From News Diversity to News Quality: New Media Regulation Theoretical Issues

Online platforms and services have challenged the press industry for more than twenty years. Internet media made possible the proliferation of news from a wide variety of sources, including non-professional sources, on different supports (tablet, smartphone, computer), and with high diversity of alternative viewpoints and opinions. In light of concentration in the media industry and lack of media independence, now is an ideal opportunity to give way to alternative media models, to social minorities, to heterogeneous political interests—that is to say to media pluralism.

However, one may notice: Diversity is not so much the problem in the digital context with the dramatic increase in the amount and variety of news available online. Information pluralism, in its normative sense of what journalism must be to accomplish its democratic role (Karppinen 2018; Rebillard and Loicq 2013; Merrill 1968), is far from achieved. Indeed, diversity and abundance are no longer synonymous with more quality, open debates, and equality. There are many obstacles in the way: circulation of fake news, hate speech and illicit content, social and political polarization, extremism in content and in users’ opinions, cultural and ideological confinement of readers because of platforms’ recommendation systems, large-scale diffusion of poor and meaningless information.

These phenomena clearly contrast with the widespread idea—notably among mainstream economists—that competition goes hand in hand with efficient outcomes. This diversity-pluralism discussion clearly necessitates an introduction of a third component, information quality, and a set of rules—regulation—to guarantee some minimal quality of the information in circulation.

Potential regulation tools may target various parts of the production value chain: economic models of news providers and middlemen (incentive regulation), media concentration (antitrust), actual conditions to access and impart news, media literacy, rights to information. One aspect rests crucial: Without any clear definition and measure of quality, media regulation will be difficult to build.

More precisely, we propose fundamental difficulties in regulation reside in three main factors: (1) multidimensional nature of the concept of information quality, widely discussed in the literature (are there some prior dimensions to be respected or all of them are of the same importance?); (2) multiplicity of agents’ wishes and needs in respect of the information quality (whose needs should be covered by the regulation taking into account that different consumers may diverge significantly in their preferences?); and (3) lack of empirical evidence about the impact of different value chain modalities (economic models, property rights, or how we access information)—i.e., potential targets of regulation—on the quality of the circulated information.

This scholarly article discusses these three elements and proposes an original framework for the analysis of the information quality with new dimensions and with respect to the heterogeneity of consumers’ preferences. This framework may be also used to get a legible presentation of the link between the existent media strategies regarding produced information and their economic models, property owners, or any other available media characteristics. Such comprehensible representation makes possible implementation of specific regulation tools in order to promote information pluralism and democracy.

In the next sections, we, first, propose a short overview of the impact of digital platforms on media industry, in general, and information characteristics, in particular. Second, we make a literature review on how the quality concept is approached in different works and some inherent limits of the proposed definitions. On this basis, we formulate, in the third section, an original approach on how to analyze news quality. The final section proposes a short empirical example of implementation of our approach to real market data, based on the French media industry.

Digital platforms such as Twitter, Facebook, Instagram, or TikTok have become information gatekeepers, prompting a new set of concerns impacting the production, diffusion, and consumption of information. They encourage new kinds of ‘alternative’ actors—readers, politicians, activists, etc.—which produce their own content in the information marketplace. This mixes up with established media and their newsrooms, which removes newsrooms as lone gatekeepers of the news selection and delivery process. As a consequence an unprecedented plethora of heterogeneous information, from professional news to info-tainment and user-generated content, now competes for readers’ attention.

Foremost, by providing direct access to news articles and new practices of ‘picking,’ digital platforms enable readers to bypass the front-page of online outlets. In so doing, they challenge the editorial line of the outlet and their traditional revenue models. This is doubly adverse for media increasingly dependent on advertising revenue. Symmetrically, online intermediaries (‘infomediaries’) have become essential and control direct relationship with readers, enabling the intermediaries to collect vast amounts of personal user data. Since access to news by individuals passes often through social media, producers of information are subject to disadvantageous pricing conditions imposed by platforms when sharing advertising revenue. Consequently, one may observe that the audiences’ capture—based on Google Trends, ‘likes,’ Google Analytics, and so forth—is the main criteria of performance. Or, that in-depth, costly, staff-written articles make way for wire service copy and the plagiarism associated with (almost) zero production costs. This pressure adds to that felt from investors, who tend to focus on short-term economic results and require newspaper managers to cut back on resources to increase profitability (already threatened by reduction in advertising revenues).

