Causes and consequences of mainstream media dissemination of fake news: literature review and synthesis

ABSTRACT

Research indicates that the reach of fake news websites is limited to small parts of the population. On the other hand, data demonstrate that large proportions of the public know about notable fake news stories and believe them. These findings imply the possibility that most people hear about fake news stories not from fake news websites but through their coverage in mainstream news outlets. Thus far, only limited attention has been directed to the role of mainstream media in the dissemination of disinformation. To remedy this, this article synthesizes the literature pertaining to understand the role mainstream media play in the dissemination of fake news, the reasons for such coverage and its influences on the audience.

KEYWORDS:

• Fake news
• disinformation
• fact checking

1. Introduction

While disinformation has circulated through media since the early days of mass communication, scholars and pundits have argued that recent years mark ‘the rise of the misinformation society’ (Pickard, 2016, p. 119) and the era of ‘alternative facts’ and ‘post truth’ (Benkler et al., 2018; Madrigal, 2017). As Higgins (2016, p. 9) explains, the term ‘post-truth’ describes not only an increase in the frequency of lies in the public sphere, but refers to a world in which truth is no longer an expectation. The rise in use of disinformation (defined as the intentional and purposive spread of misleading information, which is different from misinformation which relates to ‘unintentional behaviors that inadvertently mislead’; Chadwick & Vaccari, 2019, p. 14; see also Benkler et al., 2018, p. 24) spans across the globe from Europe and the U.S. to Brazil, Nepal, and Russia (to name a few examples) and across contexts from health issues (such as the safety of the childhood vaccinations) through environmental problems (such as global warming) to consumer topics (such as the effects of Listerine mouthwash on the severity of colds) and, of course, to diverse political and social debates.

While political actors have probably always exaggerated, misled or at times even lied, it seems both the frequency of lies and their centrality in the strategy of some political actors have increased in recent years. Politicians’ opportunities to disseminate disinformation directly to the public, bypassing the media's gatekeeping and their editorial scrutiny, have increased with the rise of social media and the possibilities they afford and with a weakening of what Graves and Wells (2019, p. 42) call ‘factual accountability.’ Another cause of disinformation lies with the rise of ‘the fake news genre’ (Egelhofer & Lecheler, 2019), referring to the intentional spread of false information, masked as traditional news, to advance political goals or generate ad revenues (Bennett & Livingston, 2018; Guess et al., 2019; Pickard, 2016). Two well-known examples are a complex of fake news websites run by teenagers from a small town in Macedonia and a U.S. company called Disinfomedia, owning many sites disguised as serious journalism (including USAToday.com.co and WashingtonPost.com.co). Both operations spread pro-Trump and anti-Clinton fake news stories prior to the 2016 U.S. elections (Allcott & Gentzkow, 2017, p. 217), as did many so-called alternative news websites in the right-wing media ecosystem (Benkler et al., 2018). As opposed to disinformation by political actors, a main characteristic of the fake news genre is that the original producers of the information are harder to trace given that they purposively hide their true identity when presenting the information as originating from a legitimate news outlet. The phenomenon is not restricted to the U.S.: similar phenomena can be found across the world, ranging from Latin America (Tedeneke, 2018) to Israel (Pfeffer, 2019), France (Farand, 2017) and Indonesia (Lamb, 2019).

While the phenomena of political actors actively spreading misleading information are well-known and discussed, the role of traditional news media, who are supposed to be the bearers of truth and factual accuracy, is less well understood. In this article, we argue that traditional news media are in fact a part of the problem, and play a somewhat paradoxical role with respect to fake news and its dissemination. On the one hand, the journalistic community's reaction to the rise of disinformation seems to be a renewed emphasis on truth and facts, with some journalistic brands around the world being more careful with facts than ever (Glasser, 2016). This response is reflected in a sharp proliferation of fact-checking (Graves, 2016), which has increased by more than 900% since 2001 in newspapers and by more than 2,000% in broadcast media (Amazeen, 2013). On the other hand, in order to correct disinformation, news media have to repeat it and repeating lies often makes it more difficult to correct them (Lewandowsky et al., 2012). In addition, few other institutions have the reach of news media (Waisbord, 2018). Research also indicates that the reach of fake news sites (Fletcher et al., 2018) and the consumption of fake news stories online (Grinberg et al., 2019; Guess et al., 2019) are limited to quite small parts of the population. Taken together, this implies that many people may hear about fake news stories through mainstream news media. Thus, the paradoxical situation arises that mainstream media covering disinformation help in its dissemination – even though their purpose is to correct disinformation.

Thus far, only limited scholarly attention has however been directed towards this paradoxical role of mainstream news media in the dissemination of disinformation, with little reflection on the causes for and the consequences of mainstream news media disseminating fake news. The effects of coverage of fake news in mainstream news media on news media audiences, including the possibility that parts of the audience learn the wrong information despite the fact that mainstream journalists cover it as ‘fake,’ have also been largely ignored by scholars. To remedy this, the purpose of this article is to review and synthesize the literature pertaining to three key questions: (a) What role do mainstream news media play in the dissemination of fake news? (b) Why do mainstream news media cover fake news? And (c) What are the potential influences of mainstream news media coverage of fake news on their audiences? Answering these questions will yield crucial knowledge not only for scholars working on disinformation, misinformation, misperceptions, knowledge resistance and related areas, but also for journalists covering misinformation and disinformation and providing fact-checking. Finally, media audiences might also need education about the paradoxical role of journalists to not fall into the trap of learning wrong information from the coverage of fake news by mainstream news media.

Before proceeding, some words of caution. Since fake news has become such a buzzword and often is used in partly different ways, it is important to define it clearly. Hence, in this article, by fake news we refer to ‘the deliberate creation of pseudojournalistic disinformation’ (Egelhofer & Lecheler, 2019), that is, to the use of journalistic rhetoric, formats and reporting styles for the intentional dissemination of false, invented, information.

