Misinformation in and about science
Open accessThe DOI matches its Crossref record; 142 of the 144 references it cites were found in a registry or an index; 19 of the 105 sources its claims cite were checked for support.
Abstract
Humans learn about the world by collectively acquiring information, filtering it, and sharing what we know. Misinformation undermines this process. The repercussions are extensive. Without reliable and accurate sources of information, we cannot hope to halt climate change, make reasoned democratic decisions, or control a global pandemic. Most analyses of misinformation focus on popular and social media, but the scientific enterprise faces a parallel set of problems—from hype and hyperbole to publication bias and citation misdirection, predatory publishing, and filter bubbles. In this perspective, we highlight these parallels and discuss future research directions and interventions.
Overview
CoChat Summary
TL;DR
This perspective argues that science faces misinformation problems parallel to those in public media—hype, publication bias, citation misdirection, predatory publishing, and filter bubbles—driven by attention and productivity incentives. It proposes incentive reform, better detection and evaluation, and data-reasoning education, while maintaining that science remains a highly trusted, self-correcting enterprise.
Key findings
- Misinformation in science arises from misaligned incentives, out-of-date publishing norms, and sociotechnical systems that concentrate attention and credit on a small subset of the literature; the paper highlights hype/hyperbole, publication bias, citation misdirection, predatory publishing, filter bubbles/echo chambers, and data distortion.
- Publication bias is substantial: only 15% of published results across the sciences are negative, and in antidepressant trials, 94% of published trials reported positive efficacy results versus 51% of FDA-registered trials yielding definitively positive results on original outcome measures.
- Citation misdirection is common: most analyses find between 1 in 5 and 1 in 10 citations support claims incongruous with the cited paper; a five-sentence opioid letter was cited over 600 times, with fewer than 20% of citations acknowledging its hospital-setting restriction; and 92% of radiation oncology articles citing retracted articles after retraction cited them as if no retraction had occurred.
- Predatory publishers produced nearly half a million articles in 2014, bringing in around $74 million in publication fees, compared with an estimated $250 million annual market for reputable open access journals and about 2.5 million Web of Science articles in 2014; predatory publishing likely comprises about 5 to 10% of the literature.
Methods
- Perspective/synthesis article; the authors state “There are no new data associated with this article.”
- The paper draws on published literature, news examples, and prior studies to build a conceptual argument about misinformation in and about science.
- Main analysis is interpretive and agenda-setting rather than systematic review or original empirical analysis.
Limitations
- The paper does not include a formal limitations section.
- Evident from the text: as a perspective, it relies on illustrative examples and cited studies rather than a systematic review, and some claims are explicitly speculative (e.g., “we surmise that most fields are susceptible”).
- The authors also state that the magnitude of publication bias is not fully known: “How bad is the problem? We do not really know.”
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
10 Valid · 3 Not valid · 59 Not checked · 33 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“endangers the well-being of the planet (3)”
What the sources say
- Partly supportedcites [3] sourceExperimental — accuracy not yet measured
The Introduction says climate-change disinformation campaigns have undermined public understanding and limited societal engagement with an issue that requires large-scale changes in behavior. This supports a risk to efforts to address climate change, but the paper does not establish that misinformation generally endangers the planet’s well-being. (LLM, conf=0.78)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“One analysis found that nearly a third of 525 papers in top obesity or nutrition journals make inappropriate causal claims in their abstracts or titles (37)”
What the sources say
- Supportedcites [37] S. S. Cofield, R. V. Corona, D. B. Allison, Use of causal language in observational studies of obesity and nutrition. Obesity facts 3 , 353-356 (2010). source
The Abstract reports that the authors scrutinized titles and abstracts of 525 observational papers in four leading obesity and nutrition journals and found misleading causal language in 161 (31%). The Discussion likewise describes inappropriate causal language in the abstracts or titles of almost one third of the reports. (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
None of the sources cited for this claim have been checked yet.
In the paper
“there is little evidence to date that it will appreciably reduce publication bias (61-63).”
What the sources say
- Not checked yetcites [61] National Academies of Sciences, Engineering, and Medicine, Reproducibility and replicability in science (2019). https://www.nationalacademies.org/ourwork/reproducibility-and-replicability-in-science. Accessed 7 December 2020.