Then, media industry has become extremely dependent upon audience/web metrics. Journalistic and professional practices are polarized around data produced by the algorithms of digital platforms. The press, newsrooms, and journalists share the same data that is publicly available or produced by the same algorithms. Speed-driven journalism illustrates the impact of such a generalized behavior where, on the one hand, the selection and treatment of news topics are determined according to their online popularity, and on the other hand, articles only survive in the light of their own popularity on the internet. Such behaviors impact the quality and diversity of information, whether in terms of the reduced variety of subjects addressed, or in focusing the attention of online readers on a small number of ‘star’ topics.

Finally, another important transformation element is situated alongside both the users’ side and the practices associated with news consumption. Importance of digital platforms as main support induces not only the availability of free offerings (and therefore the problem of profitability for producers), but also new constraints on the content format, information deepening, and complexity. Circulation of short, easy-to-read-and-share news (not necessarily fakes) attracts larger public and stimulates network effects based on interactions among platforms’ users’ on news. This permits the massive collection of users’ data by the digital giants reinforcing their market power as essential intermediaries in access to online news. As a result, today consumers often have strong preferences for short and descriptive ‘snack content.’ More generally, networking tools such as ‘likes’ or ‘retweets’ expand mimic behavior through which individuals select and share the same kinds of information, whatever the intrinsic quality of veracity, originality, or richness of that information. In addition, the well-known fact that the content selection procedures are often determined by the platform’s algorithms and user’s previous choices then reduces the user’s feasible search space. This could lead to filter bubble (Pariser 2011), where users get less exposure to diverse, conflicting viewpoints and prove to be isolated in their own informational bubble.

To summarize, these transformations lead to a higher variety of sources of information (blogs, social networks, alternative media, and more) and of conditions stipulating the production and consumption of news (end of the monopoly of traditional media, data-driven journalism, etc.). The widespread and often free provision of digital tools and services on these platforms has undeniably lowered the barriers to entry in the production and distribution of news: less capital requirements to create and sustain an outlet; decentralization of production sources; sharp reduction in distribution costs; and more yet unlisted. In this context of information abundance, the problem of news quality has become crucial in this changing environment. It is not surprising that one may observe today flourishing organizations and initiatives aimed at developing new standards in journalism, supposed to sustain and to improve information quality, and consequently, media pluralism. Among them are The Fourth Estate, The Independent Press Standards Organisation, The Organization of News Ombudsmen and Standards Editors, The Media Pluralism Monitor—this list is far from exhaustive. While being mainly focused on the supply side and the characteristics of the produced content, these initiatives create more questions than answers (Karppinen 2018; Carpentier and Cammaerts 2006; McLennan 1995).

To shed some light on these questions, we need first to clarify the concept of quality.

Beyond some differences between USA and European traditions, the concepts of pluralism and diversity are quite established in media studies literature. Whereas pluralism refers to a normative orientation and democratic role of journalism, diversity is understood as its measure, i.e., the heterogeneity at the level of various production elements (Karppinen 2018; Rebillard and Loicq 2013). The dimensions of diversity vary depending on the levels of analysis ranging from the media ownership to the content characteristics. Among the most well-known classifications of information diversity, Napoli (1999, 2003) proposes to distinguish between sources diversity, content diversity, and exposure diversity (both vertical and horizontal). Source diversity refers to ownership and nature of outlets; content diversity concerns formats and viewpoints; and exposure diversity addresses either the variety of information provided by different outlets available to consumers (horizontal diversity) or the variety of information provided by an individual outlet (vertical diversity). For McQuail (2005) the main dimensions of diversity are ‘genre, style of format in culture or entertainment; news and information topics covered; political viewpoints; and so on.’ Rebillard (2012) proposes a definition based on topics’ diversity, topics’ equilibrium, and their disparity of treatment.

