Fake news and fact-checking: The impacts of content and context

(1)

FAKE NEWS AND FACT-CHECKING I

UNIVERSITY OF AMSTERDAM 02.02.2018

Department: Graduate School of Communication Program: Master’s Program Communication Science Project: Master Thesis

Tutor: Penny Sheets Thibaut

Fake news and fact-checking: The impacts of content and context

Submitted by:

Tim Lenke

Course of studies: Political Communication (M.Sc.) Student ID: 11368152

Address: Niasstraat 1

1095 TS Amsterdam E-Mail: tim.lenke93@gmail.com

(2)

FAKE NEWS AND FACT-CHECKING 1

Abstract

Today, there is an ever-increasing amount of information online that flies the flag of

journalism but does not qualify by professional standards. Fake news - information designed to mislead - keep journalists and researchers busy and 137 active fact-checking groups around the world deal with factual inaccuracies nowadays. This paper explores specific aspects of news – both content and context which might affect credibility perceptions of real and fake news. An experiment conducted in Germany finds that respondents do perceive fake news content as significantly less credible. This result refutes common concerns about the risk of fake news being completely mistaken for real news, as well as the resultant loss of informed citizens. At the same time, such fake news content is perceived as more credible when in a traditional news mediated context, versus a non-traditional, fake-news-style website. This indicated the key role traditional media outlets can play in combatting fake news. Finally, fact-checking is proclaimed and contested as a remedy to fake news, this paper finds a surprisingly high intention to seek fact-checking sources regardless which news version was received, but little evidence to explain this trend.

(3)

Fake news and fact-checking: The impacts of content and context

“Trees cause four times more pollution as automobiles and factory chimneys put together” (Dobbs, 2012, p.4). This was one of the claims used in Ronald Reagan's presidential campaign in 1980 among several doubtful assertions, which turned out to be incorrect. On these grounds, modern-day “fake news” can be dated back to Ronald Reagan's presidency, because reporters began to regularly check factual inaccuracies after his press conferences and television statements (Dobbs, 2012). The phenomenon of fake news - information designed to mislead - is sowing mass confusion (Barthel, Mitchell & Holcomb, 2016). Usually, citizens have relied on publishers, editors, and experts to vet their consumed information (Wineburg, McGrew, Breakstone & Ortega, 2016). But with the rise of the unregulated Internet, especially participatory media including social media and blogs, the job of assessing the credibility of information has shifted from editors to online information users (Hu & Sundar, 2010). The philosopher Michael Lynch observed that the Internet is “both the world’s best fact-checker and the world’s best bias confirmer – often at the same time” (Lynch, 2016). Today, news information is highly contested and it has become increasingly difficult to discern reliable information from the rest (Graves, 2016). Never before was so much information available at our fingertips (Wineburg et al., 2016). But at the same time, “there is an ever-increasing amount of material that flies the flag of journalism but does not qualify by our standards” (Giles, 2005, para. 9). Whether this huge amount of available information will make us better informed or narrower-minded will depend on the awareness of the problem and the educational response (Wineburg et al., 2016). At present, fake news keeps journalists and researchers worried about the threat it poses to democracy due to the ease with which it spreads and flourishes (Wineburg et al., 2016).

(4)

While much has been written about fake news, not much academic research has been published on the subject (Ordway, 2017). Recent studies have covered the effects of fake news on social media (Allcott & Gentzkow, 2017; Gupta, Lamba, Kumaraguru & Joshi, 2013; Zubiaga, Liakata, Procter, Hoi & Tolmie, 2016) with the main result that fake news is largely spread and shared via social media (Barthel et al., 2016). Other studies have focused on the efficacy of fake news to explore ways to resolve fake news online (Chan, Jones, Hall Jamieson & Albarracin, 2017) or political misinformation (Rojecki & Meraz, 2016; Nyhan & Reifler, 2015) with the outcome that fake news is ideally dismantled by refuting the false information through a causal explanation of context information (Nyhan & Reifler, 2015). And lastly, some researchers attempt to debunk fake news through detecting algorithms or machine learning (Conroy, Rubin & Chen, 2015; Figueira & Luciana, 2017).

This paper aims to untangle the circumstances under which citizens give credibility to fake news and when citizens are most likely to consult fact-checking. With the key concept of news credibility, this paper explores both context and content of fake news through an online experiment. The context was manipulated either to be professionally mediated (by traditional news outlet) or unmediated (an independent website bearing little semblance to professional journalism), which is often where fake news originates (LaCapria, 2017), whereas the content was either real or fake. This 2x2 design allowed exploration of citizens’ credibility judgments about online news. Moreover, this paper assumes that fact-checking might be the most

promising remedy for fake news. On this account, it is expected that the perceived credibility of online news influences the intent to seek fact-checked information. These assumptions were studied for the case of Germany because the country has not been scrutinized

extensively regarding fake news and Germany is unique experiencing with fake news mainly spread by vocal supporters of fringe movements (Cerulus, 2017). With the continuing growth of information sources and channels online, especially for fake news about right-wing

(5)

populists in Germany (Neudert, Kollanyi & Howard, 2017), it becomes crucial to gain an understanding of how credible fake news is perceived.

Theoretical Background The rise of fake news

The fully functioning democracy needs an informed citizenry to deliberate about important issues in open and free discussions (Habermas, 1992). Without adequate

information, citizens are unable to participate in public debates, make judgments on the basis of characters instead of issues and feel less inclined to participate in politics (Galston, 2001). While the importance of adequate information for healthy decision-making is undisputed, it has proven challenging to find ways to provide such information (Delli Carpini & Keeter, 1996). In previous centuries, information was distributed through “one-to-many” broadcast technologies like newspapers, television or the radio. With the rise of the Internet and especially participatory media like Twitter or Blogs, the quality of information increasingly estimated by peer-to-peer networks (Wardle, 2017). Although participatory media was encouraged in early years for their unmediated and diverse public debate (Gillmor, 2004; Bowman & Willis, 2003), they are now criticized for the overload of available information including the growing amount of so-called “fake news” (Rochlin, 2017). Fake news is defined by the absence of a factual basis and its appearance in writing style and design like real news (Allcott and Gentzkow, 2017; Rochlin, 2017; Tandoc, Lim & Ling, 2017; Horne & Adali, 2016). Moreover, fake news is content that is “intentionally and verifiably false and could mislead readers” (Allcott & Gentzkow 2017, 213). Although the fake news has been disseminated for centuries from ancient forms of propaganda to the sorts of campaign examples highlighted in the introduction, it has reached new heights in recent years due to a number of factors.

(6)

One major factor is the use of news bots, which are basically software applications that compile and create online news relatively unsupervised and in some cases solely dedicated to establishing entire ecosystems of fake news sites (Tandoc et al., 2017).

Another factor is the direct entry of news articles through social media feeds, which leads to bypassing the trusted news outlet for the selection of online news, for instance, the news homepage, which opens the floor to publish online articles for everyone (Gottfried & Shearer, 2016). On this account, citizens increasingly attempt to take over the role of professional journalists by publishing online news on their own website or blog with large audiences, but without any affiliation to a newsroom (John & Johnson, 2016). Previous research has argued that citizen’s opinions have been perceived as more credible than expert’s opinions (Gillmor, 2004; Bowman & Willis, 2003) because regular citizens are seen as more credible due to their authenticity (Wang, Walther, Pingree, & Hawkins, 2008). This phenomenon has arguably changed the process of gatekeeping of online news (Bruns, 2005). Traditionally, trained editors assessed and selected credible information, but with the rise of the Internet, the role of gatekeepers shifted from editors to online information users (Hu & Sundar, 2010).

