What about building 7? A social psychological study of online discussion of 9/11 conspiracy theories

Recent research into the psychology of conspiracy belief has highlighted the importance of belief systems in the acceptance or rejection of conspiracy theories. We examined a large sample of conspiracist (pro-conspiracy-theory) and conventionalist (anti-conspiracy-theory) comments on news websites in order to investigate the relative importance of promoting alternative explanations vs. rejecting conventional explanations for events. In accordance with our hypotheses, we found that conspiracist commenters were more likely to argue against the opposing interpretation and less likely to argue in favor of their own interpretation, while the opposite was true of conventionalist commenters. However, conspiracist comments were more likely to explicitly put forward an account than conventionalist comments were. In addition, conspiracists were more likely to express mistrust and made more positive and fewer negative references to other conspiracy theories. The data also indicate that conspiracists were largely unwilling to apply the conspiracy theory label to their own beliefs and objected when others did so, lending support to the long-held suggestion that conspiracy belief carries a social stigma. Finally, conventionalist arguments tended to have a more hostile tone. These tendencies in persuasive communication can be understood as a reflection of an underlying conspiracist worldview in which the details of individual conspiracy theories are less important than a generalized rejection of official explanations.

The Internet was made for conspiracy theory: it is a conspiracy theory: one thing leads to another, always another link leading you deeper into no thing and no place.

Conspiracy theories, defined as allegations that powerful people or organizations are plotting together in secret to achieve sinister ends through deception of the public (Abalakina-Paap et al., 1999; Wood et al., 2012), have long been an important element of popular discourse. With the advent of the Internet, they have become more visible than ever. Although the psychological literature on conspiracy belief has a relatively short history, with most of the relevant research having been conducted only within the past twenty years, it has revealed a great deal regarding individual differences between those who generally believe conspiracy theories (whom we call conspiracists) and those who prefer conventional explanations (whom we call conventionalists). Conspiracy beliefs have been shown to be positively correlated with mistrust of other people (Goertzel, 1994) and authorities (Swami et al., 2010); feelings of powerlessness and low self-esteem (Abalakina-Paap et al., 1999); superstition, beliefs in the paranormal, and schizotypy (Darwin et al., 2011); a perceived lack of control (Hamsher et al., 1968; Whitson and Galinsky, 2008); a Machiavellian approach to social interaction (Douglas and Sutton, 2011); and openness to experience (Swami et al., 2010; but see Swami et al., 2011).

At the present time, questionnaire-based investigations of individual differences make up the bulk of the existing research, although experimental approaches are emerging (e.g., Douglas and Sutton, 2008; Jolley and Douglas, 2013). A fairly recent development in the field has been an acknowledgement that in addition to trait-like variables and transient psychological states, ideologies and broad belief systems play a substantial role in conspiracy theory belief. For example, in an examination of conspiracy theories regarding an alleged cover-up of the divinity of Mary Magdalene and the bloodline of Christ, Newheiser et al. (2011) demonstrated that the plausibility of these theories hinged largely on broader beliefs about the world. People with traditional Christian beliefs were likely to reject such theories out of hand, while those with a more New Age approach were much more receptive. In a similar vein, Lewandowsky et al. (2013b) demonstrated that rejection of climate science (though not explicitly conspiracist) is determined in part by ideological concerns, with libertarian free-market ideology, apparently predisposing people to believe that anthropogenic global warming is an unscientific hoax. It is clear, then, that individual conspiracy theories or related counter-normative explanations can seem more or less likely depending on how they comport with other beliefs held by the audience.

Some researchers have gone further, proposing the existence of a conspiracist worldviewa belief system conducive to conspiracy beliefs in general (e.g., Goertzel, 1994; Swami et al., 2010; Wood et al., 2012). This proposal stems primarily from the finding that beliefs in unrelated conspiracy theories tend to intercorrelate: for example, someone who believes that Princess Diana was deliberately assassinated is also more likely to believe that the moon landing was a hoax. Indeed, Wood et al. (2012) demonstrated that even beliefs in directly contradictory conspiracy theories were positively correlated with one another, indicating that conspiracy beliefs may be held together not by direct agreement with one another, but by mutual agreement with higher-order beliefs about the world. One particularly important element of the conspiracist worldview is thought to be a generalized opposition to official or received narratives. In this view, conspiracy belief is not about believing in particular alternative theories, but in disbelieving in whatever the official story is. This tendency has been informally noted by Dean (2002), who described most conspiracy theories as bits and pieces without a plot [that] fail to delineate any conspiracy at all. They simply counter conventional narratives with suspicions and allegations that, more often than not, resist coherent emplotment (p. 92). Likewise, Clarke (2007) observed that conspiracy theories are often extremely vague, particularly in the Internet age.

