Here’s the short version from the research: people adopt and stick with conspiracy theories because these narratives meet psychological needs (to understand, to feel safe/in-control, and to belong to a valued group). Once adopted, several well-documented processes make the beliefs resistant to correction—even sometimes stronger when challenged.

Why people believe in conspiracy theories

• Three core motives (epistemic, existential, social). Conspiracy theories promise hidden “explanations” for confusing events (epistemic), restore a sense of control in threatening situations (existential), and flatter one’s in-group while derogating out-groups (social). See the broad review by Douglas, Sutton & Cichocka (2017).
  https://pmc.ncbi.nlm.nih.gov/articles/PMC5724570/. (PMC)
• Motivated / identity-protective reasoning. People assess evidence in ways that protect their group identity and existing worldview, selectively crediting supportive information and attacking disconfirming sources. See Kahan’s synthesis of identity-protective cognition.
  https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2973067. (SSRN)
• Cognitive tendencies that make conspiracies attractive. Pattern perception and agency detection under uncertainty; proportionality bias (big events “must” have big causes); lower tolerance for uncertainty; and, for some, a need for uniqueness (“I know what others don’t”).
  Van Prooijen (2018): https://pmc.ncbi.nlm.nih.gov/articles/PMC6282974/; Lantian et al. (2017): https://econtent.hogrefe.com/doi/abs/10.1027/1864-9335/a000306. (PMC)
• A “conspiracy mentality” / belief network. Endorsing one conspiracy often predicts endorsing others (a partially “monological” system), creating a self-reinforcing web of explanations. See Goertzel (1994) and newer examinations.
  Goertzel (1994): https://www.jstor.org/stable/3791630; Enders & colleagues (2021): https://jspp.psychopen.eu/index.php/jspp/article/view/5649/5649.html. (JSTOR)

Why corrections often fail (and can sometimes entrench beliefs)

• Continued-influence effect. Even after a correction is accepted, the original misinformation keeps shaping inferences unless the gap is filled with a coherent alternative explanation.
  Debunking Handbook 2020 (overview): https://digitalcommons.unl.edu/scholcom/1247/; background review: https://cssn.org/wp-content/uploads/2020/10/DB2020paper-1.pdf. (UNL Institutional Repository)
• Familiarity & repetition. Repeating a false claim (even to debunk it) can increase its familiarity and perceived truth, especially over time or when memory is taxed—hence the “illusory truth” problem.
  Prike et al. (2023): https://cognitiveresearchjournal.springeropen.com/articles/10.1186/s41235-023-00492-z. (SpringerOpen)
• Motivated counter-arguing / reactance. When a correction threatens one’s identity or group, people actively generate counter-arguments, question source motives, and feel “pushed,” which can harden attitudes—classic identity-protective cognition.
  Kahan (2017): https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3046603. (SSRN)
• “Backfire effect” — what the field now says. Early studies suggested corrections might increase misbelief (backfire). Larger, newer syntheses show true backfire is rare, though corrections often yield only partialimprovement and can fade.
  Nyhan (2021) on durability of misinfo & limited correction effects: https://www.pnas.org/doi/10.1073/pnas.1912440117; Swire-Thompson et al. (2022) test a canonical backfire prediction and do not find it: https://pmc.ncbi.nlm.nih.gov/articles/PMC9283209/; Ecker et al. (2023) fail to replicate familiarity/fear backfires: https://www.emc-lab.org/uploads/1/1/3/6/113627673/ecker.2023.plosonepp.pdf. (PNAS)

Putting it together (the mechanism in plain language)

1. Adoption: In moments of uncertainty, threat, or loss of control, a conspiracy story offers a simple, pattern-rich narrative that protects identity and elevates one’s in-group (and sometimes one’s own status via uniqueness). (PMC)
2. Shielding: Once inside the belief, identity-protective cognition filters new information—friendly sources are trusted, opposing sources are impugned; contradictions are re-framed as part of the plot (“that’s what they _want_you to think”). (SSRN)
3. Resistance to change: Corrections help, but because misinformation remains familiar and fills causal gaps, its influence lingers; if the correction feels identity-threatening, people counter-argue, which can leave attitudes largely intact or even more confidently held. True “backfire” is uncommon, but limited correction and attitude polarization are common. (UNL Institutional Repository)

Practical implications (evidence-based)

• Prebunk when possible. Pre-exposure “inoculation” and clear warnings reduce later susceptibility; when debunking, give a concise fact, then the myth once, then an alternative causal explanation. See Debunking Handbook 2020.
  https://digitalcommons.unl.edu/scholcom/1247/. (UNL Institutional Repository)
• Affirm identity, avoid zero-sum threats. Frame corrections in ways that don’t force a choice between “my group” and “the truth,” reducing defensive processing. (SSRN)
• Limit needless repetition of the myth and fill the causal gap so people aren’t left with a vacuum that the conspiracy neatly filled. (Brown Climate Social Science Network)