News credibility

At the heart of the problem of fake news is arguably the issue of credibility, which raises the question: under which circumstances do citizens believe or question fake news? And what is the role of such credibility judgments in prompting citizens to seek out fact-checking? Historically, both credibility and trust have been central in studies to approach this question of when audiences believe the news (Bakker, Trilling, de Vreese, Helfer &

Schnbach, 2013). Since Hovland, Janis, and Kelly (1953) first examined source credibility, communication research has focused on credibility and trustworthiness as two key

(7)

have been used interchangeably in a number of studies (Hellmueller & Trilling, 2012), there is a theoretical distinction between the concepts and an inconsistency in their measurement (Jackob, 2010). It has been noted that credibility usually precedes trust (Bakker et al., 2013). Based on the definition that “credibility can be defined as ‘believeability’” (Wathen & Burkell, 2002, p. 135), this paper utilizes credibility as the key concept.

Two crucial types of credibility are context credibility, and content credibility

(Appelman & Sundar, 2015; Chung, Nam & Stefanone, 2012; Lucassen & Schraagen, 2012; Metzger, Flanagin, Eyal, Lemus & Mccann, 2003). Scholars understand that both the content of the message and the context in which it is conveyed impact the receivers’ credibility judgment (Lucassen & Schraagen, 2012). This paper compares content that is either real or fake, presented in a mediated or unmediated context. In this way, both context and content credibility should be affected.

The context of online news

There is not much known about the context credibility of political news, but it has been noted that context in which a message is conveyed is crucial for the perceived credibility of citizens (Bakker et al., 2013). Previous research argued that the context of information is one of the essential indicators to judge credibility (Lucassen, & Schraagen, 2012). For this reason, the context can be understood as a shortcut to evaluate online information.

It has been argued that the judgment of context credibility does not differ between the Internet and traditional media (Chung et al., 2012; Reih & Belkin, 2000). While the context of traditional media was primarily judged on news sources such as the Washington Post, The New York Times, BBC or CNN, the online context is similarly assessed according to sources and in some cases even with a consideration of suffixes of URL addresses, for example,

(8)

“.edu, .org and .gov” (Chung et al., 2012; Reih & Belkin, 2000). This paper investigates two versions of online news contexts, which I refer to as mediated and unmediated. This

distinction is based on the shift in gatekeeping and the opportunity of online publishing, which enable trained journalists and ordinary citizens alike to report news online (Bruns, 2005). Consequently, this paper studies if citizens prefer political information from an identifiable, traditional news source which I term ‘mediated’ because of the traditional journalistic experience or ‘unmediated’ contexts of independent websites that are not affiliated with traditional news outlets (Kang, Bae, Zhang, Sundar, 2011).

In a mediated online context, a trained journalist is responsible for the assessment of the information before publishing and the news outlet bears responsibility for the content (Bakker et al., 2013; Lucassen, & Schraagen, 2012). Previous studies indicate that journalists compared to ordinary citizens are able to increase the credibility of a message in certain settings, especially for young people (Bakker et al., 2013). Regardless, it is the idea that a mediated context suggests some institutional responsibility and credibility that is essential here.

In contrast, an unmediated online context represents an independent or unaffiliated website, which is directly accessible to citizens without any warning that it may contain non-credible information (Bakker et al., 2013). Most characteristic for the unmediated context is the absence of any journalistic training of the publisher, hence the news production is made by ordinary citizens on their own platform, for instance, their website or blog (Mejias & Vokuev, 2017). Neither pre-existing experience or training of the publishers nor any known news outlets are associated with the unmediated online context (Lucassen, & Schraagen, 2012). In Germany, such unmediated sites are the origin of fake news mainly spread by fringe movements or political blogs, which are boosted on Facebook and other social media platform (Cerulus, 2017; Shuster, 2017). During the campaigning of the general election

(9)

2017, German public opinion polls indicated that 59% of Germans have reported exposure to such fake news sites (Cerulus, 2017; Shuster, 2017).It is not the case, of course, that all such unmediated sites are necessarily fake; instead, the empirical point here is to compare for citizens whether a traditional mediated context is perceived as more credible than an independent, unmediated one.

Given these robust findings of source credibility in the past literature

(Appelman & Sundar, 2015; Chung et al., 2012; Lucassen & Schraagen, 2012; Metzger et al., 2003), we can assume that credibility associated with traditional news outlets will make that mediated context perceived as more credible by respondents. Therefore I pose my first hypothesis:

H1: The context credibility news (real or fake) is higher in the mediated than in the unmediated context.

The content of online news

Since Lazarsfeld, Berelson, and Gaudet (1944) used the term selective exposure, much communication research has focused on citizens’ choice of information. Previous studies have shown that citizens prefer attitude-consistent messages to attitude-discrepant messages, also called confirmation bias (Taber & Lodge, 2006). The confirmation bias is one of the reasons for citizens’ lack of factual knowledge in politics (Delli Carpini & Keeter, 1996), which also affects their expressed opinions (Althaus, 1998; Kuklinski, Quirk, Jerit, Schwieder, & Rich, 2000; Gilens, 2001). Taken to its most extreme, these patterns could lead to ill-informed citizens. From a normative perspective, it is important to distinguish between misinformed and disinformed citizens (Kuklinski, et al., 2000). While misinformed citizens are unable to answer questions on a factual basis, disinformed citizens hold demonstrably

(10)

false beliefs about the correct answer (Kuklinski, et al., 2000). Such disinformation has become an imminent concern of scholars because it distorts citizen’s opinions and actions about consequential issues not only in politics but also in science or medicine (Flynn, Nyhan & Reifler, 2017).

So far academic research has neglected to offer a clear distinction between real and fake news. Early studies have categorized fake news into types including parodies, satire and propaganda (Tandoc et al., 2017). Today, one of the most known non-profit organizations researching fake news is First Draft, which outlined content characteristics of observed fake news, namely misleading content, imposter content, fabricated content, false connection and manipulated content (Wardle, 2017). Eventually, the existing work only captures a

characteristics of fake news mostly the intentional ‘creation of something new’ to

disseminated demonstrably false information with a clear intention to harm (Wardle, 2017). Therefore, the distinction between real and fake information can be described through an overproportionate use of cognitive shortcuts, mainly short articles with loads of visuals to persuade readers (Horne & Adali, 2016). Practically speaking, fake news is targeting audiences which skim the news online because this audience is most likely to not read beyond titles (Horne & Adali, 2016).

On this account, the fake content in the experiment of this paper is based on an authentic news article of the prestigious news website “Spiegel Online” but the line of argumentation was fabricated through manipulated facts, non-existent public opinion polls, quotes from unknown experts, misleading pictures and fake graphics. In contrast, the real news article was a shortened version of the original. These two formats of content are

expected to trigger contradictory perceptions of content credibility, namely higher credibility for real content and lower credibility for fake content. I therefore pose my second hypothesis:

(11)

H2: The content credibility is higher for real than for fake content regardless the context.

The first and second hypothesis examine credibility for either the context or content of news articles, which is essential for this paper, however, there might be interesting interactions between all four experimental conditions. Therefore, this paper explores an additional research question to outline the differences across conditions, for instance, if fake news in a mediated context is perceived as more credible than real news in an unmediated context. For this reason, I pose the following research question:

RQ: To what extent does credibility of online news differ across the four experimental conditional, namely real news, fake news, mediated, and unmediated?