2. Mainstream media as principle disseminators of fake news

The first question to address is what role mainstream news media play in the dissemination of fake news. Although there is no empirical research directly addressing this question, it is evident that fake news and disinformation have been prominent topics in public and academic debate in relation to the past U.S. presidential election. Survey data demonstrate that during the 2016 election and its aftermath, large proportions of the public believed different fake news stories. For example, more than a third of American adults believed that the Pizzagate story––connecting the Clinton campaign to a pedophile ring – was ‘probably’ or ‘definitely’ true. Only 29% of respondents were sure the allegation was ‘definitely not true’ (Frankovic, 2016). As another example, Silverman and Singer-Vine (2016) report on a study that presented respondents with three false ‘fake news’ headlines and three true headlines and asked them to assess the accuracy of those headlines they had seen during the campaign. 75% of the fake news headlines recalled by respondents were perceived as ‘somewhat’ or ‘very’ accurate. The parallel figure for real news headlines was 83%. Research also suggests that during the weeks leading to that same election, the most popular ‘fake news’ stories enjoyed more audience attention compared to the most popular ‘real’ mainstream news stories on Facebook (Silverman, 2016). As yet another example, close to 20% of respondents in a study by Allcott and Gentzkow (2017) recalled seeing the leading fake news stories of the 2016 campaign reported or discussed before the elections. The corresponding rate in a study commissioned by BuzzFeed was nearly 33% (compared to 57% reporting they had seen a real news headline; Silverman & Singer-Vine, 2016). While high rates of respondents recalled seeing ‘placebo’ fake news stories as well, Allcott and Gentzkow (2017, p. 227) estimate that the average American adult saw and remembered at least 1.14 stories from their data set of 156 election-related fake news articles circulated in the three months before the 2016 U.S. presidential elections, using lists from independent fact-checkers (Snopes, PolitiFact and BuzzFeed).

Despite the scarcity and lower quality of data, it is reasonable to assume that people are exposed to and believe fake news also in other contexts, not only in the U.S. For example, one study shows that 64% of a sample of protestors in the 2015 anti-government protests in Brazil believed the Workers’ Party (PT) wanted to create a communist regime in the country. 54% also thought that a drug gang is the party's armed wing and 43% that the PT brought 50,000 illegal immigrants from Haiti to vote in the elections. These stories were all verifiably false news stories that were shared on social media (Arnaudo, 2017). As another example, in the context of the April 2019 Israeli elections, 20.5% of adult respondents in the online module of the Israeli National Election Study were ‘absolutely sure’ or ‘pretty sure’ that the wife of opposition prime ministerial candidate Benny Gantz was active in ‘checkpoint watch’ (a group many mainstream Israelis consider pro-Palestinian). Only 22% of respondents were sure that the story was not true. This story was called by observers the most widely circulated online fake news story of the campaign. In other contexts, were no data is available on public belief in fake news, we can still deduce from reports on real world consequences that fake news were believed. For example, false claims on social media have reportedly led to a woman being beaten to death in Brazil and to incitement to violence in Nigeria and Nepal (IFCN, 2016). In sum, evidence documents that a number of notable fake news stories are recalled and believed by parts of the populations in these different countries.

However, convincing evidence also shows that, in social media, exposure to fake news is heavily concentrated: For example, an analysis of Twitter activity in the context of the 2016 U.S. election shows that 80% of fake news stories were consumed by merely 1% of the population (Grinberg et al., 2019). Another study focusing on online browsing activity showed that almost 60% of the visits to fake news websites came from the 10% of the population with the most conservative online information diets (Guess et al., 2018). The authors estimated that, during the last weeks of the campaign, only 2.6% of all the visits to current affairs articles were to fake news websites (see also Guess et al., 2020 for similar results). Using trafic data from the U.S., collected during the 2016 election year, Nelson and Taneja (2018, p. 3727) demonstrated that the number of monthly visitors to an average real news site was more than 40 times larger than the number of monthly visitors to an average fake news site, and that the average visiting time to a fake news site was much shorter than the average for a real news site. Fake news websites are not only visited by a small segment of the population; they are also rarely shared. In a study linking survey data with respondents’ Facebook profiles, Guess et al. (2019) found that sharing fake news was a highly infrequent activity. Only 8.5% of respondents for whom the authors had matched Facebook data shared fake news on their feeds. In Grinberg et al's (2019) study, only 0.1% of Twitter users¹ were responsible for 80% of the sharing of information from fake news sources.

Similar findings emerge in a study that tracked 300 Italian and French websites that were identified by independent fact checkers as publishers of false news (Fletcher et al., 2018). None of these websites had an average monthly reach of over 3.5% in 2017, with most reaching less than 1% of the online population, and the total time spent on fake news websites was dramatically lower than the time spent on popular mainstream news websites (Le Monde in France and La Repubblica in Italy). A recent study conducted in Singapore (Tandoc et al., 2019) similarly suggests limited diffusion of fake news on social media. The above strongly suggests that social media are NOT the primary driver of spreading fake news among the general population.

Part of the research literature about disinformation and fake news stresses notions of selective exposure and attention, arguing that fake news stories are disseminated mainly in ideological echo chambers or cyber-ghettoes (Lewandowsky et al., 2012, p. 111). It is true that, on the supply side, people ‘selectively share’ news stories, including fake news stories, that align with their political views (Guess et al., 2019; Grinberg et al., 2019). Given the ideological homophily typical of social media, that is, that people tend to view mostly congenial content on their social media feeds (Bakshy et al., 2015), this is probably true with regards to fake news as well. The end result is that social media users with ideologically clear allegiances will be more likely to see ideologically-congruent fake news stories on their feeds. On the demand side, these users are furthermore going to be more likely to select and read these stories (Bakshy et al., 2015; Grinberg et al., 2019). Also congruent with the argument about online echo chambers, findings show that fact-checking information travels selectively online (Hameleers & van der Meer, 2019; Shin & Thorson, 2017) and, perhaps as a result of this process, Guess et al. (2019) estimate that fact-checkers’ corrections almost never reach fake news-websites consumers. It should be noted, however, that other research points out that echo-chambers may not be as prevalent and omnipresent as some have worried (Bruns, 2019; Zuiderveen Borgesius et al., 2016).

