Cognitive dissonance and motivated reasoning
Cognitive dissonance and motivated reasoning is a barrier to human progress within the psychological factors. The evidence for it, and the methods proposed to address it, follow.
Festinger (1957) described the discomfort people feel when their beliefs or actions conflict. People often reduce it by changing a belief or rejecting conflicting evidence. The theory's details are debated (rated C in the strength ratings).
Evidence
- Replication. A preregistered replication ran in 39 labs with nearly 4,900 people. It found that writing an essay against one's own view shifted attitudes. But the same study found that having a choice about writing it made no difference. This contradicts the classic prediction (Vaidis et al., 2024).
- Related prediction. The selective exposure meta-analysis, described under cognitive biases, supports a related prediction. That prediction is that people avoid information that conflicts with their beliefs (Hart et al., 2009).
- Replication authors' reply. The authors of the replication replied to commentaries. They said the wider dissonance literature shows widespread weaknesses, such as underpowered studies (Sleegers et al., 2024).
What can be done
Actions rated on the strength-rating scale (A strong and replicated, B solid but limited, C weak or debated, D contested or failed), applied to the specific claim made.
- Correct misinformation without fear of backfire (A). Across five experiments with more than 10,100 people on 52 issues, corrections did not backfire (Wood and Porter, 2019). A review concluded that backfire effects are not robust (Swire-Thompson, DeGutis, and Lazer, 2020). Corrections change factual beliefs more than attitudes.
- Deep canvassing (B). Swapping personal stories without judgment reduced exclusionary attitudes for at least four months (d = 0.04 to 0.08), while arguments alone had no effect (Kalla and Broockman, 2020).
- Accuracy prompts on social media (B). Getting people to think about accuracy improved the quality of news they shared (Pennycook et al., 2021). Across 20 experiments, prompts cut sharing of false headlines by 10% (Pennycook and Rand, 2022). An independent replication found a smaller effect (Roozenbeek, Freeman, and van der Linden, 2021).
- Teach lateral reading, which means leaving a site to check who is behind it (B). Six lessons improved students' ability to judge online sources in one school district (Wineburg et al., 2022). Lateral reading is one of nine tools in an expert review of ways to fight misinformation (Kozyreva et al., 2024). A 2025 meta-analysis coded post-test scores of 5.06 versus 2.83 points on a 14-point measure (g = 0.79, not adjusted for classroom clustering). The same meta-analysis notes that the lateral reading group was slightly ahead before the lessons (2.93 versus 2.27) (Fendt, Muth, and Edelsbrunner, 2025).
- Registered reports in science (B). Only 44% of registered reports had positive results, against 96% of standard papers. This suggests the format reduces publication of false positives (Scheel, Schijen, and Lakens, 2021).
- Weak or not shown to work: Prebunking videos helped people recognize manipulation techniques in tests, some of them run with a platform partner (C) (Roozenbeek et al., 2022). But gamified prebunking mostly made people more skeptical of all news, true and false (D) (Modirrousta-Galian and Higham, 2023, Graham et al., 2023, Seabrooke, Modirrousta-Galian, and Higham, 2025). Adversarial collaboration has no controlled outcome evidence, only participants' reports (C) (Isch, Tetlock, and Clark, 2025). The findings of the main meta-analysis of devil's advocacy could not be confirmed (Schwenk, 1990).
Proposed and experimental methods
Methods that are proposed, under trial, approved in some places, or tried and then failed. Each shows a stage label and an evidence rating. A stage label shows how far a method has progressed, not whether it works. The stage labels are explained on the psychological factor page.