In the aforementioned approaches, diversity is used as synonym to news quality: The presence of this characteristic in the journalistic information supply available on the market is supposed to be associated with higher quality and further, the necessary and sufficient condition for normative media pluralism. However, especially in the digital context, diversity does not necessarily imply any quality (e.g., circulation of poor or wrong information, content promoting illicit activities, etc.). Moreover, the problem arises when one tries to deconstruct the value of pluralism: Are there some limits to diversity and whether at some point ‘healthy diversity’ may turn into ‘unhealthy dissonance’? Are all interests equal in defining the quality attributes or some issues are of higher priority? How to connect individual-level criteria to social-level outcomes which are more than just a summing of individual utilities? In Karppinen (2018) one can find a detailed discussion on this matter. The author argues that in the context of growth of digital media, characterized more by abundance than scarcity, the conceptual ambiguity and divergence of definitions of normative and political frameworks are stronger than ever, making difficult their measures and promotion of some quality standards.

To tackle this problem, some authors elaborate on the idea of the quality of the journalistic information, complementary to diversity. According to Meyer and Kim (2003), quality definition comprises two levels—organizational and content ones. The former level focuses on the media or press outlet as unit of analysis, whereas the latter one is inherent to the produced information. Often these levels are somehow mixed up in media studies literature.

At the organizational level, quality indicators can correspond to the reputation of the media (Stone et al. 1981) or its organization characteristics such as integrity, staff enterprise, community leadership, editorial independence, staff professionalism, editorial courage, decency, influence, and impartiality (Gladney 1990). Among other features, Merrill (1968) proposes such criteria as (1) financial stability; integrity; social concern; good writing and editing; (2) strong opinion and interpretive emphasis; world consciousness; nonsensationalism in articles and makeup; (3) emphasis on politics, international relations, economics, social welfare, cultural endeavors, education, and science; (4) concern with getting, developing and keeping a large, intelligent, well educated, articulate and technically proficient staff; (5) determination to serve and help expand a well- educated, intellectual readership at home and abroad; desire to appeal to, and influence, opinion leaders everywhere.

Bogart (1989) conducted a large survey asking newspaper editors how they rate different attributes of newspaper quality¹: accuracy, impartiality in reporting, investigative enterprise, specialized staff skill, individuality of character, civic-mindedness, and literary style. Bogart justifies for the choice of such subjective criteria because they are allegedly commonly shared by newspaper editors themselves when assessing the quality of their own outlets.²

Rather in line with the content level, abundant literature proposes originality, diversity of topics and comprehensive coverage, comprehensive coverage, accuracy of reporting and expert judgment, timeliness, and novelty as indicators to characterize news quality. Picard (2000) considers quality as related to the amount of journalistic work in terms of investigation, verification, and sourcing that have been carried out before writing the news.³ Abdenour and Riffe (2019) note that, in general, academics tend to focus on strong investigative reporting to infer news quality indicators. McQuail (2005) holds honesty and checkability as the characteristics of journalism products. In addition, quality can be defined by negative indicators like shallowness, incompleteness, inaccuracy, bias, or misinformation (Craig 2011; Urban and Shweiger 2014). For instance, Magin (2019) shows that tabloidization generates lots of news of lower quality and associated with a large share of ‘politically irrelevant topics, a focus on episodic framing and a visual, emotionalised, opinion-driven style.’ In addition, some researches put in light the value of audience engagement and digital interactivity on the news as quality dimensions (see, for instance, Bogart 2004; Blanchett Neheli 2018; Belair-Gagnon 2019).

Finally, some authors argue that, in addition to the supply-oriented approaches presented above, quality definition must take into account the type of actors who define it, i.e., academics and journalists, politicians, judges and lawyers, or final users (Meier 2019; Lacy and Rosenstiel 2015). For example, the focus on the consumers’ and recipients’ perceptions (Lacy 2000; Urban and Schweiger 2014; Rosenstiel et al. 2015; Lacy and Rosenstiel 2015) is in line with the Federal Communications Commission’s proposition: ‘As an alternative to measuring the “supply” of content to assess viewpoint diversity, should we take a “demand side” approach and utilize measures of audience satisfaction and media consumption as proxies for viewpoint diversity.’ In this demand approach, any news is considered as a bundle, and ‘[its] quality aggregates individual consumer’s perceptions of how well journalism serves their needs and wants’ (Lacy and Rosenstiel 2015).

