Abstract
For the first time in history, civic discourse commonly occurs in digital environments in which algorithms influence exposure to social information1,2. It is increasingly important to understand whether and how these algorithms affect political discourse3,4,5. Here we built custom feed-ranking algorithms with full control over their features, and randomly assigned 2,000 participants to use them for 8 weeks (before and after the 2024 US presidential election). We tested whether an engagement-based algorithm (used on major social media platforms6,7) amplifies intergroup, moralized and emotional (IME) information in ways that skew perceptions of social norms around political dialogue5,8, and whether it increased engagement with IME content and perceptions of partisan animosity (compared with a reverse-chronological feed9,10). We also developed and tested a ‘diversified extremity’ algorithm to reduce the influence of extreme users11,12,13 to improve the accuracy of social norm perception14,15,16 and reduce perceptions of partisan animosity. We found that engagement-based feeds amplified IME and toxic content relative to reverse-chronological feeds, with the largest increases in moral outrage and political content. Engagement-based feeds also reduced prescriptive norm perception accuracy (albeit in an unexpected direction) and increased perceived partisan animosity. However, they did not significantly alter users’ own engagement behaviours. The diversified extremity algorithm reduced IME and toxic content exposure, improved prescriptive norm accuracy, yet maintained comparable platform enjoyment—suggesting that reducing the influence of extreme users can curb algorithmic distortions without diminishing user experience.
protocol registration The Stage 1 protocol for this Registered Report was accepted in principle on 17 September 2024. The protocol, as accepted by the journal, can be found at https://osf.io/c9a3m.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$32.99 / 30 days
cancel any time
Subscribe to this journal
Receive 52 print issues and online access
$199.00 per year
only $3.83 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to the full article PDF.
USD 39.95
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
All raw data and materials are available at the Open Science Framework (https://osf.io/k2crm/) and GitHub (http://github.com/METResearchGroup/bluesky-research).
Code availability
All code for our data collection, storage and custom-feed ranking algorithms is available at our GitHub (http://github.com/METResearchGroup/bluesky-research). R scripts for the power analysis and data analyses are available at the Open Science Framework (https://osf.io/k2crm/).
References
Shmargad, Y. & Klar, S. Sorting the news: how ranking by popularity polarizes our politics. Polit. Commun. 37, 423–446 (2020).
Bavel, J. J. V., Rathje, S., Harris, E., Robertson, C. & Sternisko, A. How social media shapes polarization. Trends Cogn. Sci. 25, 913–916 (2021).
Carpenter, J., Brady, W., Crockett, M., Weber, R. & Sinnott-Armstrong, W. Political polarization and moral outrage on social media. Conn. L. Rev. 52, 1107 (2020).
Wilson, A. E., Parker, V. A. & Feinberg, M. Polarization in the contemporary political and media landscape. Curr. Opin. Behav. Sci. 34, 223–228 (2020).
Brady, W. J., Jackson, J. C., Lindström, B. & Crockett, M. J. Algorithm-mediated social learning in online social networks. Trends Cogn. Sci. https://doi.org/10.1016/j.tics.2023.06.008 (2023).
Braghieri, L., Levy, R. & Trachtman, H. Frictions in news consumption: evidence from social media. Preprint at SSRN https://doi.org/10.2139/ssrn.5494670 (2025).
Brady, W., Jackson, J., Doyle, M. & Baier, S. Engagement-based algorithms disrupt human social norm learning. Preprint at OSF https://doi.org/10.31219/osf.io/mgdwq_v3 (2025).
Centola, D. Change: How to Make Big Things Happen (Little, Brown Spark, 2021).
Guess, A. M. et al. How do social media feed algorithms affect attitudes and behavior in an election campaign? Science 381, 398–404 (2023).
Milli, S. et al. Engagement, user satisfaction, and the amplification of divisive content on social media. PNAS Nexus 4, pgaf062 (2025).
Papakyriakopoulos, O., Serrano, J. C. M. & Hegelich, S. Political communication on social media: a tale of hyperactive users and bias in recommender systems. Online Soc. Netw. Media 15, 100058 (2020).
McClain, C., Widjaya, R., Rivero, G. & Smith, A. The Behaviors and Attitudes of U.S. Adults on Twitter (Pew Research, 2021).
Bail, C. Breaking the Social Media Prism: How to Make Our Platforms Less Polarizing (Princeton Univ. Press, 2021).
Fernbach, P. M. & Van Boven, L. False polarization: cognitive mechanisms and potential solutions. Curr. Opin. Psychol. 43, 1–6 (2022).