- Crowd-sourced fact-check notes shown only when people who usually disagree rate them helpful (Large trial, B). This "bridging" system runs on X as Community Notes. Meta said in January 2025 that it would replace U.S. third-party fact-checking with a similar system (Associated Press, 2025). Across 40,078 posts on X, notes cut reposts by 46.1% and views by 13.5% after being attached. But over a post's whole life, notes cut them by only 11.6% and 5.5%, because notes arrive late (Slaughter et al., 2025). An independent study of 237,180 posts found a 61.2% drop in later spread and 94.3% higher odds that authors deleted the post. That study found these effects were weaker for influential accounts and political content. It also found a 14.9% fall in total engagement (Chuai et al., 2026).
- Reranking social media feeds to show less partisan hostility (Early trial, B). The bridging principle has been proposed for ranking whole feeds (Ovadya and Thorburn, 2023). A preregistered 10-day experiment with 1,256 X users ran during the 2024 U.S. campaign without platform cooperation. It used an AI model to show fewer or more posts with antidemocratic attitudes and partisan animosity. This shifted feelings toward the other party by more than 2 points on a 100-point scale (Piccardi et al., 2025). Switching Facebook and Instagram users to chronological feeds for 3 months did not change polarization (Guess et al., 2023). A simulation on Facebook data found that simple bridging ranking can push down content on controversial topics (C) (Milli et al., 2024).
- Short depolarization treatments tested in a megastudy (Large trial, B for short-term effects, C for lasting effects). A megastudy tested 25 treatments with 32,059 Americans. It found that many reduced partisan animosity. The strongest effects came from showing relatable people from the other side or shared identities. The megastudy also found that several reduced support for undemocratic practices. The strongest effects there came from correcting misperceptions of rivals' views (Voelkel et al., 2024). A news summary reports that 23 of 25 reduced animosity, 6 reduced antidemocratic attitudes, and 4 increased them. The summary also reports that effects had substantially faded two weeks later (Fried, 2026).
- AI conversations to reduce false beliefs (Early trial, C). A person states a belief, and an AI model argues against it with tailored evidence. Among 2,190 conspiracy believers, a dialogue with GPT-4 Turbo reduced belief by about 20%. The effect remained 2 months later, and belief in unrelated conspiracies also fell (Costello, Pennycook, and Rand, 2024). But in June 2026 the journal issued an expression of concern. This came after the authors reported problems reproducing some values from the public data (the authors say their reanalysis matches the original findings) (Retraction Watch, 2026). A month-long preprint study with 67 people found that AI help raised accuracy by 21% during assisted sessions. The same study found that unassisted performance on new items fell 15.3% by week 4 (Rani et al., 2025). The same skill can be misused. In debates with 900 people, GPT-4 with access to personal data was more persuasive than humans 64.4% of the time when the two differed. However, a September 2026 author correction found that the added advantage of personal data over GPT-4 without it was not statistically significant (Salvi et al., 2025).
- Intellectual humility training (Early trial, C). It aims to make people aware of the limits of their knowledge and value others' views. In one of four studies, making a growth mindset of intelligence salient increased intellectual humility. That in turn increased openness to opposing views (Porter and Schumann, 2018). But three main studies with 1,740 adolescents, including a randomized experiment, found a different result. In them, collaborative dialogue about ethics and politics did not increase intellectual humility on self-report, behavioral, or text measures (Prinzing and Vazquez, 2026).
- Large adversarial collaborations (Early trial, C). This builds on the adversarial collaboration entry above (Isch, Tetlock, and Clark, 2025). In one large example, proponents of two leading theories of consciousness preregistered rival predictions with a neutral team. They tested the predictions in 256 people with three brain imaging methods. The results substantially challenged key tenets of both theories (Cogitate Consortium et al., 2025). There is no evidence yet on whether the theories' proponents change their views.