- Could not verifycites [62] R. M. Shiffrin, K. B ¨ orner, S. M. Stigler, Scientific progress despite irreproducibility: A seeming paradox. Proc. Natl. Acad. Sci. U.S.A. 115 , 2632-2639 (2018). sourceExperimental — accuracy not yet measured
Our copy does discuss preregistration in '4. Targeting Remedies' (p. 19), but only as a proposal whose value 'might well differ for translational and exploratory research' and with a caution that it 'not reduce creativity in exploration' — it makes no statement about the state of evidence for preregistration's effect on publication bias. The paper's treatment of publication-type bias is separate: 'selective reporting of positive rather than negative results' (2.1, p. 5) and editorial rejection of non-replications (p. 19), neither linked to preregistration. Searched: preregistration + publication bias, selective reporting/evidence of remedy effectiveness, file drawer/negative results — no passage addresses the claim's point, so it can be neither confirmed nor contradicted. (LLM, conf=0.70)
- Not checked yetcites [63] S. Goldin-Meadow, Why preregistration makes me nervous. APS Observer 29 , 5-6 (2016).
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Depending on the field and the methodology, most analyses of the problem reveal that between 1 in 5 and 1 in 10 citations are used to support claims incongruous with the results of the cited paper (refs. 74-76 and references therein).”
What the sources say
- Partly supportedcites [74] P. A. Todd, J. R. Guest, J. Lu, L. M. Chou, One in four citations in marine biology papers is inappropriate. Mar. Ecol. Prog. Ser. 408 , 299-303 (2010). sourceExperimental — accuracy not yet measured
The Implications for scholarship section (p. 3) reports a median of 20% inaccurate citations in a summary of medical research. But this paper's Results (p. 2) classify 24.2% of marine-biology citations as ambiguous, unsupported, or empty; those categories are not all claims incongruous with a cited paper's results. The source does not establish that most analyses find a rate between 10% and 20%. (LLM, conf=0.85)
- Supportedcites [75] H. Jergas, C. Baethge, Quotation accuracy in medical journal articles-a systematic review and meta-analysis. PeerJ 3 , e1364 (2015). source
The Results (Main analysis) estimate that 11.9% of references had major quotation errors—roughly one in eight or nine—meaning they did not support the citing claim. The Discussion reports an error in every fifth quotation, about half of them seriously incongruous, and notes substantial variation across studies. These figures support the citing sentence’s approximate range, rather than an exact rate for every analysis. (LLM, conf=0.86)
- Partly supportedcites [76] G. De Lacey, C. Record, J. Wade, How accurate are quotations and references in medical journals?. Br. Med. J. 291 , 884-886 (1985). sourceExperimental — accuracy not yet measured
The Discussion reports that 12% of references in the six medical journals surveyed contained slightly or seriously misleading accounts of the original work, within the claimed 10–20% range. This paper reports one survey; it does not establish what most analyses find across fields and methodologies. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“many established scientific results in the social (49, 50) and biomedical (51-53) sciences cannot readily be replicated.”
What the sources say
- Supportedcites [49] R. Rahal et al. , Estimating the reproducibility of psychological science. Science 349 , aac4716 (2015). source
This source supports the social-science portion of the sentence. The Discussion (p. 12) reports that a large portion of attempts to replicate published psychology findings did not reproduce evidence supporting the original results; Evaluating Reproducibility (p. 11) reports that 14 of 53 social-psychology effects replicated under its significance criterion. The biomedical portion is attributed to other citations, not this source. (LLM, conf=0.92)
- Supportedcites [50] C. F. Camerer et al. , Evaluating replicability of laboratory experiments in economics. Science 351 , 1433-1436 (2016). source
The citation to this paper is attached to the social-sciences portion of the sentence; separate citations accompany the biomedical-sciences portion. In the Main Text, the paper reports that 11 of 18 economics experiments replicated at the stated significance threshold and that four were strong replication failures. It also reports that a psychology replication project found a significant effect in the same direction for 36% of the studies it tested. (LLM, conf=0.90)
- Partly supportedcites [51] C. G. Begley, L. M. Ellis, Drug development: Raise standards for preclinical cancer research. Nature 483 , 531-533 (2012). sourceExperimental — accuracy not yet measured
The biomedical portion of the sentence is attributed to this source; the social-science portion has separate citations. In “Confirming Research Findings” (p. 2), the authors report that findings were confirmed in only 6 of 53 selected landmark preclinical cancer papers. This supports a reproducibility problem in preclinical cancer research, but not the broader claim about biomedical sciences generally. (LLM, conf=0.85)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
56 more claims · 5 with their cited sources checked
Create a free account to see all 105 claims and to verify the 86 references still open.