If this is the case, then for people who hold a conspiracist worldview, the specifics of a conspiracy theory are less important than its identity as a conspiracy and its opposition to the official explanation. The important element is that those in power are lying and cannot be trusted, and that they are covering up something sinister. Opposition to officialdom, in this sense, parallels the generalized prejudice that Adorno et al. (1950) found to be strong enough to overcome contradictions between different anti-Jewish stereotypes. More than being a specific belief that Jews are overly secretive or overly intrusive, anti-Semitism appears to be more of a general belief that Jews are generally unpleasant people. Likewise, conspiracy theory belief appears to be more of a negative belief than a positive oneit is more concerned with saying what the cause of a condition or event was not (i.e., whatever the official explanation is) than with putting forward a specific alternative account.

An opportunity to test this idea presents itself in the form of observation of online discourse. In spite of, or perhaps because of, the lack of mainstream public acceptance for their theories, many conspiracists, both prominent and otherwise, appear to see themselves as having a duty to spread their views to the public at large. They often exhort the unthinking masses to wake up (e.g., Crane, 2008; Byers, 2009; Icke, 2012). This is a reasonable reaction: given a belief that people's lives are being manipulated by malevolent forces beyond their control, most would probably agree that trying to spread the word about that fact is a good idea. Outspoken conventionalists, such as those in the skeptic movement (e.g., Randi, 1982; Sagan, 1995; Shermer, 1997; Novella, 2009), find most conspiracy theories to be misguided at best and destructive at worst, and so make a point of arguing against them in the public sphere.

This discussion is voluminous and highly visible in many arenas, perhaps none more so than news website comment sections. Articles about topics for which popular conspiracy theories exist, such as 9/11, the moon landing, and vaccines, can have tens of thousands of comments, most of which are devoted to advancing or refuting allegations of conspiracy. These comments are often archived along with the associated articles for months or years afterward, which provides an excellent opportunity for archival research to give some insight into the thoughts and beliefs of those writing them (e.g., Fat et al., 2012; Loke, 2012; Sisask et al., 2012).

The present study consists of an examination of a large number of conspiracy theory-related persuasive comments on news stories. Such analysis of online discourse as a method of examining psychological states has increased in prominence as the Internet has become a more popular place to discuss one's ideas. The subject and pace of online discussion has been shown to be a more or less reliable barometer of public concern over social issues (Roberts et al., 2002; Scharkow and Vogelgesang, 2011), and emotional reactions expressed online can be used to consistently predict political approval ratings (Gonzalez-Bailon et al., 2012). Quantitative analysis of online discussion has also been used to gain insight into the social psychology of groups with fringe views (Douglas et al., 2005), attitudes toward Tourette's Syndrome (Fat et al., 2012), and racial views (Loke, 2012). Qualitative research on online discourse has been more common, including a study demonstrating the evolution of conspiracy theories over time in response to evidence (Lewandowsky et al., 2013a). In the context of conspiracy theories in particular, there are several advantages to content analysis of online commentary. The self-selective nature of online communication allows for the collection of a great deal of data regarding opinions that may be held by only a minority of people; moreover, the degree of anonymity facilitates the honest expression of opinions that might not be held in high social esteem elsewhere (e.g., Douglas et al., 2005; Loke, 2012).

There are some caveats associated with analyzing persuasive comments in particular. While external validity may benefit from observing behavior in a naturalistic setting, there is some degree of uncertainty regarding the internal validity of any conclusions drawn from such methods. Most obviously, there is the issue of to what degree the content of persuasive communications reflects the properties of the author rather than the demands of the situation. Rather than faithful representations of internal psychological processes, commenters' methods of argumentation might instead reflect strategic considerations regarding the audience, the venue, and the subject matter. While self-presentation is very often a concern in psychological research, even in laboratory settings, such demands may be especially salient in a situation where one's goal is implicitly (or even explicitly) to persuade others rather than to provide an honest and straightforward account of one's beliefs. Indeed, some research has shown that people do adapt their persuasive techniques according to their knowledge of the audience and the subject (Friestad and Wright, 1999; Douglas et al., 2010; Vogel et al., 2010).

The question of whether we can expect persuasive communication to accurately reflect inner psychological processes is not easily answerable, as the effect of lay persuasive knowledge on generation of persuasive arguments is fairly sparse. While there is a substantial body of research on lay persuasive knowledge, the vast majority of it focuses instead on how such knowledge affects susceptibility to the persuasive messages of others. However, it is well-established that people tend to rely heavily on projection for predicting others' behaviorthat is, they use themselves as a model for prediction. This effect is especially strong when relatively little is known about the target [for a review, see Robbins and Krueger (2005)]. In general, then, it is likely that persuaders use the self as a model for argument generation: in other words, they argue in a way that they would themselves find convincing. This, in turn, suggests that the types of arguments used by persuaders can contain information relevant to understanding how they think about the issue at hand.