The Need for Cognition

It has been noted that information processing does not work similarly for everyone. Scholars recognize that individuals differ in their “Need for Cognition” which has become one of the most utilized concepts to capture the different engagement of individuals in processing effortful information (Graham, 2007).

The Elaboration Likelihood Model outlines individual relevance to process information (Petty & Cacioppo, 1986). First, information processing depends on the involvement and motivation (Petty & Cacioppo, 1986). Highly involved and motivated individuals process information centrally through attentive consideration of arguments (Petty & Cacioppo, 1986). In contrast, individuals with little involvement and motivation process information peripherally due to signals or simple conclusions (Petty & Cacioppo, 1986). However, these assumptions of information processing due to involvement and motivation do not affect everyone to the same extent (Cacioppo & Petty, 1982; Graham, 2007). On this

(12)

account, the Elaboration Likelihood Model outlines the Need for Cognition to capture the difference of individual’s engagement in effortful information (Cacioppo & Petty, 1982; Graham, 2007). In previous literature, the Need for Cognition has been described as a personality trait that refers to people’s tendency to enjoy thinking (Bakker et al., 2013). Citizens with a high Need for Cognition are associated with the increased appreciation of complex tasks such as idea evaluation or problem-solving, while citizens with a low Need for Cognition (NFC) avoid challenging information (Graham, 2007; Kardash & Noel, 2000). Accordingly, citizens with high NFC are more likely to give credibility to information due to the quality of the message, whereas citizens with low NFC tend to use information shortcuts rather than the message itself (Martin, Lang & Wong, 2003). Therefore, the NFC is expected to moderate the effect of the context and content credibility in this paper.

H3: The difference in context credibility between the mediated and unmediated context is higher for respondents with a low Need for Cognition.

H4: The difference in content credibility between the real and fake content is higher for respondents with a high Need for Cognition.

Fact-checking: The remedy to fake news?

Fact-checking is a new genre of accountable journalism to evaluate the truth of public claims, typically but not exclusively political claims (Graves, Nyhan & Reifler, 2015). Usually, fact-checking is associated in the public with a verification of statements during political campaigns, but ”fact-checking is not just for political campaigns anymore […] this kind of journalism has become a regular feature of the news landscape” (Graves & Glaisyer, 2012, p.3). “Fact-checking should be to provide clear and rigorously vetted information to

(13)

consumers so that they may use the facts to make fully cognizant choices in voting and other essential decisions” (Elizabeth, 2014, para.11). Formally, fact-checking is dedicated to fulfilling the ideal of journalism by writing for an audience of democratic citizens (Graves, 2016). However, the mission of fact-checking is purely the provision of information, not a change in the public discourse (Graves, 2016), which makes fact-checking one of the most indispensable solutions for fake news nowadays. The main difference between modern-day fact-checking and the longstanding process of in-house controls, meaning the confirmation of facts before publishing, is the investigation of claims that have been published already

(Scriber, 2016). In other words, fact checkers perform a new variant of the reporter’s role, namely as the witness of mediated events (Graves, 2016). Nowadays, fact-checking is

contested terrain since journalists, media critics and politicians lay claim to it (Graves, 2016). Some scholars even note that fact-checking is the new form of accountable journalism, which inheres the potential to remake the news in the present ‘post-fact’ and ‘post-truth’ age

(Graves & Cherubini, 2016).

While the context and content are expected to have an influence on the credibility of online news, this paper is equally interested in studying the action recipients undertake to ensure or correct disinformation. Therefore, this paper investigates which context and content lead citizens to actively seek fact-checking. As one of the first studies analysing this relation, it is assumed that the credibility judgment of online news, especially fake news, influences the interest to consult fact-checking.

H5: The lower the context credibility of news article the higher the intention to consult fact- checking.

(14)

H6: The lower the content credibility of news article the higher the intention to consult fact- checking.

Fake news in Germany

While fake news is mainly discussed and researched in the context of the United States and the United Kingdom, this paper aims to widen the view by investigating Germany. The present debate on the fake news in the U.S. stresses automated bots that spread

propaganda across social media. Germany has to deal with fake news mainly spread by vocal supporters of fringe movements (Cerulus, 2017). International journalists argue that

Germany’s media system inheres important boundaries to resist fake news such as strict regulations on social media and the prevalence of its strong public broadcasting system (Bershidsky, 2017). However, recent research has confirmed that fake news was disseminated both before and during the general election in 2017, with most fake news working in favour of the right-wing populist party ‘Alternative für Deutschland, AfD’ (Neudert et al., 2017).

Uncertainty prevails in Germany regarding fake news some polls argue that 59% of Germans have reported exposure to fake news online and 61% witnessed fake news as the threat to democracy (Shuster, 2017). Another poll by BBC World and Global Scan has shown that Germany appears to be the least concerned country about fake news, with only 14% strongly worried and 33% somewhat worried (Cakebread, 2017). This paper aims to reveal the perception of fake news in Germany, especially the influence of context and content on the intention to consult fact-checking.

(15)

Method General Design

To test the hypotheses, an online survey experiment was conducted in Germany with N=394 cases, after the removal of all invalid cases, which are respondents that could not recall the source of the online news. The online survey experiment was designed as a 2 (mediated vs. unmediated context) x 2 (real vs. fake content) between-subjects design.

Participants were randomly assigned to one of the four conditions. The topic of the study was the context and content of fake news alongside the issue of education in Germany, which is sufficiently relevant to participants and is also regularly discussed in the German news media (Allgöwer, 2017; Munzinger, 2017, Bernewasser, 2017). In this way, this paper increased the external validity of the online survey experiment.

Procedure

First, all participants completed measures to ensure their knowledge of the German language and that they were above the age of 18. Demographic information such as gender, employment status, and education were also gathered. Afterwards, a number of media use and political interest questions was answered. Then the possible moderator ‘Need for Cognition (NFC)’ for medium and message credibility was examined with six standardized questions provided by Bullock (2011), before participants were asked to read the stimulus material, namely a news article.

For the experiment presented in this paper, the news article covers the education system in Germany. From a German citizen’s perspective, education is a permanent public debate, especially regarding the differences between federal and national level (Klovert, 2016). One of the most controversial issues in Germany’s education system is the high school graduation after eight or nine years (Klovert, 2016) Nowadays, every federal state is handling

(16)

education in their own interest, which offers a controversial issue on the national level of Germany. From an academic perspective, the discussion about education in Germany offers an untapped field for fake news, as even the right-wing Populist Party, “Alternative for Germany (AfD), discusses education rather sober-minded than emotionally with the pledge to focus on knowledge in schools (Schmoll, 2017). Therefore, education in Germany, especially the discussion on the length of high school offers an issue of national importance without significant confounding factor (gender, age, partisanship, political involvement) or an existing history of fake news.

In the experiment of this paper, the mediated context is illustrated by the news site “Spiegel Online”, which is the most used online news site in Germany according to the Digital New Report 2017 (Reuters Institute, 2017). The dubious news site Exil Press in Germany portrays the unmediated online context. The unmediated news site was highlighted in a list of fake news sites collected by professional journalists to disclose hyperpartisan and conspiracy news sites reporting about Angela Merkel in Germany (Nardelli & Silverman, 2017). However, the frequency of published articles on Exil Press is sparse, which is the reason why it can be described as underground fake news site (Nardelli & Silverman, 2017). This underground fake news site, which is published by one ordinary citizen according to the masthead is ideal to test impartial reactions towards this unmediated context (ExilPresse, 2017).

