The Working Paper examines the influence of media intermediaries on opinion formation and discusses models of how diversity in public discourse can be ensured.
From the perspective of communication studies, this working paper reflects on
- the concept and the democratic-theoretical anchoring of journalistic diversity,
- derives standards for optimal diversity from considerations of democracy and publicity theory,
- discusses the effects of intermediaries on journalistic diversity and
- concludes by showing how algorithmic recommendation systems can be designed to make a positive contribution to diversity.
Different theories of democracy and public sphere specify the task of the media and the rules for public discourse. They clarify the requirements for diverse media offerings and are central to the evaluation of the media’s mediation performance. While the provision of these services used to be the sole responsibility of the classic mass media – press, radio and television – in the digital age numerous intermediaries have emerged that now take over these mediation services as well and thus not only change the functionality of classic media, but also influence individual and public opinion-forming processes.
This working paper analyses the influence of media intermediaries on public discourses and discusses (possible) mechanisms narrowing diversity in the context of opinion formation. At the same time, it takes up the relatively new idea in research of how algorithmic recommendation systems can contribute to more diversity of media content.
In this way, it not only answers the fundamental question of why diversity is important, but also why diversity of supply in the digital age is not synonymous with diversity of use.
The most important facts in brief
Why is diversity important?
Media regulation has been linked to the guiding idea of “diversity” for decades. This is because diversity is essential for the democratic formation of opinion and is accordingly an essential criterion for determining the power of opinion. Ideally, the media provide diverse information and opinions on relevant socio-political topics.
On this basis, citizens should be able to form their own informed opinion. Accordingly, the goal of media policy is to prevent certain groups or individuals from gaining too powerful an influence – especially on the formation of public opinion. In this sense, media diversity is considered a (normative) prerequisite for the free formation of opinion, but also for tolerance in a plural society. Indeed, a high social acceptance of diversity is considered desirable and necessary for the cohesion of a society.
How is diversity measured and evaluated?
Diversity as a multidimensional concept can be measured on different levels. The systematisation used here – sources, media, content, use – shows that diversity in content and use in particular are central.
The content diversity of the media offer is mostly examined in relation to topics, opinions, actors and spaces. Against the background of models of democratic theory, there are different implications for journalistic diversity. Accordingly, they differ in terms of who is heard in the public sphere at all (including political representatives, elites, citizens, civil society actors) and in what way the exchange of arguments (rules and quality of discourse) should take place. The basic assumption of diverse points of view is ideally seen in the “marketplace of ideas” and should open up the possibility for citizens to perceive the plurality of society, but also marginalised positions in particular. Normative standards are essential in determining the performance of a media system or individual media sectors, because they ultimately avoid the fallacy of defining high diversity as the sole objective and address the question of the extent to which diversity can also be dysfunctional.
Against this background, the liberal, deliberative and participatory models are considered, three paradigms of democratic theory, each of which sets different normative expectations for optimal diversity in public discourse. In the liberal paradigm, for example, the political elites shape the public discourse and are given space in the reporting according to the political balance of power (proportional representation principle). The deliberative and participatory paradigms, on the other hand, are critical of the elite orientation of the liberal paradigm and call for a differently accentuated opening of the public discourse for civil society actors. They should participate with their topics and positions at least on an equal footing with the elites. The media should also take on a more active role here than in the liberal paradigm – moving away from a purely informative institution to active participants who create the greatest possible diversity of opinion and thus stimulate active public debate.
How does diversity protection change on the internet?
Diversity norms are not obsolete in the internet age. On the contrary: intermediaries pose a number of potential dangers, which are not only attributable to algorithmic personalisation, but also to the economic principles of attention generation in order to sell advertising. This brings to the fore the question of what content is actually discoverable or visible on the user side. For media intermediaries can not only narrow the perceptible spectrum of opinion, but also generally manipulate the perception of topics.
Furthermore, there are limits to both structural and journalistic diversity on the internet – contrary to original assumptions. On the one hand, the diversity-reducing factors on the supply side are also countered by the limited processing and reception capacity of the recipients. For the expanded structural diversity does not necessarily go hand in hand with increased journalistic diversity; moreover, diversity of use is not equally given.
In order to ensure the free formation of opinion on the internet, it is therefore necessary to focus more on the diversity of use, especially as a future design goal in media policy. This reorganisation should also take into account instruments of positive diversity assurance, for example, supporting diverse perspectives of use through the design of algorithmic recommendation systems.
What are the concrete recommendations for action?
- The legislator must make fundamental normative considerations about the desired optimal diversity. In doing so, it is recommended that the paradigms not be seen as mutually exclusive: Different normative standards can certainly complement each other in a diverse media system. For in view of the increasingly heterogeneous public sphere(s), standards from different paradigms should be used for different journalistic and media sectors (e.g. (quality) press, public service broadcasting, social media). For example, the abandonment of the classical mass media can be derived from the liberal paradigm. For other, more peripheral sectors, the standards of the deliberative or participatory paradigm can be decisive.
- With regard to the regulation of opinion power, it is important to detach the focus from the pure provider perspective and to consider both the content-related information and opinion diversity of the journalistic offer, as well as (potential) effects on the side of the users.
- An integral approach to the challenges of media policy – especially by media intermediaries – should take into account different forms of regulation at different levels and with different actors or institutions. Only then can diversity in digital media environments be secured in the long term. In addition to an urgently needed renewal of the existing media concentration law (cf. Reinemann/ Zieringer 2021) and the adjustments already made in competition law, concrete instruments must therefore be developed, for example, which guarantee the discoverability of diversity-enhancing content from the user’s perspective. This also includes the extended areas of application in the new Interstate Treaty on the Media to media intermediaries, which in particular want to create more transparency for users and ensure freedom from discrimination for journalistic-editorial content.
- In this context, the prevailing negative view of media intermediaries and their algorithmic recommendation systems in the media policy debate must be broken down and it must be taken into account that they can also make a positive contribution to diversity. Communication science offers fruitful considerations on how algorithmic recommendation systems can make a positive contribution to diversity. It is crucial that the user’s freedom of information remains guaranteed by creating transparency about the functioning of the recommender systems and leaving the user with corresponding design options.
- Irrespective of this, the need for continuous diversity or quality monitoring remains. It should inform society about which (media) providers present which diversity in form and content with which offers, to what extent they represent the diversity of society and how they are used. In future, this monitoring should also provide more systematic empirical evidence about the diversity that reaches users and its relationship to the original diversity of content supplied to the media intermediaries. For then it would be possible to make an evidence-based assessment of how certain ownership structures and automated selection decisions actually influence content and use, and the complex interdependencies of the diversity concept would be much better taken into account. Media concentration law has so far been based on the truncated assumption that ensuring diversity of providers is sufficient to ensure diversity of content and opinion in public discourse, and automatically translates into a high level of diversity in use. As this assumption can be critically questioned, further research is needed that examines the causal chain in its entirety and is thus able to precisely determine the opinion power of providers in a supply-, content- and usage-centred manner.
The working paper ties in with another working paper, which was developed within the bidt project “Measurement of Opinion Power and Diversity on the Internet” and analyses, from the perspective of political communication research, the conceptual foundations of the current legal requirements for controlling opinion power and their implementation by the KEK as well as the monitoring of media diversity and opinion power by the state media authorities (Reinemann/Zieringer 2021).