Lees, J. & Cikara, M. Inaccurate group meta-perceptions drive negative out-group attributions in competitive contexts. Nat. Hum. Behav. 4, 279–286 (2020).
Brady, W. J. et al. Overperception of moral outrage in online social networks inflates beliefs about intergroup hostility. Nat. Hum. Behav. https://doi.org/10.1038/s41562-023-01582-0 (2023).
Levy, R. Social media, news consumption, and polarization: evidence from a field experiment. Am. Econ. Rev. 111, 831–870 (2021).
Yesilada, M. & Lewandowsky, S. Systematic review: YouTube recommendations and problematic content. Internet Policy Rev. 11, 1652 (2022).
Tucker, J. A. et al. Social media, political polarization, and political disinformation: a review of the scientific literature. Preprint at SSRN https://doi.org/10.2139/ssrn.3144139 (2018).
Terren, L. & Borge-Bravo, R. Echo chambers on social media: a systematic review of the literature. Rev. Commun. Res. 9, 99–118 (2021).
Smith, J. R. & Louis, W. R. Do as we say and as we do: the interplay of descriptive and injunctive group norms in the attitude–behaviour relationship. Br. J. Soc. Psychol. 47, 647–666 (2008).
Burton, J. Algorithmic Amplification for Collective Intelligence (CBS, 2023).
Törnberg, P., Valeeva, D., Uitermark, J. & Bail, C. Simulating social media using large language models to evaluate alternative news feed algorithms. Preprint at https://doi.org/10.48550/arXiv.2310.05984 (2023).
Carmines, E. G., Ensley, M. J. & Wagner, M. W. Who fits the left-right divide? Partisan polarization in the American electorate. Am. Behav. Sci. 56, 1631–1653 (2012).
Huszár, F. et al. Algorithmic amplification of politics on Twitter. Proc. Natl Acad. Sci. USA 119, e2025334119 (2022).
Thorp, H. H. & Vinson, V. Context matters in social media. Science 385, 1393–1393 (2024).
Guess, A. M. et al. Reshares on social media amplify political news but do not detectably affect beliefs or opinions. Science 381, 404–408 (2023).
Nyhan, B. et al. Like-minded sources on Facebook are prevalent but not polarizing. Nature 620, 137–144 (2023).
Krosnick, J. A. & Petty, R. E. in Attitude Strength: Antecedents and Consequences (eds Petty, R. E. & Krosnick, J. A.) 1–24 (Lawrence Erlbaum, 1995).
Luttrell, A. & Sawicki, V. Attitude strength: distinguishing predictors versus defining features. Soc. Pers. Psychol. Compass 14, e12555 (2020).
Gelfand, M. J., Gavrilets, S. & Nunn, N. Norm dynamics: interdisciplinary perspectives on social norm emergence, persistence, and change. Ann. Rev. Psychol. 75, 341–378 (2024).
Bluesky user counter. bsky-users.theo.io https://bsky-users.theo.io/ (accessed 1 September 2024).
Budak, C., Nyhan, B., Rothschild, D. M., Thorson, E. & Watts, D. J. Misunderstanding the harms of online misinformation. Nature 630, 45–53 (2024).
Perspective API. perspectiveapi.com https://perspectiveapi.com/ (accessed 25 September 2024).
Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. BERT: pre-training of deep bidirectional transformers for language understanding. In Proc. 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Vol. 1 (Long and Short Papers) 4171–4186 (Association for Computational Linguistics, 2019).
Finkel, E. J. et al. Political sectarianism in America. Science 370, 533–536 (2020).
Luttrell, A. & Togans, L. J. The stability of moralized attitudes over time. Pers. Soc. Psychol. Bull. 47, 551–564 (2021).
Beknazar-Yuzbashev, G., Jiménez-Durán, R., McCrosky, J. & Stalinski, M. Toxic content and user engagement on social media: evidence from a field experiment. Preprint at SSRN https://doi.org/10.2139/ssrn.5130929 (2025).
Vallone, R. P., Ross, L. & Lepper, M. R. The hostile media phenomenon: Biased perception and perceptions of media bias in coverage of the Beirut massacre. J. Pers. Soc. Psychol. 49, 577–585 (1985).
Matias, J. N. Preventing harassment and increasing group participation through social norms in 2,190 online science discussions. Proc. Natl Acad. Sci. USA 116, 9785–9789 (2019).
Bagchi, C. et al. Social media algorithms can curb misinformation, but do they? Zenodo https://doi.org/10.5281/zenodo.13787981 (2024).
Scott, K. BlueSky benefits from X Brazilian ban as users hit nine million. Tech.co (9 September 2024).