The factual content was either real or fake, as described earlier. The fake content included manipulated facts, non-existent public polls, quotes from unknown experts,

misleading pictures and fake graphics. In essence, facts of the original article were reversed and supported by extreme, non-existent public polls. The line of argumentation was faked by quotes from long-since retired public officials. Moreover, the cover picture was contradicting to the headline, for instance, the headline claimed mental shortcomings of pupils, but the

(17)

cover picture portrayed physically disabled people. The same logic applied to the fake graphics, which contained exaggerated impacts of the education system on pupils’ behaviour including depression, violence, and suicide risk.

On the other side, the real news article was a shortened version of the original article with a reasonable argumentation that the change to eight years of high school has no

significant influence on pupils’ behaviour or health, which is based on studies of the prestigious University of Tubingen.

The article was presented to respondents as an image file, a sort-of screen capture-style of a website with the article on it. The created image derived from the original websites of “Spiegel Online” or “Exil Press” including the URL, logo, structure, resort, comment column, social media menu, and advertisings. The article was presented as a short news article with a headline about the new direction for the duration of high school graduation in Germany. At the time of the online survey experiment and before, the topic has not been on the national political agenda, but was discussed on federal level and by local news media (Allgöwer, 2017; Munzinger, 2017, Bernewasser, 2017), which increased the likelihood that participants were involved but did not hold strong opinions on the issue.

As a form of manipulation check, respondents were asked which news outlet published the article they read, with five different options, namely “Spiegel Online, Exil Press, Zeit Online, Sueddeutsche Online” and “don’t remember.” A successful manipulation check included participants that could recall the news source (N=394) and the failed

manipulation check included participants, who could not recall the source (N=145).

Measurement

As control variables, gender, age, employment and education were measured. Age was measured as a continuous variable, education on a 7-point-scale. On average, the sample

(18)

represents 40% male and 59% female respondents with an average age (M = 30,05; SE = 0,56) ranging from 18 to 78 years, and a high level of education including high school graduates (30,6%), undergraduate degree (35,2%) and graduate degree (25,8%).1

To measure the moderator Need for Cognition, this paper utilized the shortened 6-item scale provided by Bullock (2011) (Cronbach’s alpha = 0,7; M = 2,77 ; SD = 0,5). The dependent variables, medium and message credibility, were asked on five-item scales to tap into the different dimensions of credibility. For context credibility, this paper investigated the degree to which participants thought the context was credible, reliable, authentic, ethical and trustworthy. For the content credibility, this paper examined the degree to which participants perceived the message as clear, credible, accurate, believable and substantiated (Bakker et al., 2013).

The factor analysis of the context credibility revealed that all five items formed a single, uni-dimensional scale where one component had an eigenvalue above one (eigenvalue 3,69). All items for medium credibility correlated positively. The third item, which covered the believability of the source had the strongest association (factor loading is .94). The item “believability of the source” was reversed coded in order to establish a reliable scale

(Cronbach’s alpha = 0,91, M= 2,9, SD = 0,08). All items were averaged into one index called context credibility with a scale ranging from one to five.

The factor analysis of the content credibility revealed all five items formed a single uni-dimensional scale where one component had an eigenvalue above one (eigenvalue 3,39). All items for message credibility correlated positively. The third item, which was about the correctness of the message, had the strongest association (factor loading .87). The items “credible message” and “substantiated message” were reversed coded in order to establish a reliable scale (Cronbach’s alpha = 0,88, M = 2,85, SD = 0,19). All items were averaged into one index called content credibility with a scale ranging from one to five.

(19)

Data Collection

The experiment was run between 3rd and 20th of December, 2017. The German sample was recruited through a mixture of social media endorsement, social media groups, Email marketing and invitations in comments columns of online news. First, the social media endorsement aimed to spread the experiment through social media channels, mainly via Twitter, Facebook, Xing and LinkedIn. Secondly, the online survey experiment was shared and promoted in existing social media groups for online experiments, surveys and other online research in Germany. Thirdly, Email marketing was based on a list of professionals working in non-university research institutions with a focus on the three largest social science institutes in Germany, namely the Goethe Institute and the Leibniz Institute as well as the Hans-Bredow-Institute for media research. Furthermore, online articles with the keywords fact-checking, fake news, false statement and factual truth were detected with Google Alerts. The selection of these articles was restricted to 15 articles of conservative online news by “Die Zeit”, liberal online news by “Sueddeutsche” and tabloid online news of “Bild” to ensure participants with different political interest and from various regions in Germany. Eventually, the political ideology of the sample was quite balanced with (M = 4,0, SE = 0,27) ranging from the extreme left (1) to extreme right (10). Moreover, the sample tended to be politically interested (M = 3,85, SE = 0,05) and followed political news (M = 3,62, SE = 0,05) both measured on a scale from not very interested (1) to very interested (5).

Results

The first hypothesis of this paper presumed that respondent will rate the context credibility higher for news presented in a mediated versus unmediated context. The hypothesis was tested with an independent samples t-test, which shows a significant difference in the context credibility between the mediated and unmediated context

(20)

(Hypothesis 1). On average, respondents who were exposed to the mediated context were more likely to indicate higher context credibility (M= 2.09, SE = 0.05) than those who were exposed to the unmediated context (M= 2.95, SE = 0.06). This difference, 0.86, BCa 95% CI [0.7, 1.01], was significant t (394) = 11.02, p < 0.00. It represents a large effect size (d=1.17). On this account, Hypothesis 1 can be accepted.

Next, it was expected that the message credibility is higher for real than for fake news in both the mediated and unmediated contexts (Hypothesis 2). An independent samples t-test indicates a significant difference in the message credibility between real and fake news. On average, respondents who read the real news article were more likely to state higher message credibility (M = 1.96, SE = 0.04) than those reading the fake news article (M= 2.79, SE = 0.31). This difference, 0.82, BCa 95% CI [0.68, 0.97], was significant t (394) = 10.9, p < 0.00. It represents a large effect size (d=1.33). Finally, hypothesis 2 can be accepted.

Moreover, the additional research questions assumed differences in credibility across the four experimental conditions. The following table illustrates one interesting finding, namely that the fake news article was perceived as more credible in the mediated than in an unmediated context. Consequently, the risk of fake news is actually higher on traditional, mediated platforms or perhaps those fake outlets that start to look like ‘real’ outlets.

Table 1: The effect of online news on content and context credibility

Mediated context Unmediated context

Real News Fake News Real News Fake News (n = 126) (n = 121) (n = 75) (n = 73)

Context 3,77 a_{(0,72) 3,22}a_(0,93) _2,83a_(0,84) _2,11a_(0,81)

Credibility

Content 3,73 a_{(0,7) 2,81}a_{(1,04) 3,47}a_(0,74) _2,38a_(0,87)

Credibility

Note: Different a_{superscripts indicate a significant difference (p < .05) between all four conditions of context or content credibility;} different b _{superscripts indicate a non-significant difference (p > .05) between all four conditions of context or content credibility; higher} mean values indicate a higher credibility. n refers to the number of respondents in each condition.

(21)

Furthermore, this paper expected that the Need for Cognition moderates both the difference of context (Hypothesis 3) and content credibility (Hypothesis 4). A Manova showed an interaction of the moderator, Need for Cognition between the four experimental groups (mediated, unmediated, real and fake) and the context credibility, F(1, 394) = 43,7, p < .001, η²

= 0,31 and on the content credibility, F(1, 394) = 35,33, p < .001, η²

= 0,27 with two large effect sizes. This is in line with the hypothesized effect hence hypotheses 3 and 4 can be accepted. Figure 1 illustrates that the context credibility is lower in the unmediated than the mediated context for respondents with a lower Need for Cognition (Hypothesis 3). It can also be noted that the higher the Need for Cognition the more extreme is the perception of context credibility for both the mediated and unmediated context. Therefore, respondents with a lower Need for Cognition have a more neutral stance towards the context credibility.