In either case, social media only account for a part of the dissemination of fake news. As mentioned above, studies in both the U.S. and Europe have estimated that the vast majority of direct exposure to the original publications on fake news websites is heavily concentrated and limited to between 1% to 10% of the population (Fletcher et al., 2018; Grinberg et al., 2019; Guess et al., 2018). Despite this limited exposure, data repeatedly demonstrate that at least some fake news stories receive widespread attention and are rather widely believed. Although the evidence is circumstantial, the fact that people hear about fake news stories, but do not see their original publication, implies that mainstream media are responsible for much of the public attention fake news stories receive. In other words, while estimates of exposure to fake news stories on mainstream media are not available, it is probable that, given the vast attention dedicated to fake news stories on mainstream outlets (Al-Rawi, 2018), larger proportions of the public are exposed to the more visible and newsworthy fake news stories through the coverage that these stories garner in mainstream news media than directly through social media.

Certainly not the only conduit of fake news and other disinformation, mainstream media are thus probably a significant amplifier and disseminator of false stories – even if they, for the most part, cover fake news with an intent to set the record straight and correct the fabricated information. To illustrate, a LexisNexis search in 12 American mainstream outlets² found 2,787 hits for ‘fake news’ in 2016. The term ‘Pizzagate,’ as an illustrative example, was covered 34 times in these mainstream outlets even before the December 4, 2016 shooting at the Comet Ping Pong pizzeria in Washington DC (which of course resulted in hundreds of hits). Major news organizations, such as the B.B.C., have even set up teams or special operations with the explicit intent to report and combat fake news (Jackson, 2017). A network analysis of hyperlinks to and from 117 fake news websites found that in some cases, news media organizations even send links to the fake news sites when fact checking them (Albright, 2016). Vargo et al. (2017) also offer evidence pointing out that fake news websites have an influence over the mainstream news agenda for some of the topics (such as international news). According to their study, partisan media were more strongly and consistently influenced by the agenda of fake news websites, and the authors suggested the possibility that fake news influence mainstream news through its coverage in partisan media (p. 2043).

3. Fake news as news for mainstream media

If the answer to the first research question is that mainstream news media seem to play a significant role in the dissemination of fake news, the second research question asks why news media cover fake news. After all, ‘Journalism is expected to report, above all things, the truth’ (Tandoc et al., 2018, p. 140). Thus, Tandoc et al. (2018) see fake news as an oxymoron. Even so, going through the literature we find at least four important reasons for why mainstream news media cover fake news. These are related to (a) journalists’ role perceptions, (b) traditional news values, (c) the psychology of news decisions, and (d) the infrastructure for covering what’s happening in the online world.

3.1. Journalists’ role perceptions and reporting about fake news

One important reason for why mainstream media report about fake news is anchored in the very basic tenants of their profession. Part of journalists’ role perceptions contains the need to seek the truth and to expose what is not true (Donsbach, 2004). This role perception drives a preoccupation with authenticity and the urge to uncover backstage developments (Jamieson & Waldman, 2004, Ch. 2) as well as an aspiration to develop ‘bullshit detection’ senses (Harcup, 2014, pp. 272–273). At the core of these professional commitments and perceptions lays a belief that a functioning democratic society depends on quality information. Providing citizens with the information they need to make political decisions is among the most highly ranked professional values in surveys of journalists across political contexts (Hanitzsch et al., 2012). Therefore, journalists perceive that an essential part of their professional duty is to correct manipulation and disinformation. Verification and correction are portrayed in the literature as the essence of journalists’ professionalism (Rosenstiel & Kovach, 2001). Stories about ‘fake news’ that include correction of deception thus enable journalists to fulfill a central element of their profession.

The obligation to expose the truth and to verify information has been a cornerstone of journalists’ professional principles since the early days of profession (Godler & Reich, 2017), but its importance has grown even stronger over the past decades, as part of the rise of a professional culture that emphasizes analysis, contextualization and interpretation, in addition to simply recording and reporting on news events (Esser & Umbricht, 2014; Salgado & Strömbäck, 2012). As Graves (2016, p. 63) explains, ‘The fact-checking movement reflects and reproduces the professional culture tied to analytical journalism.’ Fact-checks are stories that, at least in some respects, deviate from journalists’ classic detached observer role. The classic journalistic mission was describing what ‘he’ and ‘she’ said. According to Graves (2016), fact-checking is instead all about deciding whether he or she are right or wrong. This goes beyond merely describing the world, as it entails interpretation and elaboration, selection of information sources that represent ‘the truth,’ against which the new information should be cross-examined, and a true/false verdict that necessitates journalists to give up a constructivist point of view that offers the option of ‘objectively’ presenting to audience competing notions of truths.

The prominence of truth and of correcting false information in journalistic professional culture has increased in recent years as arguments about ‘alternative facts’ are mounting under the ‘information disorder of the post-truth era’ (Bennett & Livingston, 2018; Graves & Wells, 2019; Pickard, 2016), and given a more complex set of issues covered by journalists, a social climate of polarization and controversy, even about scientific facts, and the ‘increasingly manipulable character of the media’ in an environment that constantly challenges journalists’ ability to discern truth and correct lies (Reich & Barnoy, 2019). The prominence of truth and facts in the professional culture of journalists is reflected in journalists’ self-criticism on the role of the news media in the disinformation order (see Patterson, 2013, e.g. p. 6). The fact that journalists complain that reporting has become ‘increasingly sloppy’ and that ‘bottom-line pressure is hurting journalism’ (ibid.) highlights the professional aspiration for un-sloppy, truthful and accurate reporting.

