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

The study investigates how belief systems shape persuasive communication in online discussions about 9/11 conspiracy theories. Building on evidence that conspiracist ideation reflects a broader worldview marked by generalized mistrust and rejection of official narratives, the authors ask whether conspiracist commenters focus more on refuting official explanations than advocating specific alternatives, compared with conventionalist commenters. They also examine whether conspiracists express more mistrust, reference other conspiracy theories more favorably, avoid the stigmatized label "conspiracy theory" for their own beliefs, and whether conventionalist comments exhibit greater hostility. The context is the widespread online discourse surrounding 9/11, with implications for understanding how belief systems guide persuasion in naturalistic settings.

Prior work shows conspiracy beliefs correlate with mistrust of authorities and others, feelings of powerlessness, lower self-esteem, superstition and paranormal beliefs, schizotypy, perceived lack of control, Machiavellianism, and certain personality traits. Conspiracist ideation is proposed to reflect a broader worldview wherein beliefs in disparate conspiracies intercorrelate, even when contradictory, suggesting adherence to higher-order beliefs (e.g., generalized opposition to official accounts). Scholars note that many conspiracy arguments emphasize negation of official narratives rather than coherent alternative plots. Online discourse research indicates that internet forums provide fertile ground for fringe views and candid expression due to anonymity, and that online discussion can reflect public concern and emotional climate. Projection processes in persuasion suggest people generate arguments they themselves find convincing, making persuasive comments informative about underlying beliefs. The 9/11 Truth Movement offers a robust case for study due to its online prominence and internal diversity of claims, alongside active conventionalist opposition in skeptic communities.

Design: Archival content analysis of persuasive comments posted to news websites about 9/11 between July 1 and December 31, 2011. Data sources: ABC News, CNN, Daily Mail, and the Independent. Articles were identified via site search (where available) or Google News using terms including: "9/11," "11/9," "September 11th," "11th September," "world trade center/centre," "wtc," "al-qaeda," "shanksville," and "building 7." Inclusion/exclusion of comments: From each article’s public comments, only original, persuasive comments were extracted. Excluded were comments consisting solely of insults/threats, solely meta-discussion about commenters/moderation, link-only posts or links with minimal description, and text copied verbatim from external sources (only original portions were coded). Sample: 2,174 comments from 1156 unique authors (321 posted more than once). Site/thread counts: ABC 65 comments (15 threads), CNN 632 (29), Daily Mail 1006 (64), Independent 471 (27). Approximately 2:1 conspiracist to conventionalist comments across sites. Coding: Primary tone (conspiracist vs. conventionalist). Counts of other, non-9/11 conspiracy theories mentioned favorably and unfavorably (excluding superconspiracies subsuming 9/11). Presence of expressions of mistrust; presence of expressions of powerlessness. Rhetorical structure: advocacy of the commenter’s favored interpretation (binary); derogation of the opposing interpretation (binary); whether the comment directly stated an explanation for what happened (binary). Use of the term "conspiracy theory" (nominal: not used; applied to opposing interpretation; applied to own interpretation; both; disputed its applicability). Hostility toward the opposing side (1–5 scale; insults, threats, sarcasm, accusations of complicity, etc.). Author identifiers, article URLs, and reply status were recorded. Data preparation: Analyses were performed at the comment level; parallel author-level analyses (means per author) replicated comment-level results. Only comment-level findings are reported. Inter-rater reliability: A random 10% of comments were double-coded. Tone κ = 0.84. For comments with agreed tone: advocacy κ = 0.64; derogation κ = 0.61; "conspiracy theory" usage κ = 0.70; mistrust κ = 0.49; direct explanatory statements κ = 0.55. Hostility ICC = 0.72. Powerlessness was too rare (n=2) to assess reliability. Statistical analyses: Group comparisons via t-tests and chi-square tests as appropriate; descriptive statistics tabulated.