Figure 1. Context credibility, depending on Need for Cognition and whether the context was illustrated in a mediated or unmediated context

(22)

On the other hand, figure 2 demonstrates that the content credibility increases nearly linearly with a growing Need for Cognition in the real news condition. Surprisingly, the credibility of fake news does not decrease linearly with a higher Need for Cognition, only a very high content credibility of fake news is associated with a lower Need for Cognition.

Figure 2. Content Credibility, depending on the Need for Cognition and whether the article contained real or fake news

Another finding of the experiment was the generally high intention to consult fact-checking in this Germany sample. On a 5-point-scale from definitely interested to definitely not interested, over 60% of the respondents were definitely interested and 27% were

interested in a fact-checked version of the article. On this account, a simple linear regression was calculated to predict the intention to consult fact-checking based on context credibility (Hypothesis 5). A significant regression equation was found F(1,394) = 10.3, p = 0,001) with an r2 _{of 0,03. Participants’ predicted intention to consult fact-checking is equal to 4,1 + 0,03} of the measured context credibility. Participants’ intention to consult fact-checking increased

(23)

0,03 for each unit of context credibility. Finally, the simple linear regression was significant, but only 3% of the variability of the intention to consult fact-checking can be accounted for by the context credibility, which represents a marginal size.

It was further examined, whether a simple linear regression can predict the intention to consult fact-checking based on the content credibility (Hypothesis 6). The equation indicated significance (F(1,394) = 8.14, p =0,005 with an r2 _{of 0,02. Participants’ predicted} intention to consult fact-checking is equal to 4,16 + 0,027 of the measured content credibility. Participants’ intention to consult fact-checking increased 0,027 for each unit of content credibility. Finally, the content credibility can only predict 2,7% of the variance in the intention to consult fact-checking. On this account, hypothesis 5 and 6 can be rejected due to the small explanation of the variance of the dependent variable namely, the intention to consult fact-checking.

Discussion

The findings of this paper portray an optimistic view towards citizens ability to

distinguish between real and fake news, as well as citizens preference towards online news by traditional mediated platforms instead of modern unmediated platforms. These results go some way toward assuaging common concerns about the risks of fake news and the

destruction of an informed citizen in the recent ‘post-fact’ and ‘post-truth’ times (Barthel et al., 2016; Flynn et al., 2017; Lynch, 2016; Ordway, 2017; Rochlin, 2017; Tandoc et. al., 2017; Wardle, 2017).

The hypothesized effects that real news content is perceived as more credible than fake news, which uses various exaggerating features like un-existent public opinion polls, quotes from unknown sources and misleading graphics and pictures can be confirmed. At the same time, the mediated platforms increase credibility while unmediated platforms reduce credibility. For these reasons, this paper takes an important step in distinguishing between

(24)

roles of content and context in affecting citizens’ credibility judgments. Although the results reassure us of citizens’ ability to detect of fake context and content, it must be noted that fake news in a mediated context was perceived as more credible than the same fake news article in an unmediated context (see Table 1). On this account, the risk of fake news is actually higher on traditional, mediated platforms or perhaps those fake outlets that start to look legitimately like ‘real’ outlets, even posing fake versions of real web pages. Consequently, traditional media outlets have not lost importance for the provision of quality online news but actually gained a central role to combat fake news (McGuire & Cormier, 2017; Cerulus, 2017). Traditional media outlets should not entertain the scenarios of the rise of fake news but focus instead on their assignment of conveying quality information, because well-established traditional media encounters more press freedom, less right-wing extremism, and corruption according to the European Broadcasting Union (Philippot & Deltenre, 2017). Consequently, citizens’ ability to detect and select quality information should be respected and not be patronized (Allcott, & Gentzkow, 2017).

The results also showed that Need for Cognition (NFC) played, again, a crucial role in information processing (Cacioppo & Petty, 1982; Petty & Cacioppo, 1986). Citizens with a higher NFC perceived the traditional, mediated outlet as more credible than the unmediated, fake-news-style website. In other words, those with a higher NFC seem to be more sceptical about the unmediated platforms, meaning they rely on qualitative content and traditional online platforms (Bakker et al., 2013; Graham, 2007; Kardash & Noel, 2000). Moreover, the higher the Need for Cognition the higher the credibility for real news. On the other side, fake news was perceived as credible to some extent even by respondents with a fairly high Need for Cognition. Only the very high content credibility of fake news was associated with a lower Need for Cognition. These findings outline that citizens with a fairly high Need for

(25)

Cognition are actually more concerned about the fake news context than the fake news content.

Additionally, this paper has pursued the motives of citizens to seek out fact-checking, but such behaviour could not be explained by either context or content credibility.

Nonetheless, the intention to consult fact-checking was in principle very high with 60% of citizens definitely interested and 27% interested. One possible explanation for the interest in fact-checking is the global boom of fact-checking since 2010 and the most recent rise in Europe (Graves et al., 2015). On this account, fact-checking inheres the potential to correct misinformation in the era of social media echo chambers, filter bubbles and fact-challenged campaigns (Graves & Cherubini, 2016). Although fact-checking is contested terrain today, academic research focused largely on the growth of fact-checking (Dobbs, 2012; Graves, Nyhan & Reifler, 2015; Lowrey 2015) and impacts and effects of fact-checking (Garrett, Nisbet, & Lynch, 2013; Gottfried, Hardy, Winneg & Jamieson, 2013; Nyhan & Reifler, 2010, 2012), but it was widely neglected to understand the elementary motivation of citizens’ to seek out fact checking. On this account, future research should enhance the examination of citizens’ motivation to seek fact-checking especially for different types of fake news.

Future research on fake news could improve on this study in a number of ways. First, it should widen the external validity through an investigation of various countries with different cultural backgrounds to generalize the results. Next, the discussion on fake news is mostly focused on channels such as social media and online news, which neglects traditional communication research practices, namely the difference of various formats such as text, audio, video or even more recent developments of instant articles, live blogs or chatbots (Allcott, & Gentzkow, 2017; Barthel et al., 2016; Wardle, 2017). Today various news formats are spread via social media, which enables citizens to share information faster than ever before (Wineburg et al., 2016). This competition of instant news coverage online poses a

(26)

threat through sensationalism and polarization (Wineburg et al., 2016). The sensationalism and polarization derive from algorithms which prioritize engagement (Tallis, 2016). For instance on Facebook it is most likely that fake news attract more user attention, thus

engaging them better (Tallis, 2016). For example, Facebook has admitted to being “too slow to recognise” Russian interference in the U.S. election in 2016, which underlines the

corrosive effect on democracy of social media (Hern, 2018).

Even governments are reacting to combat disinformation as part of modernizing defence through heavy penalties on Facebook like in Germany or by setting up security units like in the United Kingdom (Bershidsky, 2017; Walker, 2018). Furthermore, social media platforms started to combat fake news and Facebook, for instance, announced to hire over 10,000 more people in 2018 to work on safety and security, which underlines how uncertain the effects of fake news are (Hern, 2018). By considering that content and context credibility are key requirements for communication effects, it is necessary for further investigations to examine how content and context influence credibility. This paper calls for more scientific inquiry due to the rise of information sources and information channels, which make it vital to understand how citizens detect, select and perceive fake news and fact-checking.

