of judicial sources used in Colombia and the UK by reference to the high levels of crime in these countries. In their study of group appearance on television news, Danielian and PageFootnote 33 similarly found evidence of compartmentalization around issue areas. In a longitudinal study of interest groups appearing in the Danish news media, Binderkrantz found some policy areas such as agriculture, business and consumer issues to be highly dominated by business interests, while more diversity was present in news stories related to health or social affairs issues.Footnote 34 We would therefore expect the policy areas most closely related to the functioning of the economy or specific economic sectors to be more dominated by economic interests than areas of more general regulation.
Media Coverage across Newspapers
Access to the media not only depends on interest group resources and general newsworthiness, but also on journalistic norms and the ideological preferences of the news outlet in question. This relates to the more general concept of political parallelism in the news media.Footnote 35 In many European countries, individual papers historically have had ties to specific parties, and the concept of parallelism implies that a newspaper parallels a party if it is closely linked to that party by organization, loyalty to party goals and the partisanship of its readers.Footnote 36 More recent work has argued that party parallelism has transformed into a more general political parallelism following the transformation of mass political parties.Footnote 37 While the original concept did not refer to the content of newspapers, today a prominent dimension of political parallelism relates to the alignment of newspapers with different ideological, political and cultural views.Footnote 38
In this study, we compare three different models of media and politics.Footnote 39 The UK is characterized as a liberal media model, with high levels of commercialization, high journalist professionalism, and moderate partisanship or political parallelism. While there are clearly papers that could be considered left or right leaning, none have direct one-on-one alignment with political parties. Indeed, papers swing back and forth in their support for a given party during UK general elections. Spain is an example of the polarized pluralist model characterized by newspapers’ strong ideological leaning, low journalist professionalism and strong state intervention.Footnote 40 Finally, Denmark exemplifies the democratic corporatist model in which political parallelism has traditionally coexisted with journalistic professionalism, and press freedom has coexisted with a tradition of strong state intervention in the media. While historically there were close links between parties and the press, these have eroded. Yet the editorial content of papers may still privilege one party or another.Footnote 41
While differences may exist across these countries, our overall expectation is that we can find evidence of political parallelism in the inclusion of interest groups in news stories in all three countries. Ties between papers and parties have generally loosened, but in today’s overcrowded media markets, many news outlets compete not for a mass audience, but for their own niche audience as identified by cultural, ideological and political commonalities.Footnote 42 This may affect not only the choice and framing of stories, but also the type of sources used by different news outlets. Existing analysis supports this argument. For example, despite their strong ideological differences, the two leading newspapers in Spain cover the same topic areas with little variation over time, but important differences exist across these papers when they make explicit references to political parties.Footnote 43 In the same vein, studies of interest group appearances in Danish newspapers demonstrate that the more left-leaning newspapers allow more room for trade unions and public interest groups, while right-leaning papers provide business groups with more attention.Footnote 44 Based on these findings, our final expectation is therefore: interest group media appearances will differ across newspapers depending on their ideological leaning.
In conclusion, we expect a pattern of concentration in the use of interest groups as media sources. A relatively small number of groups is expected to get the lion’s share of attention, and economic groups are expected to be at an advantage vis-à-vis citizen groups. Group appearances are also expected to vary across policy areas, and we expect differing patterns of source use depending on the political orientation of the news outlet.
DATA AND RESEARCH DESIGN
The countries included in this study exemplify three different models of media and politics: the UK the liberal model, Denmark the democratic corporatist system and Spain the polarized pluralist country.Footnote 45 These systems are marked by differences in state laws and regulations, links between media and political parties, and journalistic professional traditions. Historically, they also exhibit very different patterns of interest group representation. While all our countries are European, within this limitation, the research design is a most-different-systems design. For each country, we include two of the most widely read national newspapers (one left leaning and one right leaning).