Some quality elements (e.g., news accuracy or fairness) are quite the same as the supply approach that surveys media managers and journalists. However, demand approach implies that, contrary to the supply approach where the concerned economic actors share the same concept of quality, ‘two different news consumers would evaluate the quality as being different because of their differences in information wants and needs’ (ibid.). Depending on the types of individual media uses and motives (e.g., feel connected to community, decide places to go, stay healthy, etc.), which differ from one consumer to another, the news consumer evaluates quality of the content, i.e., how well it meets individual information needs and wants.

The definitions of news quality raise a set of problems.

As demonstrated in Urban and Schweiger (2014), ‘the level of analysis in these studies is quite broad. Most of them ask for recipients’ evaluations of whole media brands like the New York Times or whole media genres like newspapers or news websites. Hence, the results cannot say much about recipients’ concrete evaluations of different news items. They rather express an aggregate opinion over a variety of articles, sections and editions. It remains unclear which part of the coverage was judged.’ Then, often in different studies, the set of criteria chosen to evaluate information quality is the result of a specific filtering through the lens of academic expertise and not directly derived by surveying recipients’ expertise. Therefore, sometimes it is unclear whether media users or journalists taking part in surveys and experiments are able or understand—at least, in the same way as researchers—the indicators they have to rate in order to evaluate quality of news (Urban and Schweiger 2014). Such research biases can negatively impact overall results.

The organization-level criteria of news quality, presented in the literature, may be called into question. To our knowledge, there is no empirical evidence in the area of causes and consequences, or more generally relationships, between different production context and organization and the characteristics of the produced news. For instance, we cannot say, without an in-depth study, when news are staff-written the characteristics of honesty or checkability of information produced by journalists are always present; or that there is a link between media’s sources of revenue (e.g., advertising) and accuracy; or again that media ownership has always a direct impact on the spectrum of the subjects that covered by this media or that actually interest the public, and so on.

At the content level, particular sets of news characteristics as quality standards puts the problem of perspective. For instance, investigative reporting may be one of the dimensions of quality from the supply-side, professional perspective. However, from the demand-side perspective, it may be of no importance for some users’ information. In the same manner, in some contexts readers may appreciate shallowness (for instance, if one wants to have a short review of current affairs) or opinion-driven style in news (for instance, if one wants to know the viewpoint of a political party or an interest group on a given situation). For the same reasons, focusing exclusively on the demand-side may also be insufficient to define the news’ quality, taking into account heterogeneity of information users and their needs and wishes in respect of the news consumption. Finally, quality criteria do not necessarily meet unanimity among both journalists and readers: ‘legitimate’ journalists will be able to put forward some criteria that otherwise are not considered as important (or even evaluated negatively) by certain categories of readers (geeks in a hurry, compulsive information sharers, and so forth) or journalists (gonzo, alternative, or others entirely).⁴

Facing these problems we suggest that some consensus can be found due to the introduction to the analysis of the information as example of economic goods. In example: a product that ‘satisfies human wants’ and is ‘exchanged on the market’ (Milgate et al. 1987). In this economic sense, goods are not ‘just physical objects, but the qualities with which they are endowed’ (ibid.); their value is the combination of ‘objective conditions of production’ and ‘subjective conditions of their consumption’ (ibid.).

When considered this way, the definition of the news quality may be reformulated as follows. First, journalistic information is produced to satisfy concrete users’ needs and wishes. Users may be final consumers, regulators, industrial actors, etc.; their needs may be of various natures (to be informed of the news, to share an opinion, to sustain political diversity or polarization, etc.). In this sense, different characteristics of journalistic information (rates of staff-written content, checkability, covered subjects, etc.) and their combinations are supposed to satisfy various wishes and needs. Second, among various users’ wishes, as for any other economic good, one can find those unanimously assumed by all actors (e.g., preference for being informed by truth information at the consumers’ side, preference to be considered as reliable media for producers) and those for which users’ (and other actors) may differ a lot. For the former needs, the associated information characteristics refer to a sort of ‘objective’ goal and underlying condition for information pluralism. The latter range of users’ needs implies the presence on the news market of diverse information characteristics attracting heterogeneous users.