The tendency to use social projection is especially relevant in online settings. Much online discussion is either fully anonymous or conducted under pseudonyms, greatly limiting the amount of information available about the other party in a discussion. As such, we assume for the purposes of the present study that people will generally tend to use arguments that they themselves would find most convincing were they the audience rather than the persuader. This, in turn, should reflect the structure of their belief systemsthe arguments that people find most convincing are those that match up with how they view the world (Darwin et al., 2011; Newheiser et al., 2011; Wood et al., 2012; Lewandowsky et al., 2013b). To that end, we systematically coded and analyzed conspiracist and conventionalist persuasive comments from four major news websites on articles relating to 9/11 from the period of July 1st through December 31st, 2011, encompassing the months surrounding the tenth anniversary of the attacks.

9/11 conspiracy theories provide an excellent research subject for several reasons. First, the community associated with these theories, known as the 9/11 Truth Movement, is noted for its substantial online presence and focus on Internet proselytizing. Bartlett and Miller (2011) observed that the movement's mass membership backbone (p. 45) devotes a substantial amount of time to producing large numbers of online comments, and Clarke (2007) saw the Truth Movement as a paradigmatic example of Internet conspiracy culture. Second, the timing was fortuitous, with the tenth anniversary of the 9/11 attacks, sure to herald a number of stories on the subject and therefore many relevant comments, having occurred shortly before data collection commenced. The recency of the materials lowered the probability of comments having been expunged from archives or lost as an unintended consequence of comment software upgrades. Third, just as the Truth Movement has a substantial online presence, so too do its conventionalist opponents in the skeptic movement. We therefore expected that there would be a good deal of debate between the two sides, providing further raw materials for analysis. Finally, the Truth Movement is a well-established community with a substantial intellectual output, including popular books (e.g., Griffin, 2004), conference circuits, several sub-organizations such as Architects and Engineers for 9/11 Truth, and at least one peer-reviewed journal, the Journal of 9/11 Studies. There is substantial debate within the Truth Movement regarding whether 9/11 was a controlled demolition, a deliberate intelligence failure, or even the result of exotic space-based weaponry (Barber, 2008). In short, its body of work is varied, voluminous, and well-developed, and should therefore be able to provide a wide range of different arguments for analysis.

If our reasoning regarding the influence of projection on persuasive tactics holds, we should see systematic differences in the characteristics of conspiracist and conventionalist arguments. Specifically, we should be able to replicate earlier results demonstrating that unrelated conspiracy beliefs are intercorrelated (e.g., Goertzel, 1994; Swami et al., 2010; Wood et al., 2012)in this case, conspiracist comments should contain more positive (and fewer negative) references to unrelated conspiracy theories compared with conventionalist comments. Examining a long-standing correlate of conspiracy belief, we also investigated the degree to which comments contained explicit expressions of mistrust, predicting that conspiracist comments would be more likely to express mistrust of authorities or other targets than conventionalist comments (e.g., Wright and Arbuthnot, 1974; Simmons and Parsons, 2005). Further, we examined expressions of powerlessness, and predicted that conspiracist comments would express more concerns about power, as feelings of powerlessness have been shown to correlate reliably with conspiracy theory belief (Abalakina-Paap et al., 1999). Replicating the previously established relationships between conspiracy beliefs, trust, and power would increase confidence in the present study's methods and help to justify any novel results derived therefrom.

In addition to verifying the utility of this archival approach by replicating previous results, we made several novel predictions. First, if we are correct in our contention that much of the conspiracist worldview is based on a generalized rejection of official explanations rather than on positing particular alternative narratives, conspiracist comments should focus on refuting conventional explanations more than on presenting or supporting specific conspiracy theories. Therefore, conspiracist comments, relative to conventionalist comments, should be more likely to derogate rival explanations and less likely to promote their own. Second, we elected to examine the veracity of the long-held contention that conspiracy theory and conspiracy theorist carry an intellectual stigma (e.g., Bratich, 2002, 2008; Coady, 2006). If this is true, people should be unwilling to apply the term to themselves and should object when others do so. As such, we predicted that conspiracists would avoid applying the term conspiracy theory to their own beliefs (or conspiracy theorists to themselves), and would attempt to dispute others' usage of the term. While this might seem an obvious predictionand indeed many authors take it as a given that the term is stigmatizedto our knowledge there have not yet been any empirical investigations of this contention.

Finally, another possible avenue by which the spread of conspiracy theories could be fruitfully understood is social influence theory (Latané, 1981). Since 9/11 conspiracy theories are (at least in the West) an opinion held by a vocal minority attempting to effect change, social influence theory (Latané, 1981) would predict that conventionalists, if they are good majority influencers, are more likely to show patterns consistent with normative social influence. In particular, Bratich (2008) has highlighted the hostility of intellectual orthodoxy toward conspiracist explanations for events and the labelling of conspiracists as paranoid or otherwise mentally ill (c.f. Hofstadter, 1964; Kalichman et al., 2010). At the same time, conspiracists are often hostile in a different way, dismissing conventionalists as naïve, gullible, and either unwitting dupes or willing stooges of the conspiracy (Crane, 2008; Byford, 2011). Therefore, we examined the hostility of each persuasive comment, whether characterized by outright insults, threats, dismissive sarcasm, accusations of complicity, or other hostile or insulting content.