Free beta access. 50 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Solves disinformation and social media propaganda on a psychological level.

Zerback 22

Zerback, Thomas ; Töpfl, Florian. “Forged Examples as Disinformation: The Biasing Effects of Political Astroturfing Comments on Public Opinion Perceptions and How to Prevent Them.” Political Psychology, vol. 43, no. 3, Oxford: Wiley Subscription Services, Inc, pp. 399–418, doi:10.1111/pops.12767.

The origin of the term “astroturfing” can be traced back to a brand of artificial grass (“AstroTurf”), which is used as a floor cover in stadiums and playgrounds (Zhang et al., 2013). AstroTurf mimics natural grass, and strategic communications have adopted the term to describe the imitation of civic grassroots movements that are in fact fabricated by other societal actors whose identities remain in the dark (Mackie, 2009). With the diffusion of digital media, new types of astroturfing have emerged and are referred to as digital or online astroturfing (Kovic et al., 2018; Zhang et al., 2013). In principle, every individual or organization in both economic and political realms can covertly commission astroturfing campaigns. In this article, we focus on political online astroturfing that has been described as “a form of manufactured, deceptive and strategic top-down activity on the Internet initiated by political actors that mimic bottom-up activity by autonomous individuals” (Kovic et al., 2018, p. 71). Astroturfing messages are not necessarily false or inaccurate, but they are deceptive in that they maliciously pretend to originate from “a genuine grassroots movement or sentiment” while they are “in fact orchestrated centrally and top down” (Keller et al., 2020, p. 257; Walker, 2014), while the principals intentionally seek to hide their true identity as well as their strategic goals. As such, they share some features with fake news (Egelhofer & Lecheler, 2019) but differ from the latter in that there is no attempt made to resemble journalistic content and factuality is not necessarily low. For political online astroturfing messages, common venues and distribution channels are websites containing information on matters of public interest (e.g., global warming), direct messages, and particularly user comments beneath news content that attract large audiences (Kovic et al., 2018). Astroturfing comments are thus frequently found on the websites and/or social media sites of news organizations offering their audiences opportunities to comment (see Hille & Bakker, 2014; Keller et al., 2020; Zelenkauskaite & Balduccini, 2017; Zelenkauskaite & Niezgoda, 2017). Astroturfing agents invade such habitats of public discourse and exploit their openness, anonymity, and reach by fabricating user comments that benefit the interests of their principal. Different approaches are used to recruit agents: In some cases, real users are simply paid to comment by using either their personal or fake accounts (sock puppets) (Chen et al., 2011). In other cases, agents create automated social bots that act like real users and spread the astroturfing messages (Ferrara, 2017; Kovic et al., 2018). In recent years, multiple instances of political online astroturfing activities have been documented around the world including campaigns across and within national borders and in various contexts including elections, protests, and others (see Woolley & Howard, 2018 for an overview). Particularly Russia and China have repeatedly used online astroturfing as a propaganda tool to influence political discourse in other nations as well as their own (Han, 2018; King, Pan, & Roberts, 2017; Lysenko & Brooks, 2018). Commissioned by Russia’s ruling elites, a collective of workers centered around the so-called St. Petersburg Internet Research Agency (IRA) (Chen, 2015; Lukito et al., 2020; Zelenkauskaite & Niezgoda, 2017) have imitated authentic users during the 2016 U.S. presidential election (National Intelligence Council, 2017) and the recent COVID-19 crisis (EU vs. Disinformation, 2020). Although the exact number of forged accounts is unknown, a dataset released by Twitter in 2018 gives a first impression of the scale of Russia’s efforts (Nimmo et al., 2018). Between 2016 and 2018, the company identified 10 million tweets stemming from 4,500 dubious accounts; 3,800 of these were believed to be affiliated with the IRA. Similarly, the Chinese government instructs and coordinates users (mostly state employees) to post government-friendly comments online. King et al. (2017) estimated the yearly number of such government-induced social media comments targeted at the domestic audience to be almost half a billion. Other political astroturfing campaigns have been observed in the context of the U.K. referendum on membership of the European Union (Brexit) (Llewellyn, Cram, Hill, & Favero, 2019) as well as in the run-up to national elections in France (Ferrara, 2017), South Korea (Keller et al., 2017, 2020), and the United States (Bessi & Ferrara, 2016; Woolley & Guilbeault, 2017). The goals of political astroturfing activities are diverse and encompass the change of beliefs, opinions, and attitudes; the creation of uncertainty, distrust, and political cynicism (Zhang et al., 2013); or simply a distraction of the population from other issues (Golovchenko et al., 2020; Keller et al., 2020; King et al., 2017; Linvill, Boatwright, Grant, & Warren, 2019; Linvill & Warren, 2020). The specific content and persuasive tactics applied within the astroturfing messages can vary widely and depend on the goals of the campaign. They range from supporting specific policy positions to attacking minority groups—from undermining trust in specific institutions to spreading confusion by disseminating a flurry of alternative and contradictory explanations for a key political event (EU vs. Disinformation, 2019, 2020; Linvill et al., 2019; Linvill & Warren, 2020; Xia et al., 2019). However, in each case, the underlying tactic is to pretend that the message communicated reflects the people’s voice, which in turn is expected to serve as a source of social influence by suggesting a broader social consensus regarding certain opinions (Zhang et al., 2013). This notion of consensus is central because it constitutes an important link between astroturfing comments and a long-standing line of research on media effects called exemplification. The exemplification approach has specifically focused on the question of how single incidents or opinions depicted in the media can affect our opinions and notions of social reality, particularly our perceptions of public opinion (Daschmann, 2000; Zillmann, 2002; Zillmann & Brosius, 2000). In the following, we explain how exemplification connects (1) to the assumed