The journalistic commitment to correct wrong information and to help the public learn the truth is well expressed in journalistic discourse about fake news, that treats the journalistic reaction to fake news as a ‘combat’ and reiterates that journalists should not stand aside and should ‘weigh in on the battle over lies, distortions and exaggerations’ (in the words of B.B.C. news chief James Harding; cited by Jackson, 2017). This discourse views old-school journalism practices of fact-checking and verification as the main strategy in this battle (O’Carroll, 2019).

It is noteworthy, however, that this practice stands in some contrast to the journalistic principle of not publishing rumors, which is also perceived by journalists as an important value (e.g. Tsfati & Meyers, 2012, pp. 451–452). That is, reporting about fake news entails publishing an online rumor and explaining why it is not true, whereas, at least by the strict codes of the profession, news does not include unsubstantiated rumors. One reason why news media still report about fake news is likely that journalists feel compelled to ‘clean’ the public sphere from fake news given that they perceive that the false information has already been widely distributed. This stands in contrast to ‘regular’ rumors that they are not supposed to correct (according to some professional codebooks), given the relatively small circle of people that have been exposed to the rumor, and the likelihood that attempting to correct these rumors would only result in further disseminating them.

In addition, journalistic reports on fake news not only enable journalists to fulfill the very basic imperatives of their profession – exposing and correcting lies. The emphasis journalists put on truth and truthful reporting when they expose and correct fake news also enables them to differentiate between their work (which they would call ‘real news’) and fake news. Gieryn's (1983) idea of ‘boundary work’ refers to ‘a field's attribution of certain characteristics to itself to create a social boundary between it and adjacent fields’ (Coddington, 2012, p. 380) in an attempt to construct a favorable public image for the field by contrasting it with what it is not (see Carlson, 2015). Similarly to journalistic coverage of WikiLeaks (Coddington, 2012) and blogs (Carlson, 2007) and to journalistic discourse about entertainment (Winch, 1997), the coverage of fake news by established media enable mainstream journalists to preserve their journalistic authority using dichotomous distinctions between ‘honest’ and ‘deceitful,’ ‘ethical’ and ‘unethical’ and ‘works for the public’ or ‘against it.’

3.2. The news value of fake news

A second reason why mainstream news media report about fake news is because they meet news criteria and values that stem from journalistic role perceptions and dominate journalistic cultures across geographic and cultural contexts (despite their constantly evolving nature; Harcup & O’neill, 2017). We know from studies that, at least in English speaking countries, the targets of fake news stories tend to be political actors (Humprecht, 2018) and the content of the false information tends to be counter-intuitive, negative and emotional (Bakir & McStay, 2018). This meets Harcup and O’neill's (2017) and Galtung and Ruge's (1965) news criteria of ‘negativity’ and ‘elite people.’ The ‘relevance’ criterion is also met, especially when the object of the fake news item is a prominent political figure or the fake news could otherwise have an impact on public opinion or the results of an upcoming elections or referendum. While the fact that someone says something and someone else refutes the same claim does not meet Galtung and Ruge's (1965) criterion of ‘unambiguity,’ the inherent ambiguity in news stories about fake news reports is related to a conflict about whether the information is true or false, and this conflict is another aspect that makes stories about fake news newsworthy. Harcup and O'Neil's ‘magnitude’ criterion is also satisfied, given the potentially large number of people who are assumed to already have been exposed to the fake news story when mainstream news media report it. Given that the topics of fake news stories in different countries reflect variations in the local news agendas (Humprecht, 2018), it is possible to assume that the ‘continuity’ criterion is also met, that is, that the false stories echo and complement other factual stories in the news. So, overall, stories about fake news tick many of the boxes of newsworthiness.

On top of the fact that stories about fake news reports are almost by definition newsworthy, the content of fake news reports often satisfy additional news criteria that further increase their news value. To demonstrate this one could examine the stories in a dataset collected by Allcott and Gentzkow (2017), containing 2016 election related fake news complied from lists from three independent websites (Snopes, PolitiFact, and Buzzfeed). These stories very often meet the ‘surprising’ (‘Hillary Sold Weapons to ISIS’; ‘Pope Francis Endorses Donald Trump’), and ‘negative’ (‘FBI Agent Suspected in Hillary Email Leaks Found Dead in Apparent Murder-Suicide’) news criteria. In addition, many of these fake news stories carry what Harcup and O’neill (2017, p. 1471) call ‘entertainment value’ relating to stories concerning sex (‘The FBI Uncovered Evidence of a Pedophile Sex Ring Run Under the Guise of the Clinton Foundation’; ‘WikiLeaks Releases Candid Photos of Hillary Clinton Grabbing a Man's Crotch’), showbusiness (‘Clinton Campaign Paid Beyonce and Jay Z $62 Million for Cleveland Concert to Secure Black Votes’), human interest (‘Physician Confirms Hillary Clinton Has Parkinson's Disease’), an unfolding drama (‘President Obama Confirms He Will Refuse to Leave Office If Trump Is Elected’), or stories offering opportunities for humorous treatment (‘Trump Threatens to Skip Remaining Debates If Hillary Is There’). In addition, many of them dealt with celebrities (‘Denzel Washington Switches to Trump’; ‘Trump to Deport Lin-Manuel Miranda’; ‘Kurt Cobain Predicted Donald Trump Presidency in 1993’), another news value mentioned by Harcup and O'Neil.

The fact that fake news stories meet news criteria is not surprising. Economists make a connection between lower entry barriers to the news media industry (e.g. it is easy and cheap to set up a website, and to mobilize content through advertising platforms) and the spread of fake news. Whereas in the past publishing false stories would entail a risk for future revenues of media outlets, currently, even false news articles that go viral on social media can draw significant revenues when users click to the original site (Allcott & Gentzkow, 2017, p. 217; see also Bakir & McStay, 2018) These clicks are easily translated to revenues, using web services (such as Google AdSense) that match between content producers and targeted advertizers around the world. Given the economy of fake news, it is possible to assume that the creators of fake news stories create them in a way that would maximize their spread. That is, fake news stories can be assumed to be produced with their news values in mind, which further helps explain the fact they meet news values, as discussed above. The fact that the content of fake news stories differs across national contexts in ways that matches and reflects national journalistic styles and news agendas (Humprecht, 2018) also implies that creators of fake news intuitively or deliberately write their stories in ways that match mainstream journalists news values.