The quality of information exchanged on the market (or its value) resides in the existence of sustainable economic models which lead to the availability of the market of journalistic information capable to satisfy the both criteria: unanimous features and diversity criteria.

We propose to consider this condition, which, contrary to previous definitions goes beyond the characteristics of the content or its production conditions, and which makes a bridge between supply-side and demand-side perspectives, as necessary requirement for information pluralism. Our next section develops in detail this approach.

We applied our conceptual framework to make an empirical study on the news quality produced by French news producers. Using a linguistic discourse analysis method on 31 striking events and 93,648 articles published over the 2015–2019 period in France, we characterized their editorial choices of 55 representative media outlets. This study shows that media strategies nowadays make it possible to produce journalistic information that is sufficiently differentiated horizontally to meet distinct consumer needs, albeit with significant disparities in vertical quality.

Our sample is made of traditional press media, all online digital media players (e.g., Yahoo News!), and ‘alternative’ or ‘partisan’ left- and right-wing media. Each selected event meets two main criteria: is singularized without ambiguity as well as containing associated key words in our publishing window. To measure the two types of news quality available in the market, we used the following dimensions as criterion. On the one hand, the originality of the information was used to assess the vertically differentiated quality. We assumed that to avoid copy-and-paste of wires or plagiarism is commonly accepted among readers. The doc2vec method (Le and Mikolov 2014) was applied to measure the semantic distance between press articles and all previous AFP wires on the same news topic. On the other hand, argumentation or analysis, as an added value proposed by journalists in the articles, was used to assess horizontal differentiation. A rhetorical analysis of documents (Roze 2013) quantifies the relationships between sentences and the presence of morphosyntactic indicators referenced in the literature. We calculated an argumentative index that makes it possible to distinguish between articles containing an analysis, consensual content or a discussion of a current subject, articles based on facts or precise positions leaving less room for analysis. This index is implicitly associated with heterogeneous tastes and preferences of news recipients.

Crossing the vertical dimension (originality) and the horizontal dimension (analysis) makes it possible to visualize the variety of editorial strategies for media outlets. On the following graph, the horizontal axis intersects the vertical axis at the median value of the maximum distance between press articles and AFP wires for the same topics. The AFP is logically located at the bottom of the southwest square associated with no significant analysis and no originality. This mapping of news providers presented above is based for each title on the aggregate values of originality and analysis for several events treated during the period 2015–2019. Depending on the nature of the event as well as the political, thematic, and geographical orientation or the level of their own resources, media may cover news topics in very different ways (Fig. 4).

Each quadrant corresponds to a type of news providers—two categories are rather specialized and then located mainly in one isolated zone. Magazine presses (monthly or weekly/paid-for publications) are mostly present in the North-East quadrant where high vertical quality (argumentation) is associated with high production costs that require important funding. The alternative media (‘partisan’ news) are also specialized but prevail in the North-East quadrant: They are characterized by the originality of information they produce but not the treatment of all viewpoints and aspects of the topics they cover (analyze). Their readers belong to well-identified and stable communities sharing similar opinions so that those media do not systemically try to set out arguments to convince their readership by discussing the opposite stakes, points of view, etc. By contrast, national daily press is present in all quadrants in a quite balanced way with some of the newspaper opting for two different strategies. Publishing strategy can also lead national newspapers to replicate agency wires (low originality) and after to significantly enrich some articles according to the importance of the topics, sometimes to differentiate from rivals. As for regional newspapers, they are slightly more present in the South-West quadrant. While they publish news close to the AFP wires, those media outlets differentiate horizontally in terms of argumentation. This result can be explained by the fact that our survey covers only national and international events less treated by regional newspapers which allocate more human resources to local events and topics.