Felicity, T. Bluesky warns of tech outage after usage traffic surges in brazil following X Ban. Venture Capital Post (4 September 2024).
Neely, S. R. Politically motivated avoidance in social networks: a study of Facebook and the 2020 presidential election. Soc. Media Soc. 7, 20563051211055438 (2021).
Boulianne, S. & Larsson, A. O. Engagement with candidate posts on Twitter, Instagram, and Facebook during the 2019 election. N. Media Soc. 25, 119–140 (2023).
Bluesky Social. feed-generator: ATProto Feed Generator Starter Kit. GitHub https://github.com/bluesky-social/feed-generator (2024).
Twitter. Twitter’s recommendation algorithm. blog.x.com https://blog.x.com/engineering/en_us/topics/open-source/2023/twitter-recommendation-algorithm (2023).
Zuckerberg, M. et al. Dynamically providing a news feed about a user of a social network. US patent US7669123B2 (2010).
Yang, J. et al. Mixed negative sampling for learning two-tower neural networks in recommendations. In Companion Proceedings of the Web Conference 2020 (eds El Fallah Seghrouchni, A. et al.) 441–447 (Association for Computing Machinery, 2020).
Covington, P., Adams, J. & Sargin, E. Deep neural networks for YouTube recommendations. In Proceedings of the 10th ACM Conference on Recommender Systems (eds Sen, S. et al.) 191–198 (Association for Computing Machinery, 2016).
Muja, M. & Lowe, D. G. Scalable nearest neighbor algorithms for high dimensional data. IEEE Trans. Pattern Anal. Mach. Intell. 36, 2227–2240 (2014).
Liu, T., Moore, A., Yang, K. & Gray, A. An investigation of practical approximate nearest neighbor algorithms. In Advances in Neural Information Processing Systems vol. 17 (eds Saul, L. K., Weiss, Y. & Bottou, L.) https://proceedings.neurips.cc/paper_files/paper/2004/file/1102a326d5f7c9e04fc3c89d0ede88c9-Paper.pdf (MIT Press, 2005).
Quelle, D. & Bovet, A. Bluesky: network topology, polarization, and algorithmic curation. PLoS ONE 20, e0318034 (2025).
Bail, C. A. et al. Exposure to opposing views on social media can increase political polarization. Proc. Natl Acad. Sci. USA 115, 9216–9221 (2018).
Stats for Bluesky by Jaz (jaz.bsky.social). bsky.jazco.dev https://bsky.jazco.dev/stats (accessed 25 September 2024).
Matsunaga, M. Familywise error in multiple comparisons: disentangling a knot through a critique of O’Keefe’s arguments against alpha adjustment. Commun. Methods Meas. 1, 243–265 (2007).
Guess, A., Munger, K., Nagler, J. & Tucker, J. How accurate are survey responses on social media and politics? Polit. Commun. 36, 241–258 (2019).
Brady, W. J. et al. Redesigning algorithms to intervene on social norm misperceptions during a national election [registered report stage 1 protocol]. OSF https://osf.io/c9a3m (2024).
Acknowledgements
The authors received funding from the Litowitz Center for Enlightened Disagreement and discretionary funds of W.J.B., E.J.F., N.K., J.T. and J.C.J. We thank the members of the Kellogg Political Psychology group (FLOPP), the Mind and Technology Lab and the Kellogg Computational Social Science Lab for feedback on previous versions of the manuscript.
Author information
Authors and Affiliations
Contributions
W.J.B. developed the research with input from A.E., E.J.F., J.C.J., N.K., V.P., C.P., T.S., J.T. and M.T.; W.J.B., M.D. and M.T. collected data and W.J.B. and M.T. performed research. W.J.B. and M.T. conducted formal analyses with input from A.E., E.J.F., J.C.J., N.K., C.P., J.T. and M.T.; W.J.B. wrote the original draft of the manuscript and A.E., E.J.F., J.C.J., N.K., C.P., T.S., J.T. and M.T. contributed to revisions.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature thanks Rafael Jimenez Duran and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Supplementary information
Supplementary Information (download PDF )
Detailed documentation of participant procedures, platform details, algorithm development, pilot testing and comprehensive statistical results for all research questions. It contains Supplementary Information 1–9, Supplementary Figs. 1–46 and Supplementary Tables 1–139.
Supplementary Information (download PDF )
An overview of Supplementary Information 1–9, and the titles and legends for the Supplementary Figures and Supplementary Tables.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Brady, W.J., Doyle, M., Elnakouri, A. et al. Redesigning algorithms to intervene on social norm misperceptions during a national election. Nature (2026). https://doi.org/10.1038/s41586-026-10536-1
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41586-026-10536-1