3.3. The psychology of news reporting about fake news

A third key reason why fake news is reported by mainstream news media has to do with the psychology of news decisions. When examining the psychological factors that shape news decisions, Donsbach (2004) argues that, beyond seeking the truth, a major factor that shapes journalists’ decision is social validation, that is, that journalists are extremely attentive to what other journalists are saying and doing when making their own news decisions. This process takes place in part through interaction with peers and colleagues and in part by observing colleagues’ journalistic outputs and internalizing common assumptions, norms, practices and ‘news ideologies’ (pp. 141–143). In that context, it is important to note that the topic of ‘fake news’ has received ample attention in mainstream media with 1,416 news items referencing the term in 23 U.S. and British mainstream newspapers between 2010 and 2017, mostly referring to social media's negative role in spreading disinformation (Al-Rawi, 2018). The ample attention dedicated by mainstream news outlets to fake news and the salience of the fact-checking concept in the coverage of fake news by these news outlets (p. 6) signal to journalists that fake news is perceived by the journalistic community as an important social problem, one that should be addressed by correction.

A second psychological factor influencing journalistic role perceptions and decisions, beyond social validation, is the power of journalists’ existing attitudes and predispositions (Donsbach, 2004, pp. 146–150). According to Donsbach (2004), as all human beings, journalists pay more attention to attitude-confirming information than to disconfirming information. When processing information, journalists too perceive attitude-confirming information as more important than attitude-disconfirming information (Kepplinger et al., 1991). Hence, their news decisions tend to be influenced by their own subjective beliefs (Patterson & Donsbach, 1996), although professional training and the fact that journalists are nested within news organizations and norms emphasizing the importance of not letting subjective beliefs influence the news acts as a counterforce (Shoemaker & Reese, 2014). At least this holds true for mainstream news media. This helps explain findings demonstrating that partisan media – where the norms are quite different from mainstream news media – are influenced more heavily by fake news than mainstream media (Benkler et al., 2018; Vargo et al., 2017). Both systematic and anecdotal evidence also demonstrates how fake news receive the attention of partisan media when it fits the ideological tendencies of the partisan outlets (Benkler et al., 2018). Bennett and Livingston's (2018) example of Fox News reports about immigrants’ crime in Sweden illustrates this point. All in all, some journalists cover fake news not only because it is newsworthy, fits their role perception, and because other journalists cover it, but in some cases also because the content of fake news fits their ideological tendencies.

3.4. The infrastructure for reporting about fake news stories

The fourth and final reason why mainstream news media report about fake news is that media already have an infrastructure to locate and report about viral online phenomena. To become news, a story must not only meet the criteria of newsworthiness and fit into journalists’ role perception. A story must first of all be noticed by reporters and editors. From that perspective, it is important that the online arena – not least the realm of social media, the originating platform for the dissemination of fake news stories – have been constantly and systematically monitored by news organizations over the past 15 years (Boczkowski, 2010). Reporters have started to regularly follow what is going on in social and online media in areas that are part of their beats, and perceive this as an important part of their jobs (Jordaan, 2013). Some news organizations have even developed special beats focused on following what is happening on social media (Broersma & Graham, 2012), while others have hired or trained staff and experts to monitor the online world (Schifferes et al., 2014). The methods used by these staff and experts in order to spot stories vary between simply trying to follow what's going on, using geo-location tags in order to drill down information about specific newsworthy events, searching for specific keywords, and using sophisticated specialist software to try to spot stories (Schifferes et al., 2014, p. 409). In other words, routines for covering the online world and to use social media as sources for mainstream news have been developed (Lecheler & Kruikemeier, 2016).

Various content analyses also show that news media increasingly and routinely cover social media (Paulussen & Harder, 2014). Paulussen and Harder (2014), for example, report an increase in the use of Facebook, Twitter and YouTube as sources by two elite Flemish newspapers. In their study, social media were used as a news source in 70% of the news reports referring to either of these platforms. However, only a minority of these stories were based on social media only, and the authors conclude that social media have not yet become a major primary news source for Flemish print journalists (p. 546). Other researchers have documented the flow of attention from social media to mainstream media in the context of the vaccine-autism controversy. Jang et al. (2019), using time-series analysis, show that Twitter drives news media coverage of the controversy and not the other way around. The authors rule out the possibility that it is tweets that merely link to news coverage that underlie the association. While the study does not document the flow of specific items, the fact that mainstream news follow Twitter debates on this topic is meaningful given journalists’ attempts to convey journalistic balance. This may lead to inaccurate depictions of scientific evidence that overwhelmingly rules out a vaccine-autism connection (p. 111).

Thus, given the infrastructure developed by mainstream news organizations for monitoring the online world, it is possible to assume that when fake stories spread on social media and receive some traction, they are very likely to be discerned by journalists, who will examine their veracity and report them, using procedures they have developed (Schifferes et al., 2014). Covering fake news does not necessarily entail a complex operation for news desks. According to Graves (2016), over the past two decades, independent fact-checkers have developed at the intersection of academic, nonprofit and journalistic sectors. These supplement smaller scale ad-financed initiatives, such as Snopes.com (pp. 28–29). These fact-checking outlets act as ‘information subsidies’ (Gandy, 1980) for news organizations who cover the intersection of the online and political world. The implication is that news organizations do not have to spend much energy on locating newsworthy fake news stories and examining their veracity; at least in some countries, they can rely on credible and transparent fact-checkers, follow their leads and craft their reports based on their examinations.