(27)

FAKE NEWS AND FACT-CHECKING II

Notes:

1_{The randomization into the four experimental conditions was successful. The} participants were equally distributed into the four conditions, namely mediated real

(n = 126) mediated fake (n = 121) unmediated real (n = 75) and unmediated fake (n = 73) A Pearson chi-square test of independence was calculated comparing the four experimental groups to gender, age, education, and political ideology.A

non-significant interaction was found for gender ( χ2 (1) = 10,1, p = 0,35), age ( χ2 (1) = 14,5, p = 0,33), education ( χ2 (1) = 16,15, p = 0,58), and political ideology ( χ2 (1) = 24,5, p = 0,6). These results represent a successful randomization among the

experimental conditions.

References:

Allcott, H., & Gentzkow, M. (2017). Social Media and Fake News in The 2016 Election. Journal of Economic Perspectives, Vol. 31(2), 211–236, doi:10.1257/jep.31.2.211.

Althaus, S. (1998). Information Effects in Collective Preferences.

American Political Science Review, Vol. 92(3), 545-558, doi:10.2307/2585480. Allgöwer,R. (2017, March 26). G9 wird um fünf Jahre verlägert.

Retrieved from https://www.stuttgarter-zeitung.de/inhalt.gymnasien-bleiben- zweigleisig-g9-wird-um-fuenf-jahre-verlaengert.fb05db5e-6ebf-4631-9bec- 6d86a92f7faa.html

Appelman, A., & Sundar, S. (2015). Measuring message credibility: Construction and validation of an exclusive scale. Journalism & Mass Communication Quarterly, Vol. 93(1), 59-79, doi: 10.1177/1077699015606057.

(28)

FAKE NEWS AND FACT-CHECKING III

Bakker, T.P., Trilling, D.C., de Vreese, C.H.., Helfer, L., & Schnbach, K. (2013). The context of content: the impact of source and setting on the credibility of news. Recherches en Communication, 2013, Vol.40, 151-168.

Barthel, M., Mitchell, A., & Holcomb, J. (2016). Many Americans believe fake news is sowing confusion. Pew Research Centre. Retrieved from http://www.journalism.org/ 2016/12/15/many-americans-believe-fake-news-is-sowing-confusion/

Bernewasser, J. (2017, May 3). Lange genug. Die Zeit. Retrieved from:

http://www.zeit.de/2017/19/gymnasium-g8-g9-turbo-abi-abitur-nrw-schulreform Bershidsky, L. (2017, October 24). Why Germany is better at resisting fake news.

Retrieved from https://www.bloomberg.com/view/articles/2017-10-24/why-germany-is-better-at-resisting-fake-news

Bowman, B. S., & Willis, C. (2003). We Media: How audiences are shaping the future of news and information. Retrieved from http://www.hypergene.net/wemedia/download/ we_media.pdf

Bruns, A. (2005). Gatewatching: Collaborative Online News Production. New York: Peter Lang Publishing.

Bullock, J. (2011). Elite influence on Public Opinion in an Informed Electorate. American Political Science Review, 2011, Vol.105(3), 496-515, doi:

10.1017/S0003055411000165

Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42(1), 116–131. doi:10.1037//0022-3514.42.1.116

Cakebread, C. (2017, September 26). Brazil is more worried about fake news than any other country. Retrieved from http://www.businessinsider.de/brazil-is-

(29)

FAKE NEWS AND FACT-CHECKING IV

Cerulus, L. (2017). Germany’s anti-fake news lab yields mixed results. Retrieved

from https://www.politico.eu/article/fake-news-germany-elections-facebook-mark- zuckerberg-correctiv/

Chan, M. S., Jones, C. R., Hall Jamieson, K., & Albarracin, D. (2017). Debunking: A meta- analysis of the psychological efficacy of messages countering misinformation. Psychological Science, Vol. 28(11), 1531-1546, doi: 10.1177/0956797617714579 Conroy, N.J., Rubin, V. L., & Chen, Y. (2015). Automatic deception detection: Methods for

finding fake news. Proceedings of the Association for Information Science and Technology, Vol.52(1), 1-4, doi: 10.1002/pra2.2015.145052010082

Chung, J.C., Hyunjung, K. & Jang, H.K. (2010). An anatomy of the credibility of online newspapers. Online Information Review, Vol. 34(5), 669-685, doi:

10.1108/14684521011084564

Chung, C., Nam, Y., Stefanone, M. (2012). Exploring Online News Credibility: The Relative Influence of Traditional and Technological Factors. Journal of Computer Mediated Communication, Vol.17(2), 171-186, doi: 10.1111/j.1083-

6101.2011.01565.x

Delli Carpini, M., & Keeter, S. (1996). What Americans know about politics and why it Matters. New Heaven, United States: Yale University Press.

Dobbs, M. (2012). The Rise of Political Fact-checking: How Reagan Inspired a Journalistic Movement: A Reporter’s Eye View. Retrieved from http://www.issuelab.org/

resources/15318/15318.pdf

Elizabeth, J. (2014, May 20). Who Are You Calling a Fact-Checker? Retrieved from

https://www.americanpressinstitute.org/fact-checking-project/fact-checker-definition/ Exil Presse (2017). Masthead. Retrieved from http://exil-presse.de/impressum

(30)

FAKE NEWS AND FACT-CHECKING V

Figueira, A., & Luciana, O. (2017). The current state of fake news: challenges and opportunities. Procedia Computer Science, Vol. 121, 817-825, doi:

10.1016/j.procs.2017.11.106

Flynn, D., Nyhan, B., Reifler, J. (2017). The Nature and Origins of Misperceptions: Understanding False and Usupported Beliefs about politics.

Political Psychology, Vol.38(1), 127-150, doi: 10.1111/pops.12394.

Galston, W. A. (2001). Political knowledge, political engagement, and civic education. Annual Review of Political Science, Vol. (4), 217-234.

Garrett, R. K., Nisbet, E. C., & Lynch, E. K. (2013). Undermining the corrective effects of media-based political fact checking? The role of contextual cues and naïve

theory. Journal of Communication, 63, 617–637. doi: 10.1111/jcom.12038 Gilens, M. (2001). Political Ignorance and Collective Policy Preferences. American

Political Science Review, Vol. 95(2), 379-396, doi: 10.1017/S0003055401002222 Giles, B. (2005). Mainstream media and the survival of journalism. Retrieved from

http://niemanreports.org/articles/mainstream-media-and-the-survival-of-journalism/ Gillmor, D. (2004). We the media. Grassroots journalism by the people, for the people.

Sebastopol, CA: O’Reilly.

Gottfried, J., & Shearer (2016, May 26). News Use Across Social Media Platforms 2016. Pew Research Centre. Retrieved from http://www.journalism.org/2016/05/26/news-use- across-social-media-platforms-2016/

Gottfried, J. A., Hardy, B. W., Winneg, K. M., & Jamieson, K. H. (2013). Did fact checking matter in the 2012 presidential campaign? American Behavioral Scientist, 57, 1558–1567. doi:10.1177/0002764213489012

(31)

FAKE NEWS AND FACT-CHECKING VI

Graham, L.M. (2007). Need for Cognition and false memory in the

Deese-Roediger-McDermott paradigm. Personality and Individual Differences, 42(3), 409-418, doi: 10.1016/j.paid.2006.07.012

Graves, L. (2016). Deciding what’s true. The rise of political fact-checking in American Journalism. United States, New York: Columbia University Press.