In the UK, the study compares media coverage on interest organizations in The Daily Telegraph, which has a relative conservative readership, with that of the left-leaning The Guardian.Footnote 46 For research design purposes the UK is our ‘liberal’ case, however we leave open the possibility that it is not perhaps the paradigm example in relation to the degree of parallelism. In Spain we compare the two most widely read newspapers: El Pais (left leaning) and El Mundo (right leaning). Finally, in Denmark we assess the left-leaning Politiken and the Jyllands-posten, which is self-described as an independent liberal paper.Footnote 47
In each case, to identify interest groups appearing in the papers, all articles in the first section and the business section were read, and articles mentioning interest group sources were recorded. Articles that were clearly non-political were left out of the analysis. We also omitted articles in which a group was framed negatively (for example, misconduct by group leaders), as these do not constitute meaningful group access to the media.
All groups identified were coded by the research group and student coders into the group types specified above (with a reliability test resulting in a Cohen’s kappa of 0.91 for Denmark; 0.76 for the UK and 0.82 for Spain). The coding periods were chosen to include a full year in which no parliamentary election took place. For Spain and the UK the period covers 2010–11, while the period for Denmark is 2009–10. For this period we coded all front pages and half of all editions from the two newspapers. Specifically, we coded two full weeks, skipped two weeks, coded two weeks and so forth. This strategy was chosen to maximize the spread of news stories across different policy areas. In total we identified 3,266 group appearances in the UK, 1,754 in Spain and 3,672 in Denmark.
While the research strategy is designed to spread our sample of articles – and thus of group appearances – over a full year and across as many different issues as possible, it is evident that some issue areas may be more prominent on the news agenda during this period than others. Most notably this is the case for stories related to the financial crisis. Overall, this may increase the number of appearances by groups representing, for example, the financial sector and relevant trade unions, but also – as will be seen in the case of Spain – groups that were formed in opposition to official policies in response to the crisis. The interpretation of results must be made in light of this potential bias; yet owing to the media’s focus on a smallish number of policy issues at a time, any time period has the same potential for bias.
ANALYZING INTEREST GROUPS IN THE MEDIA
How Attention Is Distributed between Different Group Types
This section analyses group appearances in newspapers across the three countries. We first address the issue of concentration in media attention, both in the distribution of attention to individual groups and across group types. Secondly, we examine the differences across newspapers and policy areas.
The first issue concerns the distribution of media appearances in our sample of newspapers. Even in a political system resembling a so-called pluralist heaven, we would hardly expect all groups appearing in the news media to get equal amounts of attention. After all, some groups represent very broad and encompassing interests and may therefore continually be relevant for news stories. Other groups are much more narrowly oriented, and their interests coincide with the news agenda only rarely.Footnote 48 So what did we find?
Figure 1 displays the relationship between the number of groups and accumulated share of appearances for each country. It indicates a heavy concentration of mentions among a small number of interest groups. In all three countries, the graphs rise quickly, indicating that a small number of groups accounts for a large percentage of the attention.
Although the curves for each country share the same shape, meaningful differences are found across the three countries. In Spain, the pattern of inequality is particularly marked. Just seven groups (corresponding to 2 per cent of all groups) get half of all attention, and the total number of groups identified in the articles is lower than in the two other countries. The UK has the most diverse pattern of attention: forty-three groups (6 per cent of all groups) account for 50 per cent of all attention, and 693 groups were found in the articles. The pattern in Denmark is in between: twenty-five groups (5 per cent of all groups) get half of the attention, and about 500 unique groups were identified. This pattern is also evident in different levels of Shannon’s H (normalized) across the three countries: for UK 0.84, Spain 0.72 and Denmark 0.80. This measure captures the diversity in attention across all groups appearing in the media; the numbers indicate that the diversity in media appearances is highest in the UK and lowest in Spain.Footnote 49
Notwithstanding this variation in degree across our three national cases, the general pattern is for a small number of groups to account for a disproportionate number of appearances. With so few groups getting such a large share of all media attention, it is interesting to take a closer look at these groups. Most of the groups at the top of the list in each of the three countries are large and encompassing economic groups. In the UK, the three most commonly appearing groups are the Confederation of British Industry, Unite the Union and Trade Union Congress. In Denmark, the Confederation of Danish Industry received the most coverage, and it is also noteworthy that Local Government Denmark, which organizes the local authorities in the highly decentralized Danish government structure, is very prominent in the news media. Finally, the Spanish trade unions receive much media coverage, with the General Union of Workers occupying the first spot on the list. The Spanish list also testifies to the importance of the financial crisis, with the anti-austerity 15-M Movement attracting much media attention.