4. Implications: audience processing of news coverage of fake news

Thus far, our analysis suggests that an important part of the dissemination of fake news takes place through mainstream news media, and that journalistic role perceptions, news values, social validation, and the fact that news media institutions have the infrastructure both for the detection and for the correction of fake news stories are reasons why mainstream news media pay attention to fake news. In this section we will address the third research question, namely, what are the potential influences of mainstream news coverage of fake news on their audiences?

Important to remember is that for the most part, when mainstream media cover fake news stories, they do so with an explicit intention to correct the false information. Are journalists successful in their attempts to set the record straight? Is it possible that parts of the audience retain the wrong information despite the fact that news media had covered it as fake news? While no empirical research to date has examined precisely how people respond to media reports about fake news, quite a bit of research in psychology and adjacent areas has addressed the question of misinformation correction. While this research relates to misinformation in general (without distinguishing between misinformation and disinformation³) and the current article relates to a specific type of disinformation, we argue that it is possible to import the findings to the context of mainstream news correction of fake news. First, because the studies we cite below were mostly experimental and used invented scenarios and (dis)information. Second, because in the psychological studies on misinformation correction, participants were typically not presented with information regarding the motivation of the disseminator of the information, not knowing whether these disseminators knowingly and intentionally deceived. Similarly, information about the motivation of the disseminator or intentionality is not always available to audiences reading news media reports about fake news outside the lab.

While findings on the effects of the correction of misinformation are not fully consistent (Walter & Murphy, 2018), social psychological research has demonstrated time and again that retractions often fail to completely eliminate the influence of misinformation (Lewandowsky et al., 2012, p. 112). For example, a recent meta-analysis aggregating 32 studies (Walter & Tukachinsky, 2019) found evidence for the continued influence of misinformation, meaning that mis- or disinformation continued to shape people’s beliefs in the face of correction (r = -.05, p = .045).The findings suggest that after people are exposed to misinformation, corrective messages cannot fully revert people’s beliefs to baseline.

Theoretically, the empirical observation that exposure to correction does not always result in correct attitudes is argued to be the result of different information processing mechanisms. In order to comprehend a statement, people must at least temporarily accept it as true (Gilbert et al., 1993). From this perspective, believing even false information is part of processing it. The cognitive explanation for this argument states that when we encounter a report that tries to correct wrong information, for example that ‘Hillary Clinton is related to a pedophile ring’ is a wrong report, we first create a mental model that connects the nodes for ‘Hillary Clinton’ and ‘pedophile,’ and this mental model persists, especially in the absence of a complete and coherent alternative mental model. If the fake news item is reported with an explanation of why the fake news might be true or if the disinformation is consistent with an explanation that is already stored in other mental models (e.g. connecting, ‘Hillary Clinton’ with ‘crooked’), the psychological literature on misinformation suggests that eliminating the mental model will be particularly difficult (Anderson et al., 1980).

An extension of this cognitive explanation for the resilience of misinformation to correction is related to retrieval failures due to negation (Lewandowsky et al., 2012, p. 115). According to this explanation, ‘people encode negative memories by creating a positive memory with a negative tag’ (Walter & Tukachinsky, 2019, p. 8). That is, the statement ‘Hillary Clinton is not connected to pedophile ring’ is encoded as ‘Hillary Clinton, pedophile – not.’ The negation tag on the connection (‘not’) can however be forgotten or otherwise lost, especially for audiences suffering from impaired strategic memory or in situations of high cognitive load (Lewandowsky et al., 2012, p. 115). The ramification is that, for example, old people consuming news reports about fake news may retain the false information in spite of its negation (Wilson & Park, 2008), given the higher likelihood of suffering from strategic memory impairments in old age. In addition, if negations are lost in the retrieval process and thus confirm the misinformation, it stands to reason that corrective messages will be more effective if they attempt to debunk misinformation without explicitly negating it.

One advantage of news reports about fake news stories over the more thoroughly studied ‘fact checks’ or ‘adwatches’ is however that for the most part, there is no time lag between the delivery of misinformation and its correction. This is important, as the literature points out that corrections are less effective when time passes between exposure to misinformation and its negation (Walter & Tukachinsky, 2019, p. 15). With some notable exceptions, in which media organizations mistakenly fall prey to the fake news and report it as real news, the false ‘fake news’ information is reported by mainstream media together with a retraction. This explains why only a minority of respondents, around 30% in the U.S. (Frankovic, 2016; regarding Pizzagate) or 20% in Israel (regarding the wife of opposition leader Benny Gantz being a member of a pro-Palestinian group) believed widely reported fake news stories. However, it is also important to mention that only 29% in the U.S., and 22% in Israel, were sure that these are fake news stories. The modal audience response, in both cases, was thus uncertainty. This suggests that despite media refutations, sizeable shares of the audience deduce that there is a chance that the ‘fake’ information might be right. Thus, doubt (instead of an outright rejection) may be the undesirable consequence of mainstream news coverage of fake news.

Despite the advantage of simultaneously receiving the false information with its correction, a major problem with news coverage of fake news is that, in order to report about fake news stories, mainstream journalists have to repeat the false information. This is problematic, as repetition is known to be a major problem in attempts to correct disinformation (Lewandowsky et al., 2012; Walter & Tukachinsky, 2019). This can be explained by the ‘mere exposure’-effect (Zajonc, 2001) and the ‘truth effect’ (Dechene et al., 2010), according to which mere exposure and repetition of statements increases the likelihood that statements are perceived as true. One key explanation is that repetition breeds familiarity and people tend to perceive familiar information as correct and trustworthy, given the sense of ease and processing fluency that accompanies familiar information (Schwarz et al., 2007; Dechene et al., 2010). As succinctly argued by Schwarz et al. (2016), ‘when thoughts flow smoothly, people nod along’ (p. 85). Repeating false information, even as part of a retraction or a correction, enhances its familiarity, and thus retractions can backfire. Remarkably, studies have also found that people infer the accuracy and consensus of an opinion from the number of times it has been repeated, even when the repeated expression is associated with only one person (Weaver et al., 2007; see also Dechene et al., 2010).