Graves, L. & Cherubini, F. (2016). The rise of fact-checking sites in Europe. Retrieved from http://reutersinstitute.politics.ox.ac.uk/sites/default/files/ The%20Rise%20of%20Fac-Checking%20Sites%20in%20Europe.pdf Graves, L., & Glaisyer, T. (2012). The Fact-Checking Universe in Spring 2012.

New America Foundation Media Policy Initiative Research Paper. Retrieved from http://newamerica.net/publications/policy/the_fact_checking_universe_in_

spring_2012.

Graves, L., Nyhan, B., & Reifler, J. (2015). The diffusion of factchecking: Understanding the growth of a journalistic innovation. Retrieved from

http://www.americanpressinstitute.org/wp-content/uploads/2015/04/The-Growth-of-Fact-Checking.pdf

Gupta, A., Lamba, H., Kumaraguru, P., & Joshi, A. (2013). Faking Sandy: Charaterizing and identifying fake images on Twitterduring Hurricane Sandy. Proceedings of the 22nd International Conference on World Wide Web, 729-739.

Habermas, J. (1992). Faktizität und Geltung. Beiträge zur Diskurstheorie des Rechtes und des demokratischen Rechtsstaats. Frankfurt am Main: Suhrkamp.

Hellmueller, L.C., Trilling, D. (2012, June) The credibility of credibility measures: A

meta-analysis of credibility research in communication journals, 1951 to 2011. Paper presented at the World Association of Public Opinion Research Conference, Hong Kong.

(32)

FAKE NEWS AND FACT-CHECKING VII

Hern, A. (2018, January 22). Facebook: We were too slow to recognise our ‘corrosive’ effect on democracy. Retrieved from http://www.theguardian.com/technology/

2018/jan/22/facebook-too-slow-social-media-fake-news-hiring

Hersh, E. D. (2015). Hacking the Electorate. How Campaign perceive voters. Cambridge: Cambridge University Press.

Horne, B., & Adali, S. (2016). This just in: Fake news packs a lot in title, uses simpler, repetitive content in text body, more similar to satire than real news. Retrieved from https://arxiv.org/pdf/1703.09398.pdf

Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion:

psychological studies of opinion change. New Heaven, United States: Yale University Press.

Hu, Y., & Sundar, S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37(1), 105–132. doi:10.1177/0093650209351512

Jackob, N. G. E. (2010). No alternatives? The relationship between perceived media

dependency, use of alternative information sources, and general trust in mass media. International Journal of Communication, 4, 589–606.

John, B., & Johnson, K. (2016). Perspectives of an underconsidered stakeholder group: Citizen journalists’ views of public relations practitioners and their materials. Public Relations Review, Vol.42(1), 184-191, doi:10.1016/j.pubrev.2015.09.010

Kang, H., Bae, K., Zhang, S., & Sundar (2011). Source cues in online news: Is the proximate source more powerful than distal sources? Journalism & Mass Communication Quarterly, Vol 88(4), 719-736.

(33)

FAKE NEWS AND FACT-CHECKING VIII

Kardash, C.M., & Noel, L.K. (2000). How organizational signals, need for cognition, and verbal ability affect text recall and recognition. Contemporary Educational

Psychology, Vol. 25(3), 317-331, doi: 10.1006/ceps.1999.1011 Klovert, H. (2016). Das irrationale Geschacher um G8. Retrieved from

http://www.spiegel.de/lebenundlernen/schule/abitur-nach-zwoelf-jahren-was-hat-g8- gebracht-a-1097395.html

Kuklinski, J., Quirk, P., Jerit, J., Schwieder, D., & Rich, R. F. (2000). Misinformation and the currency of democratic citizenship. Journal of Politics, 62(3), 790-816. LaCapria, K. (2017). Snopes’ field guide to fake news sites and hoax purveyors.

Retrieved from https://www.snopes.com/2016/01/14/fake-news-sites/

Lazarsfeld, P., Berelson, B. & Gaudet, H. (1944). The people’s choice. United States, New York: Columbia University Press.

Lodge, M., & McGraw, K. M. (1995). Political judgement: Structure and process. Ann Arbor: University of Michigan Press.

Lowrey, W. (2015). The Emergence and Development of news fact-checking sites:

Institutional logics and population ecology. Journalism Studies, Vol. 18 (3), 376-394. doi: 10.1080/1461670X.2015.1052537

Lucassen, T., & Schraagen, J. (2012). Propensity of trust and the influence of source and medium cues in credibility evaluation. Journal of Information Science, Vol. 38(6), 566-577, doi: 10.1177/0165551512459921

Lynch, M. (2016, March 9). Googling is believing: Trumping the informed citizen. Retrieved from https://opinionator.blogs.nytimes.com/2016/03/09/

(34)

FAKE NEWS AND FACT-CHECKING IX

Martin, B.A.S., Lang, B., & Wong, S. (2003). Conclusion explicitness in advertising: The moderating role of Need for Cognition (NFC) and argument quality (AQ) on persuasion. Journal of Advertising, 32(4), 57–65.

McGuire, J. & Cormier, M. (2017, February 9). The public broadcaster’s role in the fake news era. Policy Opinion. Retrieved from: http://policyoptions.irpp.org/magazines/ february-2017/the-public-broadcasters-role-in-the-fake-news-era/

Mcleod, D.M., Kosicki, G.M., & Mcleod, J. M. (2002). Resurveying the boundaries of political communication. InJ Bryant and D. Zillmann (Eds.), Media effects: Advances in theory and research (2nd ed.), 215-267. Hillsdale, NJ: Erlbaum.

Mejias, U.E., & Vokuev, N.E. (2017). Disinformation and the media: The case of Russia and Ukraine. Media, Culture & Society, Vol. 39(7), 1027-1042, doi:

10.1177/0163443716686672

Metzger, M., Flanagin, A., Eyal. K., Lemus, D., & Mccann, R. (2003). Credibility for the 21st century: Integrating perspectives on source, message, and media credibility in the contemporary media environment. Annals of the International Communication Association, Vol. 27 (1), 293-335, doi: 10.1080/23808985.2003.11679029

Munzinger, P. (2017, April 6). Das bayrische G8 wird abgeschafft – endlich.

Retrieved from http://www.sueddeutsche.de/bildung/gymnasium-das-bayerische-g-ist-tot-endlich-1.3451899

Nardelli, A., & Silverman, C. (2017, January 14). Hyperpartisan sites and facebook pages are publishing false stories and conspiracy theories about Angela Merkel. Retrieved from https://www.buzzfeed.com/albertonardelli/hyperpartisan-sites-and-facebook- pages-are-publishing-false?utm_term=.mspXZAyqX#.kiEDaKGkD

(35)

FAKE NEWS AND FACT-CHECKING X

Neudert, L., Kollanyi, B., & Howard, P. (2017). Junk News and Bots during the German Parliamentary Election: What are German Voters Sharing over Twitter? Retrieved from http://comprop.oii.ox.ac.uk/wp-content/uploads/sites/89/2017/09/ComProp_ GermanElections_Sep2017v5.pdf

Nyhan, B,, & Reifler, J. (2010). When Corrections Fail: The Persistence of Political Misperceptions. Political Behavior, Vol. 32 (2), 303-30. doi:10.1007/s11109-010-9112-2

Nyhan, B., & Reifler, J. (2012). Misinformation and Fact-Checking: Research Findings from Social Science. Retrieved from

http://www.dartmouth.edu/~nyhan/Misinformation_and_Fact-checking.pdf

Nyhan, B., Reifler, J. (2015). Displacing misinformation about events: An experimental test of causal corrections. Journal of experimental Political Science, Vol. 2(1), 81-93, doi:

10.1017/XPS.2014.22

Ordway, D. (2017). Fake news and the spread of misinformation.