effects of astroturfing comments on public opinion perception and (2) why we expect inoculation messages to be an effective countermeasure against this influence Effects of Political Online Astroturfing Comments on Public Opinion Perception The first question we have to answer is why online user comments should influence perceptions of public opinion. Exemplification theory is a specifically useful theoretical approach to answer this question (Zillmann, 1999): It states that people’s perceptions of social reality partly depend on single incidents that they encounter in the media—so-called exemplars. Exemplars can be events (e.g., car accidents) but also opinions voiced by individuals. Empirical research has shown that the recipients’ notion of how the public thinks about issues can be altered by exposing them to opinion exemplars in journalistic news stories (e.g., Brosius & Bathelt, 1994; Zerback & Peter, 2018). More recent research also confirmed this effect with regard to user-comment exposure on social media (e.g., Ross & Dumitrescu, 2018; Zerback & Fawzi, 2017) or in the comment sections of journalistic news outlets (e.g., Lee, 2012). For example, Zerback and Fawzi (2017) exposed participants to user comments on Facebook, which either supported or opposed the eviction of violent immigrants and afterwards asked them to assess how the general population thinks about the issue. They showed that those who were exposed to comments opposing eviction perceived less support for eviction in the general population. Similar results have been observed by other researchers (e.g., Lee & Jang, 2010). From a theoretical point of view, exemplar effects are assumed to originate from cognitive heuristics in that exemplars depicted in the media are perceived as representatives of a broader group. Consequently, the audience tends to generalize certain features of the exemplar (e.g., an opinion) to that group (representative heuristic). This tendency to generalize intensifies when similar exemplars (e.g., similar opinions) are encountered repeatedly because the incidents become cognitively available and more easily accessible (Zillmann, 2002). Given the theoretical and empirical insights to date, we therefore assume that exposure to online astroturfing comments will also influence readers’ perceptions of public opinion. H1: Online astroturfing comments influence perceptions of public opinion in the direction of the comments. Immunizing Citizens Against Political Online Astroturfing If Hypothesis 1 holds, then astroturfing comments pose a threat to public discourse because they convey an intentionally manipulated and biased picture of opinion distributions in society. Hence, the question arises of what can be done to prevent this potentially harmful impact. Inoculation theory is a promising approach in this regard (McGuire, 1964). It is a theoretical framework that rests on the idea that people can be immunized against the effects of persuasive messages by exposing them to a preemptive inoculation message. Inoculation messages typically consist of two elements: a description of a threat and a preemptive refutation. The threat component includes a warning that there will be a persuasive attack in the future intended to alter one’s attitudes. The following preemptive refutation then educates the individual about the persuasive tactics employed by the sender. Two common forms of preemptive refutations can be distinguished: refutational-same messages educate the recipient about the exact arguments employed by the persuader and refute them. Refutational-different messages present and refute arguments that are not part of the following persuasive attempt. Both variants have been shown to effectively hamper persuasive effects of attack messages (Banas & Rains, 2010; McGuire, 1964) and can serve as a powerful tool to protect the audience from the effects of persuasive messages (see also the meta-analysis by Banas and Rains, 2010, and the extensive literature reviews by Compton and Pfau, 2005, and Compton, 2012). Although most studies on inoculation concentrate on preventing message-induced attitude change, there is reason to believe that immunization can also be achieved regarding public opinion perceptions. To substantiate this assumption, we have to recall the mechanisms by which exemplars affect our assessments of public opinion: People perceive individual opinions depicted by the media as representatives of the group that the individual typifies. For example, if a person in the media is perceived as a common citizen, then his opinion is likely to be generalized to the opinion held by the population in general. This also means that if a person is not perceived to be part of the general population, then his or her opinion will not be generalized—at least not to the same extent. In fact, studies have shown that the similarity between an exemplar and the target population to be assessed affects the strength of the generalization effect (Peter & Zerback, 2017; Zerback & Fawzi, 2017). Zerback and Fawzi (2017) asked participants to assess public opinion in two different target populations: the German general population and the German Internet users. They found that the online exemplars’ opinions were generalized to both target populations, but the effect was more pronounced in the case of the Internet population. In accordance with other researchers (e.g., Zillmann & Brosius, 2000), the authors explain this effect by referring to the work of Tversky (1977) who assumed that the degree of correspondence between an object (the exemplar) and a class of objects (the target population) determines if a person judges the object to be a member of that class. Hence, people should generalize from single exemplars to broader populations particularly in cases when the exemplar and the target population share various characteristics. What can we conclude from these theoretical assumptions and empirical findings regarding the effect of inoculation messages on the tendency to generalize from online comments? As we have already seen, the two key ingredients of an inoculation message are educating people about a possible threat (the persuasive attack) and a refutational preemption. In our context, these ingredients involve informing people about the existence of fake comments intended to alter their political attitudes. As soon as they have acquired this knowledge, they are more likely to consider the possibility that some online comments may in fact be manufactured and therefore do not represent actual citizen voices. Consequently, inoculation should hamper or even annihilate the exemplification effect simply because the online comments lose their status as an exemplar.

Sign in to recut, update, or save this card.