In addition to increasing the familiarity of wrong information, there are other reasons to believe that some audiences will retain the disinformation despite the fact that news media report about it as ‘fake news.’ Research on the psychology of truth assessment has for example found that people tend to believe not only familiar, but also simple and coherent statements (Lewandowsky et al., 2012). Media refutations of fake news stories may thus be less effective when the fake news story is clear and coherent, and the refutation complicated and detailed. In line with this logic, Walter et al. (2019) found that lexical complexity (calculated using such indicators as noun and verb variation) of fact-checking messages was negatively associated with correction of misinformation. Simply put, the more complex the correction, the less fluent the processing will be, and the less likely it is to be effective.

The literature also suggests that the richness of the incorrect mental model makes it harder to substitute it with a correct model (Walter & Tukachinsky, 2019). For example, a model connecting ‘Hillary’ with ‘pedophile’ and ‘pizzeria’ is richer than the model connecting just ‘Hillary’ and ‘pedophile.’ Creating an alternative mental model will necessitate including a coherent explanation for the pizzeria aspect of the story, which is a more mentally difficult task. Research also points out that simple refutations leave remnants of the misinformation untouched (Cappella et al., 2015). To fully remove the false linkages, more ‘enhanced’ refutations are needed, which engage emotions or offer causal linkages. While some news stories about fake news seem to offer refutations that include narratives that provide a fuller account of the misinformation, such coverage is typically found in magazines and other in-depth genres, that do not reach all the audience of mainstream news media.⁴

Congruence between the misinformation and audiences’ prior attitudes, beliefs and opinions also shapes audience retention of the misinformation from media reports about fake news, given findings demonstrating that the ability to correct misinformation is attenuated by audience's preexisting beliefs (Walter & Tukachinsky, 2019). Unsurprisingly, audiences who already view Hillary Clinton negatively will be more likely to accept a false story demonstrating her negative behavior, as well as more likely to resist corrective information that highlights her positive actions. These findings can be understood as an extension of the motivated reasoning approach – whereby people can process information either with accuracy goals (i.e. trying to reach the most accurate conclusion) or directional goals (i.e. trying to reach a conclusion that is consistent with their broader worldview; see Kunda, 1990). When it comes to value-laden beliefs and polarizing issues, research suggests that information is processed with directional rather than accuracy goals in mind.

Also important in this context is that news stories reporting about fake news in an attempt to correct misinformation are not necessarily perceived as more credible than the fake news they try to expose or correct. Audience trust in the mainstream media is low in many countries, and in about half of the countries studied in the World Values Surveys and European Values Surveys, it is decreasing (Hanitzsch et al., 2018; For a recent overview of research on media trust, see Strömbäck et al., 2020). Even in societies where media trust is high, such as the Philippines or Japan (Hanitzsch et al., 2018, Table 1), rather large segments of society (25% of adults or more) distrust the media. This is important since research has demonstrated that trusting audiences have a stronger tendency to accept media facts and narratives, compared to audience scoring low on trust in media (Ladd, 2012). In a situation in which news outlets try to refute an ideologically-congruent claim (and label it ‘fake news’), and when the audience member does not trust the media, it is thus possible that the retraction will fail or even backfire among large segments of the public, given that it comes from a non-credible source (the news media).

5. Conclusion

While a lot has been written about fake news during the last few years, the role of mainstream news media in the dissemination of fake news has received significantly less attention. To help remedy this, the purpose of this article has been to review and synthesize the literature pertaining to three key questions: (1) What role do mainstream news media play in the dissemination of fake news? (2) Why do mainstream news media cover fake news? (3) What are the potential influences of mainstream news coverage of fake news on their audiences?

The results of our analysis suggest that mainstream news media in fact play a significant and important role in the dissemination of fake news. While no empirical estimates of the exposure to fake news stories through mainstream news media exist, based on research documenting the very concentrated and relatively limited exposure to fake news on social media (Grinberg et al., 2019), and the fact that some of the stories are remembered, recognized and even believed by large segments of the audience (Frankovic, 2016), strongly suggests that, at least when it comes to the most heavily covered stories, more people learn about these stories from mainstream news media than from social media. The evidence might be circumstantial, based on inference, rather than on direct evidence, but we still consider it quite convincing.

Why then do mainstream news media cover fake news? As suggested by our analysis, there are several answers to that question. One key reason is that mainstream news media feel compelled to cover fake news stories because some of these stories carry enormous news values, and given their role perceptions as the guardians of the truth. Another reason why mainstream news media feel compelled to cover fake news stories is that other news media cover them, and for partisan media, it also matters that some of the fake news stories fit their ideological tendencies. Indeed, research demonstrated that partisan media are more influenced by mainstream media by fake news (Vargo et al., 2017). At any case, an underlying reason is that many fake news stories are designed to fit important criteria of newsworthiness, regardless of whether these are shaped by journalistic considerations only or partisan considerations also.

Turning to the influences of mainstream news coverage of fake news on their audiences, based on theory and research in social psychology and the psychology of truth assessment, it can be inferred that it is more than likely that the result of news coverage of fake news in mainstream media is that significant parts of the audience (who did not see the original publication, just their coverage in mainstream news) internalize the wrong information or at least become less certain regarding the truth. The most likely victims are those with a high ‘latitude of acceptance’ of the disinformation (to borrow a classic concept from Sherif & Hovland, 1961; see also Eagly & Telaak, 1972), that is, those who already hold preexisting negative attitudes towards the object of negative information, or those for whom the fake information echoes preexisting schemata. Some audiences are probably also more likely to retain the wrong information although it was covered as ‘fake news,’ if they for example suffer from strategic memory impairments or of information overload. It is hard to speculate about those who are heavy consumers of mainstream news media. On the one hand, the repeated exposure increases the familiarity of the wrong information contained in fake news, which has been found to be related to retention of wrong information in psychological research. On the other hand, heavy consumers of mainstream news media are also more likely to be exposed to the corrections and, even more importantly, to elaborate explanations of why and how the wrong information has been circulating. Because of this, they may be less vulnerable to the disinformation.