Retrieved from: https://journalistsresource.org/studies/society/internet/fake-news-conspiracy-theories-journalism-research

Petty, R., & Cacioppo, J. (1986). The elaboration likelihood model of persuasion. In Communication and persuasion (pp. 1-24). New York, NY: Springer.

Philippot, J. P., & Deltenre, I. (2017). Annual Report European Broadcasting Union. Retrieved from: http://www.ebu.ch/files/live/sites/ebu/files/Publications/ Activity%20reports/EBU-Annual-Report-16-17_EN.pdf

Reih, S., & Belkin, N. (2000) Interaction on the Web: Scholars’ judgment of information quality and cognitive authority. In Kraft (Ed.), Proceedings of the 63rd ASIS Annual Meeting, 25–33. Silver Spring, MD: American Society for Information Science.

(36)

FAKE NEWS AND FACT-CHECKING XI

Roberts, C. (2010). Correlations among variables in Message and Messenger Credibility Scales. American Behavioral Scientists, Vol 54 (1), 43-56,

doi:10.1177/0002764210376310.

Reuters Institute (2017). Digital News Report. Retrieved from

http://www.digitalnewsreport.org/survey/2017/germany-2017/ Rochlin, N. (2017). Fake news: belief in post-truth.

Library Hi Tech, Vol. 35(3), 386-392. doi: 10.1108/LHT-03-2017-0062

Rojecki, A., & Meraz, S., (2016). Rumors and factitious informational blends: The Role of the Web in speculative politics. New Media & Scoiety, Vol. 18(1), doi:

10.1177/1461444814535724

Schmoll, H. (2017, August 29). Innovativ oder inklusiv? Retrieved from

http://www.faz.net/aktuell/politik/bundestagswahl/die-plaene-der-parteien/was-die- parteien-bei-der-bildungspolitik-wollen-15129233-p2.html

Scriber, B. (2016, September 8). Who decides what´s true in politics? A history of the rise of political fact-checking. Retrieved from http://www.poynter.org/2016/who-decides- whats-true-in-politics-a-history-of-the-rise-of-political-fact-checking/429326/

Shuster, S. (2017, August 9). Russia has launched a fake news war on Europe. Now Germany is fighting back. Retrieved from http://time.com/4889471/germany-election-russia- fake-news-angela-merkel/

Taber, C., & Lodge, M. (2006). Motivated scepticism in the evaluation of political beliefs. American Journal of Political Science, Vol. 50(3), 755-769. doi: 10.1111/j.1540-5 907.2006.00214.x

Tallis, J. (2016, December 21). How big is the fake news problem for Facebook.

Retrieved from: https://www.forbes.com/sites/greatspeculations/2016/12/21/how-big- is-the-fake-news-problem-for-facebook/#34f3602c5bd1

(37)

FAKE NEWS AND FACT-CHECKING XII

Tandoc, E., Lim, Z. & Ling, R. (2017): Defining “Fake News”, Digital Journalism, 1-17. doi: 10.1080/21670811.2017.1360143

Walker, P. (2018, January 23). New national security unit set up to tackle fake news in UK. Retrieved from https://www.theguardian.com/politics/2018/jan/23/new-national- security-unit-will-tackle-spread-of-fake-news-in-uk

Wang, Z., Walther, J. B., Pingree, S., & Hawkins, R. P. (2008). Health information, credibility, homophily, and influence via the Internet: Web sites versus discussion groups. Health communication, 23(4), 358–68. doi:10.1080/10410230802229738 Wardle, C. (2017, February 16). Fake News: It’s Complicated. Retrieved from

https://medium.com/1st-draft/fake-newsits-complicated-d0f773766c79

Wathen, C. N., & Burkell, J. (2002). Believe it or not: Factors influencing credibility on the Web. Journal of the American Society for Information Science and Technology, 53(2), 134–144. doi:10.1002/asi.10016

Week, B. (2015). Emotions, Partisanship, and Misperception: How anger and anxiety moderate the effect of partisan bias on susceptibility to political Misinformation. Journal of Communication, Vol. 65(4), 699-719, doi: 10.1111/jcom.12164

Wineburg, S., McGrew, S., Breakstone, J. & Ortega, T (2016). Evaluating Information: The Cornerstone of Civic Online Reasoning. Stanford Graduate School of Education. Retrieved from: http://stacks.stanford.edu/file/druid:fv751yt5934/SHEG%20 Evaluating%20Information%20Online.pdf

Zubiaga A, Liakata M, Procter R, Wong Sak Hoi G, Tolmie P. (2016). Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads. PLoS ONE, Vol. 11(3): e0150989, doi: 10.1371/journal.pone.0150989.

(38)

FAKE NEWS AND FACT-CHECKING XIII

Appendix A

Manipulation Material

1. Mediated context real (left) versus fake news (right):

(39)

FAKE NEWS AND FACT-CHECKING XIV

2. Unmediated context real (left) versus fake news (right):

3. Translation of the real news article:

The irrational discussion about G8

G8 is proscribes as turbo high school diploma, a stress test for pupils. Now several federal states experiment with 9 years period to the High School graduation again. Time to ask: What did the reforms do?

(40)

FAKE NEWS AND FACT-CHECKING XV

It is one of the most bizarre plays on the stage of school reforms in Germany: the transition from high school in the 13th grade (G9) to high school in the 12th grade (G8) and back. Even the Bavarian government is thinking about leading pupils to high school graduation in nine years.

The turnaround is different in the federal states. ‘Niedersachsen’ is the only federal state that completely sets back to nine years high school - although the first G8-year just graduated five years ago. Since 2013, schools in ‘Hessen’ can decide for themselves whether they want to switch back to G9, set up individual G9 classes or stick to G8. In Baden-Wurttemberg, just 44 high schools are testing the return to G9.

The “Jako-o education study”, which interviewed 3000 parents two years ago, shows that eight out of ten parents would choose the nine-year high school for their child if they could decide. The collapse of G8 began with parents, educators, pupils and politicians who complained about the growing stress in schools in an allegedly over-optimized society. The protest against G8 led in ‘Hamburg’ and ‘Bavaria’ to referendums, in ‘Nordrheinwestfalen’, ‘Schleswig-Holstein’ and ‘Hessen’ parent initiatives and civic alliances were formed.

What about the many fears that cling to the high school graduation in the 12th grade? Are G8 students really unhappy? Do they have less time to develop before they have to find their way into the working life? Do you arrive earlier at work, according to the reforms plan?

One should not dramatize these results, says study director Ulrich Trautwein from the University of Tubingen. "Maybe the eight year high school is actually more exhausting for some students, but maybe many just feel more stressed out because the media has reported so often, how bad G8 is - or because their teachers are critical of the reform."

In some subjects, pupils perform worse, but in which? The research results contradict each other. In ‘Sachsen-Anhalt’, the direct comparison between G8 and G9 pupils, showed that G8 pupils perform a little worse in math, but equally good in German. In ‘Baden-Württemberg’, however, there was no difference in math and physics. In English, G9 students were much better and a bit better in biology. On the final graduation marks, however, this scheme hardly worked. In ‘Bavaria’, ‘Saarland’, ‘Baden-Wuerttemberg’ and ‘Niedersachsen’ for example graduates of the double years were almost equally on the high school average.

And now?

One thing the reforms definitely wasted: power. Scientists warn against turning back to G9. "I do not see any indicators that suggest that we have to get away from G8," says education researcher Klaus Klemm. Ulrich Trautwein from Tubingen says similarly: "We should make