Sources cited on this page
- Festinger, L. (1957). A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson. C Limited: D for the effect of choice in induced compliance
- Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009). Feeling validated versus being correct: a meta-analysis of selective exposure to information. Psychological Bulletin, 135(4), 555-588. DOI B Moderate
- Vaidis, D. C., Sleegers, W. W. A., van Leeuwen, F., et al. (2024). A multilab replication of the induced-compliance paradigm of cognitive dissonance. Advances in Methods and Practices in Psychological Science, 7(1), 1-26. DOI A Strong
- Guess, A., & Coppock, A. (2020). Does counter-attitudinal information cause backlash? Results from three large survey experiments. British Journal of Political Science, 50(4), 1497-1515. DOI B Moderate
- Gal, D., & Rucker, D. D. (2018). The loss of loss aversion: Will it loom larger than its gain? Journal of Consumer Psychology, 28(3), 497-516. link C Limited
- Graham, M. E., Skov, B., Gilson, Z., Heise, C., Fallow, K. M., Mah, E. Y., et al. (2023). Mixed news about the Bad News Game. Journal of Cognition, 6(1), 58. link B Moderate
- Isch, C., Tetlock, P. E., & Clark, C. J. (2025). Reflections on adversarial collaboration from the adversaries: Was it worth it? Theory and Society, 54(6), 929-946. link C Limited
- Kalla, J. L., & Broockman, D. E. (2020). Reducing exclusionary attitudes through interpersonal conversation: Evidence from three field experiments. American Political Science Review, 114(2), 410-425. link B Moderate
- Kozyreva, A., Lorenz-Spreen, P., Herzog, S. M., Ecker, U. K. H., Lewandowsky, S., Hertwig, R., et al. (2024). Toolbox of individual-level interventions against online misinformation. Nature Human Behaviour, 8(6), 1044-1052. link B Moderate
- Li, Y., & Bates, T. C. (2019). You can't change your basic ability, but you work at things, and that's how we get hard things done: Testing the role of growth mindset on response to setbacks, educational attainment, and cognitive ability. Journal of Experimental Psychology: General, 148(9), 1640-1655. link B Moderate
- Modirrousta-Galian, A., & Higham, P. A. (2023). Gamified inoculation interventions do not improve discrimination between true and fake news: Reanalyzing existing research with receiver operating characteristic analysis. Journal of Experimental Psychology: General, 152(9), 2411-2437. link B Moderate
- Pennycook, G., Epstein, Z., Mosleh, M., Arechar, A. A., Eckles, D., & Rand, D. G. (2021). Shifting attention to accuracy can reduce misinformation online. Nature, 592(7855), 590-595. link B Moderate
- Pennycook, G., & Rand, D. G. (2022). Accuracy prompts are a replicable and generalizable approach for reducing the spread of misinformation. Nature Communications, 13(1), 2333. link B Moderate
- Roozenbeek, J., Freeman, A. L. J., & van der Linden, S. (2021). How accurate are accuracy-nudge interventions? A preregistered direct replication of Pennycook et al. (2020). Psychological Science, 32(7), 1169-1178. link B Moderate
- Roozenbeek, J., van der Linden, S., Goldberg, B., Rathje, S., & Lewandowsky, S. (2022). Psychological inoculation improves resilience against misinformation on social media. Science Advances, 8(34), eabo6254. link C Limited
- Scheel, A. M., Schijen, M. R. M. J., & Lakens, D. (2021). An excess of positive results: Comparing the standard psychology literature with Registered Reports. Advances in Methods and Practices in Psychological Science, 4(2). link B Moderate
- Schwenk, C. R. (1990). Effects of devil's advocacy and dialectical inquiry on decision making: A meta-analysis. Organizational Behavior and Human Decision Processes, 47(1), 161-176. link C Limited: (findings not confirmed)
- Seabrooke, T., Modirrousta-Galian, A., & Higham, P. A. (2025). Re-examining the bad news game: No evidence of improved discrimination of Indian true and fake news headlines. Psychonomic Bulletin & Review, 33(1), 13. link B Moderate