But are economic groups in general getting more attention than other types of groups? Our next core question was to assess the patterns of group appearances in the national print media to determine whether a dominance of economic groups could be found. A difficulty of assessing bias in interest group representation is that there is no way to know what an unbiased pattern would look like.Footnote 50 Our aim is not, therefore, to evaluate the degree to which the media presents a biased picture of the universe of organized interest, but rather to investigate the media logics that affect the diversity in the range of voices heard in news stories. While we will analyze patterns of appearances within specific policy areas below, Figure 2 presents the predicted number of appearances for each type of group across all policy areas. The figure is based on a multivariate regression analysis that includes dummy variables for each group type as well as for countries, and for the interaction between group types and countries. In this model, significant differences (at the 0.001 level) were present when comparing business groups to all other group types, except institutional groups. Regression results may be found in the online appendix.
The figure illustrates that, on average, three types of economic groups – trade unions, business groups and institutional groups – receive more media coverage than other types of groups. In particular, trade unions – with an average appearance of almost nineteen articles – are more heavily used as sources than other types of groups. The analysis also shows that professional groups are on a par with citizen groups in their average amount of media coverage. When it comes to country differences (not shown), the regression finds that Danish groups appear more often than groups in the UK (difference significant at the 0.05 level). Overall, this supports the expected dominance of economic groups: each individual economic group appearing in the media is likely to receive more attention than citizen groups. How this affects the level of diversity overall and within specific policy areas is the subject of the next section.
Diversity in Group Appearances: Policy Areas and Newspaper Types
Patterns of group appearances have so far been analyzed across the range of policy areas covered in news stories. This may mask divergence between different issue areas, as the relevant sources for news stories obviously depend on the topic of the story. Figure 3 displays diversity in appearances in eight different policy areas (in which at least fifty groups appeared in each country) and compares this with diversity within the data as a whole. While above we calculated diversity based on the appearances of individual groups, here we focus on the distribution of attention across the seven different types of groups.
It is evident that diversity varies across policy areas, and that the overall pattern is similar across the three countries. Diversity is at its lowest in news stories related to the labor market, and at its highest in coverage of health issues, legal and justice issues, and issues related to education. It is also interesting to note that the spread of attention across group types is lower than the spread of attention in the full set of news stories for most policy areas.
To further analyze which groups are getting attention in different policy areas, Table 1 displays the types of groups appearing in four different areas: macroeconomics, labor market issues, legal and justice issues, and social and family issues. These areas have at least one hundred interest group sources appearing in each country.
Note: a chi-square test indicates that differences between policy areas are significant (0.001 level).
The overall patterns of group appearances in the three countries correspond to the analysis above, which finds that economic groups are more prominent than citizen groups. While this trend is most pronounced in the Danish case (possibly due to the country’s corporatist traditions), trade unions are also particularly successful at gaining attention in Spain. As expected, the patterns of group appearances differ across the four policy areas.Footnote 51 Economic groups are most dominant in issues related to the labor market – and in all countries this is largely explained by the extensive coverage of trade unions. Citizen groups are more prominent in articles covering legal and justice issues in the UK and Denmark because of the high levels of attention given to public interest groups, while identity groups are more prominent in Spain.
In the two other policy areas, country-specific patterns are more pronounced. It is particularly noticeable that public interest groups are widely mentioned in Spanish articles about macro-economic issues and in UK stories about social and family issues. In the Spanish case this is largely due to a high level of citizen mobilization in reaction to the financial crises (for example, the 15-M Movement), while the UK pattern is probably related to the engagement of a large number of citizen groups in providing social support for vulnerable citizens.Footnote 52 Alongside the overall dynamics related to policy characteristics, more situational factors also affect patterns of group appearance across policy areas. The overall pattern, however, supports our expectation that the policy areas most closely related to the functioning of the economy are more dominated by economic interests than are areas of more general regulation.
