Supranational and intergovernmental organisations
Supranational governmental organisations provide another rich source of data and analysis. These include the UN and its agencies such as UNESCO, supranational regulatory bodies like the ITU, intergovernmental authorities such as the European Union and SADEC, and political and human rights bodies like the Council of Europe. These organisations have varying forms of regulatory power but have in common a policy orientation so their relevant work on misinformation tends to combine data collection, research, legal-regulatory analysis, and policy proposals (Bayer 2019). For example, the EU s high-level expert group on fake news and online misinformation (European Commission 2018) made proposals concerning transparency, media literacy, empowering users and journalists, safeguarding European news, and promoting research.
Livingstone (2012, 421) draws on Kohn’s (1989) framework identifying four types of crossnational comparative research: ideographic, hypothesis-testing, system-sensitive, and transnational. The first three focus on the nation-state as unit; the fourth, transnational, treats countries as the locus for global trends. To date, comparative studies of misinformation fall mainly into ideographic and transnational categories. An example of the latter is Vosoughi et al.’s (2018, 1146) study finding false news stories on Twitter ‘diffused significantly farther, faster, deeper, and more broadly’ than true stories. There has been some development of hypothesis-testing but negligible system-sensitive, ‘media systems’ research to observe and explain ‘how and why nations vary systematically’ (Livingstone 2012, 419).
The vast majority of studies have been ideographic, single country studies, mostly of the US, providing resources for comparability and system analysis but not explicitly combining both. A recent study noted that outside the US, ‘we lack even the most basic information about the scale of the problem in almost every country’ Fletcher et al. 2018, 1). This is being remedied by an increasing number of cross-national studies, but most focus on discrete phenomena for the purposes of comparison. For instance, Fletcher et al. (2018, 7) examine false news websites in France and Italy, finding that ‘false news has more limited reach than is sometimes assumed, in line with US research’. Yet this short study does not compare the findings cross-country, much less relate these to system-level features.
A comparative study of fact-checking does incorporate hypothesis testing: namely, that high-source transparency will correspond to high journalistic professionalism and that countries with low trust in media will have fact-checkers who provide comparatively ‘higher levels of source transparency’. The study concludes (Humprecht 2019, 14):
Although online disinformation is a global phenomenon, practices of correction still seem to be shaped by national news cultures — in newsrooms as well as in independent organizations. Consequently, from a user’s perspective, the possibility of coming across transparent, professional fact-checking depends on the country in which one lives.
More commonly, misinformation is addressed in relation to discrete aspects of media performance, politics, or policies and usually with a focus on two of those three domains, although with recognition of their interconnectedness. Here, misinformation fits within research specialisms, such as news coverage, journalistic practices and cultures, and the relationship between media structures and content and the public affairs knowledge of citizens (Aalberg and Curran, 2012; Esser et al. 2012). More recent studies examine how the production and circulation of misinformation is shaped by the economic drivers of platformisation and the challenge and opportunities of monetisation across online publishing and social media. This includes studies of the commercial dynamics driving politically oriented misinformation, whereby ‘political disinformation succeeds because it follows the structural logic, benefits from the products, and perfects the strategies of the broader digital advertising market’ (Ghosh and Scott 2018). The digital advertising model has rewarded maximising clicks and undermined the value and resources for quality reporting while social media gateways have disaggregated news content from suppliers and retained advertising revenues. Misinformation has been aided by the growth of programmatic advertising, automating ad buying, and placement processes (Bakir and McStay 2018). In addition, political marketers have used behavioural advertising and micro-targeting, with studies (Bayer 2019, 33) showing
many UK political parties use third-party digital campaigning platforms (such as NationBuilder), which enables parties to match voters’ contact information with the data on Facebook and Twitter. . . [while] political micro-targeting was also reported in Italy, Germany and the Netherlands, although to a lesser degree.
While media systems research has closest affinities with studies of media institutions and production, misinformation analysis includes rich resources comparing users’ communication activities. WhatsApp is used more widely across Southern than Northern countries: with nearly 60 percent in Brazil in discussion groups with strangers and nearly a fifth discussing politics, compared to 2 percent in the UK (Newman et al. 2019, 2020). More broadly, studies of misinformation examine digital materialities, socio-cultural and psycho-cognitive aspects that are expanding areas of comparative communication research but barely incorporated into ‘media systems’ literature. Disinformation is often intentionally affective and thrives on the generation of feelings of superiority, fear, and anger (Wardle and Derakhshan 2017; Bakir and McStay 2018). Those are mobilised by a mix of loosely affiliated or tightly structured networks, such as across the alt-right (Lewis 2018) to sustain racism and racial violence, misogyny, and other forms of hate speech. This makes studies of gender, sexuality, race, and the communicative resources used by or against all minorities of vital importance. Again, that is reflected across comparative communication research (Esser and Hanitzsch 2012) but not yet integrated into ‘system sensitive’ media systems analysis.
Greater advance has been made examining the relationship between misinformation flows and key variables concerning the political system; patterns of media and information availability; levels of trust across politics, media and sources of authority; and user profiles, including demographics, access, education levels, literacy, and usage patterns across legacy media, social media, and search. For instance, studies suggest that ideologically motivated disinformation is more prevalent in polarised societies with low trust in media (Humprecht 2019). Audience research conducted across the US, the UK, Spain, and Finland found low trust in media was common but varied (Nielsen and Graves 2017, 6):
[L]ess than half of online news users in the US and the UK think you can trust most of the news most of the time. While the figures are higher in Spain and in Finland, very large parts of the population still have limited trust in news in general, and — strikingly — only somewhat higher trust in the news that they themselves use.
Trust illustrates the significance of another key resource. Commercial sector data analysis or large-scale surveys usually exceed the capacity available through university-funded research. The Edelman (2020) Trust Barometer has charted falling trust in media worldwide, although with differences between countries and among media vehicles and platforms and class position, finding a widening gap in trust levels between the more affluent and general populations. This is also an example of integration, with Edelman’s research widely cited by academics. In the US, overall positive trust in media was only around half (53 percent) in 1997, yet had fallen to less than a third (16 percent) in 2016 (Newman and Fletcher 2017, 7). In research conducted by the Reuters Institute for the Study of Journalism, a corporate-sponsored academic research centre, Denmark, Germany, Ireland, and Spain were at the higher end of trust in media, and the US, France, and Greece had comparatively low trust scores, with Australia and the UK in between. Newman and Fletcher (2017, 26) conclude, ‘[t|he reasons for lack of trust in the media are remarkably consistent across countries. Bias and agendas are rife and are perceived to have worsened with the advent of the internet’.
Trust and confidence in journalistic processes are associated with journalism that supports claims with evidence, conducts and highlights fact-checking, and is transparent about sources. There is evidence of growing distrust in the commercial bias in news. Newman and Fletcher (2017, 24) report that ‘[ajnother reason for mistrust identified by our respondents is the belief that media companies are distorting or exaggerating the news to make money’.
The reshaping of news for commercial agendas, including brand-sponsored content, native advertising, content recommendation, and clickbait, is being examined, yet there is a need for integrative studies encompassing sources of misinformation across political and commercial speech, one of many key tasks for future research.