This analysis highlights the editorial strategies of French news producers. There is a relation between those strategies and news quality. We also show that variety and viability of editorial choices of media can be closely linked to their economic models. These media have expanded in the market during two decades of digital transformation. For example, national and regional media, in order to foster their position as leaders, systematically explore diversified editorial strategies to generate new kinds (and streams) of revenues. Their strategies, despite their substantial resources, are concentrated in areas with higher demand (and therefore profitability) and not necessarily with higher added value (as suggested by the overrepresentation of these players in the north-eastern quadrant in our graph). For alternative media, the only way to survive this competition is to position itself in highly targeted niche markets. These actors must unite loyal and stable communities of readers to finance their production so as not to disappear or be bought by bigger media groups. But an excessive dependence upon a readers’ community might lead them to an overproduction of news consistent with the opinions of the target readership. For more generalist new media creators, which were largely free when they emerged in the media landscape, a choice is currently being made between adopting a pay-for model or remaining free on condition of producing low added value information. Advertising revenues, which drive this decision, prove insufficient to profitably finance production in the press sector. The problem with these choices is that the financial success of the pay-for strategy is not guaranteed if the number of actors adopting it increases significantly.

Finally, our study suggests that the criterion of horizontal quality (diversity) of the information provided by the French media is satisfied because the large supply of vendors is likely to meet the readers’ tastes and needs for different types of information. Readers can thus find any type of information by focusing on one type of media or by combining several media according to their preference. There are also all types of media in terms of vertical quality. However, a low willingness to pay of readers might lead them to select a (too) low vertical quality, raising a crucial issue for liberal democracies and therefore for regulation: How can media be encouraged to improve the quality of their supply or readers to increase their willingness to pay for higher vertical quality? For instance, regulation could aim to increase the readers’ willingness to pay for higher vertical quality news without restraining their access to horizontally differentiated information. This could consist in influencing their individual preferences—training young and easily influenced people to better identify fakes or misleading news—or increasing their real income by awarding virtuous news producers (reduced taxation) and overtaxing those providing recipients with excessive amounts of copy-paste from wires or proven fakes. Implementing such a cross-taxation might be possible precisely by being able to identify and target outlets according to the nature of news quality they provide in the media market. The range of such tools then may cover a large area of instruments from education, promotion campaigns, taxation, and subsidies to sanctions and rewards.

In this chapter, we discussed the issue of news quality as a key aspect of media pluralism. We elaborated upon a theoretical framework that permits to evaluate the news quality in an economic perspective. Depending on the definition of quality (veracity, independence, etc.) and the stakeholders’ expectations, the criteria used for the analysis can be classified according to the vertical and horizontal axes. As a tool, the resulting media mapping helps identify editorial areas with high media concentration against underdeveloped areas. This approach can serve to explore different phenomena in the media ecosystem and, from a regulation perspective, to correct new and traditional market failures. Once we agree on the description of the news by means of needs and features (i.e., as an economic good), understanding of the factors determining users’ preferences and producers editorial choices opens the discussion on potential regulation tools.

On the side of readership preference: Cultural capital, the social influence by peers (the media and type of news that social neighbors look up), the social media used by readers, and more generally their demographics play a key role in forming their preferences. Thus, these factors are among the main factors that could impact the quality expectations of the news’ consumers. For instance, a reader with a high cultural capital is likely to be more demanding on certain characteristics of news (i.e., in-depth news making, references…) and will discredit the media producing unverified or tabloid information that does not meet their expectations. This factor may also determine how other actors like academics, journalists, politicians, and legal experts appraise the quality of news.

In dealing with supply: Media competition, ownership, and business models are capable of determining which characteristics of news will be produced by the media. For instance, the business models⁸ choice implies particular market positioning, customer segments serving by the media, the customer relationships established and maintained with each customer segment, the revenue and pricing model, the cost structure and financing, and the partnership network (in-house production vs. outsourcing). All these elements may be determinant for the editorial choice—as in, the set of information features produced by the media to satisfy demand, expectations, and needs.

Finally, the perception of quality by different actors can be impacted by more general factors—i.e., the institutional framework, actions of civil society such as NGOs, (medi)activist groups, trade unions, and political parties—whose actions and decisions could influence some news features to the detriment of others (veracity against sensationalism) (Fig. 5).

A general mapping of media, according to horizontal and vertical axes of different news features, crossed with the factors underlying producers’ (e.g. economic models) and consumers’ (e.g. number of subscribers) editorial choices substantiate potential efficiency of various regulation tools. Distinguishing news according to their quality—horizontal or vertical—by using transparent and measurable criteria could permit to finely design and advice for regulation instruments (media education and prevention campaigns, taxation and subsidies, rewards, sanctions, and prohibitions…) to preserve and promote media pluralism.