Our last question to explore is the issue of political parallelism. We expected right-leaning newspapers to disproportionately favor business groups, and left-leaning papers to give priority to groups traditionally related to the left wing such as trade unions and public interest groups. Table 2 disaggregates the data on media appearances for the two different newspapers examined in each country. For each group type it reports the mean number of mentions in the left-leaning and right-leaning papers among all groups appearing in our dataset.
Parallelism, defined as significant differences between the mix of groups used as sources between left- and right-leaning newspapers, is found in all three countries. We focus our attention on salient differences between our newspaper sources and among group types. First, as Table 2 illustrates, the differences in the media coverage of interest organizations in the UK are as expected. In this case, mean differences are significant for trade unions, business groups and public interest groups (see significance levels reported Table 2). In Denmark, business groups receive more attention in the right-leaning JP and public interest groups more in Politiken, but a difference is also found for identity groups that get more coverage in JP. This is even more marked in Spain, where the right-leaning paper El Mundo pays more attention to identity groups and the left-leaning El Pais to institutional groups. This pattern may be explained by the prevalence of some issues on the political agenda, and more specifically to the importance that El Mundo places on the victims of ETA terrorism.Footnote 53 Overall, these findings illustrate that media coverage is distributed unevenly across newspapers in all three countries.
DISCUSSION AND CONCLUSION
Media appearances can be a valuable asset for interest groups; however, there are clear limits on how many groups can be included in news stories. When explaining the patterns of group appearances in the news media, it is therefore necessary to draw on theories about media selection. Based on such theories, we developed expectations about: concentration in media attention, bias in the types of groups getting coverage, and differences across issue areas and newspapers of different political leanings. The article tested these expectations in the first multi-country study of media appearances by interest groups. By comparing three countries with different media systems, we have been able to demonstrate similarities as well as differences across countries.
In all countries, media attention is highly skewed. A small number of groups receive the vast majority of attention, while most other groups appear only once or twice in the material. From a normative perspective, this means that the range of voices available to citizens is far from an equal reflection of the interest group society.Footnote 54 In particular, citizen groups are much less commonly referred to in the news than economic groups such as trade unions and business groups. With regard to the political process, to the extent that politicians rely on the media for input into their political work, it also means that the input received will be skewed towards certain interests.Footnote 55 While many of the groups – such as major trade unions or business groups – cited most often in the media represent broad encompassing interests, it is evident that biases are present in the way interest groups are used as sources, and that these biases are detrimental to the level of diversity in group media appearances.
The use of sources is highly dependent on the topics reported on, and large differences in patterns of group appearance are found across policy areas. In some issue areas economic groups are close to exercising complete dominance, while more diversity is found in other policy areas. Here, situational differences and country differences are also at play. The three countries have rather divergent interest group systems, and the policy agenda in the analyzed period differed. Most notably, the Spanish media paid a great deal of attention to the economic crisis and terrorism. These patterns also affected the comparison of right- and left-leaning papers, as the Spanish El Mundo paid a lot of attention to identity groups representing the victims of ETA terrorism. In addition, evidence of parallelism was found across all countries, with systematic differences in the patterns of group appearance in different papers.
In a situation in which ties between political parties and newspapers have loosened, and interest groups and parties are much less closely aligned than in previous decades,Footnote 56 it is interesting that parallelism is present in the news coverage of groups. The three countries represent different models of media and politics, and thus there is reason to believe that these results may be generalized to other European countries. Parallelism may come about not as a result of organizational links, but rather as an effect of competition between papers seeking their audience in niches identified by cultural, ideological and political commonalities.Footnote 57 For readers of newspapers, this means that the selection of topics reported on, the framing of the stories and – as illustrated in this article – the set of organized interests they hear from depend on their choice of news outlet. If most citizens refer to only a limited set of media outputs, this will limit the diversity in the viewpoints to which they are exposed.