Sample composition: Of 2,174 comments, 1,459 conspiracist and 715 conventionalist. Site proportions of conspiracist vs conventionalist were similar (about 2:1), χ²(3) = 1.514, p = 0.68. Cross-theory references: Conspiracist comments mentioned more unrelated conspiracy theories favorably than conventionalist comments (M = 0.12 vs 0.02 per comment), t(2172) = 3.82, p < 0.001; and fewer such theories unfavorably (M = 0.02 vs 0.18), t(2172) = -7.51, p < 0.001. Mistrust: Conspiracist comments more often expressed mistrust (10.6%) than conventionalist comments (1.4%), χ²(1) = 57.22, p < 0.001. Powerlessness: Only two comments expressed powerlessness; hypothesis could not be tested. Rhetorical structure: Advocacy of own explanation appeared in 31% of conspiracist vs 56% of conventionalist comments, χ²(1) = 121.69, p < 0.001. Derogation of opposing explanation appeared in 64% of conspiracist vs 44% of conventionalist comments, χ²(1) = 80.13, p < 0.001. Direct explanatory statements were more common in conspiracist comments (52%) than in conventionalist comments (19%), χ²(1) = 53.56, p < 0.001. Hostility: Conventionalist comments were more hostile (M = 2.08, SD = 1.02) than conspiracist comments (M = 1.44, SD = 0.79), t(2172) = 16.22, p < 0.001. Use of the label "conspiracy theory": Conspiracists rarely self-applied the label (2.1%); more often applied it to the opposing (official) account (4.3%) or disputed its applicability (4.5%). Conventionalists frequently labeled opposing beliefs as conspiracy theories (23.2%) and almost never self-applied (0.1%) or disputed (0.1%); none labeled both sides as conspiracy theories.

Findings align with the notion of a conspiracist worldview centered on generalized rejection of official narratives. Conspiracist commenters preferentially derogated the official account rather than advocating specific alternatives, while conventionalist commenters more often advocated their own (official) explanation. Conspiracists also expressed more mistrust and referenced other conspiracy theories more favorably, conceptually replicating prior evidence that conspiracy beliefs intercorrelate and are associated with lower trust. Although conspiracists more frequently issued direct explanatory statements (e.g., "inside job"), these often functioned as slogans or as conclusions following refutations of the official account rather than as elaborated alternative theories. Conventionalist comments exhibited greater hostility, consistent with majority-norm enforcement in social influence dynamics. The pattern of reluctance to self-apply the term "conspiracy theory" and disputation of its use supports the view that the label is stigmatized and considered derogatory or vague by conspiracists. Alternative interpretations include rhetorical strategy: commenters may select favorable or unfavorable analogies (e.g., JFK vs. moon landing) to bolster plausibility or discredit the opposing side, and some commenters present themselves as nominal conventionalists or conspiracists to increase persuasive impact.

This study shows that in naturalistic online discussions about 9/11, conspiracist persuasion emphasizes opposition to official narratives, heightened mistrust, and favorable cross-referencing of other conspiracy theories, while conventionalist persuasion emphasizes advocacy of the mainstream explanation and exhibits higher hostility. The reluctance to self-apply the term "conspiracy theory" underscores its social stigma in discourse. These results extend survey-based findings to real-world persuasive communication and illuminate how belief systems guide rhetoric online. Future research could experimentally test causality and audience effects, examine broader platforms and time periods, analyze the persuasive effectiveness of different rhetorical strategies, and refine measurement of constructs like powerlessness that were rare in this dataset.

Archival, observational design limits causal inference; persuasive comments may reflect strategic self-presentation and audience adaptation rather than underlying beliefs. The sample is limited to four mainstream news sites within a six-month period around the 10th anniversary of 9/11, potentially affecting generalizability. Excluding insult-only, meta-only, link-only, and copied comments may omit relevant expressions of hostility or belief. Inter-rater reliability for some variables (e.g., mistrust, direct explanatory statements) was moderate. Powerlessness could not be analyzed due to rarity. Although author-level analyses were conducted, only comment-level results are reported. Anonymity and lack of demographic data prevent assessment of user characteristics and selection effects.