- Swire-Thompson, B., DeGutis, J., & Lazer, D. (2020). Searching for the backfire effect: Measurement and design considerations. Journal of Applied Research in Memory and Cognition, 9(3), 286-299. link B Moderate
- Wineburg, S., Breakstone, J., McGrew, S., Smith, M. D., & Ortega, T. (2022). Lateral reading on the open Internet: A district-wide field study in high school government classes. Journal of Educational Psychology, 114(5), 893-909. link B Moderate
- Wood, T., & Porter, E. (2019). The elusive backfire effect: Mass attitudes' steadfast factual adherence. Political Behavior, 41(1), 135-163. link A Strong
- Chuai, Y., Pilarski, M., Renault, T., Restrepo-Amariles, D., Troussel-Clément, A., Lenzini, G., et al. (2026). Community-based fact-checking reduces the spread of misleading posts on X (formerly Twitter). Nature Communications, 17, 4070. link B Moderate
- Cogitate Consortium, Ferrante, O., Gorska-Klimowska, U., Henin, S., Hirschhorn, R., Khalaf, A., et al. (2025). Adversarial testing of global neuronal workspace and integrated information theories of consciousness. Nature, 642(8066), 133-142. link B Moderate
- Costello, T. H., Pennycook, G., & Rand, D. G. (2024). Durably reducing conspiracy beliefs through dialogues with AI. Science, 385(6714), eadq1814. link C Limited
- Fried, C. (2026, July 15). Reducing partisan hostility can be surprisingly simple. UCLA Anderson Review. link B Moderate
- Guess, A. M., Malhotra, N., Pan, J., Barberá, P., Allcott, H., Brown, T., et al. (2023). How do social media feed algorithms affect attitudes and behavior in an election campaign? Science, 381(6656), 398-404. link B Moderate
- Milli, S., Thorburn, L., Bouchaud, P., Wang, Y., Garg, N., & Pierson, E. (2024). Bridging in social media feeds censors controversial topics [Workshop paper]. Pluralistic Alignment Workshop, NeurIPS 2024. link C Limited
- Ovadya, A., & Thorburn, L. (2023). Bridging systems: Open problems for countering destructive divisiveness across ranking, recommenders, and governance (arXiv:2301.09976). arXiv. link B Moderate
- Piccardi, T., Saveski, M., Jia, C., Hancock, J., Tsai, J. L., & Bernstein, M. S. (2025). Reranking partisan animosity in algorithmic social media feeds alters affective polarization. Science, 390(6776), eadu5584. link B Moderate
- Porter, T., & Schumann, K. (2018). Intellectual humility and openness to the opposing view. Self and Identity, 17(2), 139-162. link C Limited
- Prinzing, M., & Vazquez, M. (2026). Does collaborative dialogue increase intellectual humility about ethics and politics? American Psychologist. Advance online publication. link B Moderate
- Rani, A., Danry, V., Liang, P. P., Lippman, A. B., & Maes, P. (2025). Dialogues with AI reduce beliefs in misinformation but build no lasting discernment skills (arXiv:2510.01537). arXiv. link C Limited
- Salvi, F., Horta Ribeiro, M., Gallotti, R., & West, R. (2025). On the conversational persuasiveness of GPT-4. Nature Human Behaviour, 9(8), 1645-1653. link B Moderate
- Slaughter, I., Peytavin, A., Ugander, J., & Saveski, M. (2025). Community notes reduce engagement with and diffusion of false information online. Proceedings of the National Academy of Sciences, 122(38), e2503413122. link B Moderate
- Voelkel, J. G., Stagnaro, M. N., Chu, J. Y., Pink, S. L., Mernyk, J. S., Redekopp, C., et al. (2024). Megastudy testing 25 treatments to reduce antidemocratic attitudes and partisan animosity. Science, 386(6719), eadh4764. link B Moderate
- Fendt, M., Muth, X., & Edelsbrunner, P. A. (2025). Judging a text by its author: A meta-analysis of interventions to foster source credibility assessment. Learning and Individual Differences, 124, 102782. link B Moderate
Every source for this factor is listed on the psychological factor page.