While we believe this review and synthesis of the literature on the role of mainstream news media in the dissemination of fake news has covered important ground, it is also clear that research in this area suffers from several deficits. To begin with, one clear conclusion stemming from this review is that systematic research on news media coverage of fake news, and how audiences encounter such coverage, is abysmally missing. What kind of fake news stories receive widespread attention and coverage in mainstream news media? How do mainstream news organizations cover these fake news stories? Given previous research on fact-checking, it is particularly important to learn if the refutations are complex or simple (as simple refutations were found to be more effective; Walter et al., 2019), whether visual representations are included (as these were found to be counter-productive; ibid.) and whether they correct only parts of the claims of fake news (which also attenuates the effects of fact-checks; ibid). While several studies have examined the frequency and salience of coverage of fake news and the flow of stories about fake news between different types of media (Al-Rawi, 2018; Vargo et al., 2017), too little is known about the style or styles of reporting about fake news in mainstream news media. Systematic content analyses of mainstream news reports about fake news is thus warranted.

A second type of research needed to better understand mainstream news media coverage of fake news should target journalists in order to investigate why they cover fake news the way they do. Newsroom ethnographies, including both participant observation and in-depth interviews, could for example shed light on procedures and practices, but also on potential tensions between the conflicting journalistic values of not disseminating rumors, on the one hand, and seeking to clean the public sphere from wrong information, on the other. Analyses of professional journalistic discourse, for example in journalists’ conferences and trade magazines, could supplement the information gathered using observation and interviews in highlighting professional dilemmas regarding what stories to cover, and how to approach correction. Such research could also reveal whether journalists are at all aware of the role they potentially play in the dissemination of fake news and whether they fear falling pray and taking part in such dissemination. Given findings that partisan media are more heavily influenced by fake news (Vargo et al., 2017) and thus, play a more central role in spreading fake news, such studies of partisan media newsrooms are also warranted.

Finally, too little is known about audience response and processing of mainstream news media reports on fake news. This article has inferred and theorized based on aggregate reports about public belief in heavily covered fake news items, and based on empirical research about audience processing in the related area of fact-checking. One thing that is missing is however experimental research focusing specifically on audience response to the correction of online disinformation by mainstream news media. The dependent variables in such studies should include not only indicators of what facts the public believes, but also indicators of media trust and attitude certainty. Studies that link content analysis (on coverage of fake news in various outlets) to survey data (tapping exposure to various outlets, on the one hand, and beliefs about the accuracy of the fake news claims, as well as the certainty of these beliefs) will likewise help us assess the effects of media coverage of fake news on public knowledge, attitudes and behaviors. Such studies should not ignore audiences of partisan media, given the potential role of such media in spreading fake news (Vargo et al., 2017).

While we argue that mainstream news organizations somewhat paradoxically are important disseminators of fake news, we should emphasize that we do not argue that ignoring fake news is necessarily more normatively desirable than reporting about such stories. Nor do we see media coverage of fake news as a symptom of a dysfunctional media system, being too focused on negativity. The important question of mainstream news media’s coverage of fake news is partly a matter of if, but also of when, how, and to what extent. Journalists will have to work with academic scholars and rely on empirical research in developing formats that will enable reporting on, and correcting fake news, despite the potential negative effects of repeating the wrong information. This could be done in ways that are similar to past empirical research dedicated to ad watches (Jamieson, 1993, pp. 281–288) and misleading political claims (see Walter et al., 2019), that partly shaped journalistic practice.

Journalistic reports about fake news are important, because they are a matter of upholding or strengthening what Graves and Wells (2019) call norms of factual accountability, that is norms that truth matters, that those who peddle untruths should be held accountable, and that deviations from truth should have negative repercussions. In that context, mainstream news media are and remain an important institution, and its commitment to exposing fake news and correcting disinformation should be celebrated and used, but in ways that do not backfire by spreading lies in the public sphere.

The contribution of the current essay is that it turns the spotlight on the role played by mainstream news media in the dissemination of disinformation. Despite its importance, this is just a small step forward. In order to inform mainstream news media on how to best tackle the challenges fake news pose to the citizenry, much more theoretical and empirical research is required.

Acknowledgements

The authors thank Nathan Walter for his comments on a previous version.

Disclosure statement

No potential conflict of interest was reported by the author(s).

ORCID

Yariv Tsfati http://orcid.org/0000-0001-7395-3257

H. G. Boomgaarden http://orcid.org/0000-0002-5260-1284

J. Strömbäck http://orcid.org/0000-0001-7392-9791

A. Damstra http://orcid.org/0000-0001-7753-018X

E. Lindgren http://orcid.org/0000-0002-9032-1347

Additional information

Funding

This work is part of the research project ‘Knowledge Resistance: Causes, Consequences, Cures,’ funded by Riksbankens Jubileumsfond for the Advancement of the Humanities and Social Sciences [grant number M18-0310:1].

Notes

1 They used automatic tools to filter bots and matched the Twitter accounts with the voter registry, and hence their data relate to Twitter users who are potential voters

2 The news outlets included ABC News Transcripts, CBS News Transcripts, NBC News, PBS News Hour, National Public Radio, The Associated Press, USA Today, The New York Times, The Washington Post, The Philadelphia Inquirer, and the Los Angeles Times.

3 In most psychological research misinformation is considered a broader category that also includes disinformation, and no differentiation is made between mis- and dis-information. For example, Pennycook and Rank (2019) studied stories from the 2016 elections saying that Hillary Clinton is dying (i.e. disinformation) but they referred to it as misinformation.

4 Unfortunately, the content of mainstream news coverage of fake news has not been systematically studied. However, when one examines, for example, the coverage of Pizzagate in prominent mainstream news outlets, it seems that much of the time, the headlines do not contain the refutation at all. While the full refutation could be found inside the article, not in the headline, some readers will not pay the same amount of attention to what is written inside the story.