- Probably not a substantial claim. 27
- It may state two separate claims as one. 24
Citations
Cites
3 retracted
Uncanny similarity of unique inserts in the 2019-nCoV spike protein to HIV-1 gp120 and Gag
Pradhan, P., Pandey, A. K., Mishra, A., Gupta, P., Tripathi, P. K., Menon, M. B., Gomes, J., Vivekanandan, P., & Kundu, B. (2020) · doi:10.1101/2020.01.30.927871 · Cited by 198 · cited 2×
“The withdrawn bioRxiv preprint mentioned previously was promoted so broadly that it garnered one of the highest Altmetric scores of all time (28).”
BackgroundRetractedMedical Students’ Perception of Their Education and Training to Cope with Future Market Trends
Iesa, M. A. M. (2020) · Advances in Medical Education and Practice · doi:10.2147/amep.s233494 · Cited by 3
RetractedIesa, M. A. M. (2020) · Advances in Medical Education and Practice · doi:10.2147/amep.s261451 · Cited by 2
RetractedThe file drawer problem and tolerance for null results.
Rosenthal, R. (1979) · Psychological Bulletin · doi:10.1037/0033-2909.86.3.638 · Cited by 9,068
The spread of true and false news online
Vosoughi, S., Roy, D., & Aral, S. (2018) · Science · doi:10.1126/science.aap9559 · Cited by 8,847
Estimating the reproducibility of psychological science
Open Science Collaboration (2015) · Science · doi:10.1126/science.aac4716 · Cited by 8,515
Social Media and Fake News in the 2016 Election
Allcott, H., & Gentzkow, M. (2017) · Journal of Economic Perspectives · doi:10.1257/jep.31.2.211 · Cited by 6,767
The Matthew Effect in Science: The reward and communication systems of science are considered.
Merton, R. K. (1968) · Science · doi:10.1126/science.159.3810.56 · Cited by 6,720
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011) · Psychological Science · doi:10.1177/0956797611417632 · Cited by 6,567
The ASA Statement on p -Values: Context, Process, and Purpose
Wasserstein, R. L., & Lazar, N. A. (2016) · The American Statistician · doi:10.1080/00031305.2016.1154108 · Cited by 6,306
Research works printed but not linked: 6
- T. C. Bergstrom, C. T. Bergstrom, Can ‘author pays’ journals compete with ‘reader pays?’ Nature (2004).
- A. Cairo, How Charts Lie. (WW Norton & Company, 2019).
- K. Parker, The growing partisan divide in views of higher education. Pew Research Center (19 August 2019). https://www.pewsocialtrends.org/essay/the-growing-partisan-divide-in-views-of-higher-education/. Accessed 7 December 2020.
- C. Funk, Mixed messages about public trust in science. Issues Sci. Technol. 34, 86–88 (2017).
- N. M. Krause, D. Brossard, D. A. Scheufele, M. A. Xenos, K. Franke, Trends—Americans’ Trust in Science and Scientists. Publ. Opin. Q. 83, 817–836 (2019).
- N. S. Board Science and Engineering Indicators (NSF 2018) vol. 1.
Cited by
Romanello, M., Walawender, M., Hsu, S.-C., Moskeland, A., Palmeiro-Silva, Y., Scamman, D., Ali, Z., Ameli, N., Angelova, D., Ayeb-Karlsson, S., et al. (2024) · The Lancet · doi:10.1016/s0140-6736(24)01822-1 · Cited by 897
Towards a standard for identifying and managing bias in artificial intelligence
Schwartz, R., Vassilev, A., Greene, K., Perine, L., Burt, A., & Hall, P. (2022) · doi:10.6028/nist.sp.1270 · Cited by 548
Trust in scientists and their role in society across 68 countries
Cologna, V., Mede, N. G., Berger, S., Besley, J., Brick, C., Joubert, M., Maibach, E. W., Mihelj, S., Oreskes, N., Schäfer, M. S., et al. (2025) · Nature Human Behaviour · doi:10.1038/s41562-024-02090-5 · Cited by 251
“nature human behaviour https://doi.org/10.1038/s41562-024-02090-5”
InfluentialStewardship of global collective behavior
Bak-Coleman, J. B., Alfano, M., Barfuss, W., Bergstrom, C. T., Centeno, M. A., Couzin, I. D., Donges, J. F., Galesic, M., Gersick, A. S., Jacquet, J., et al. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2025764118 · Cited by 289
Self-citationZhou, J., Zhang, Y., Luo, Q., Parker, A. G., & De Choudhury, M. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581318 · Cited by 288
“Oftentimes, such information can be fawed or unrepresentative, or the conclusions and interpretation are unjustifed [110].”
Background
Metadata
- Authors
- West, J. D., & Bergstrom, C. T.
- Affiliations
- University of Washington (US)
- Corresponding author
- West, Jevin D.
- Venue
- Proceedings of the National Academy of Sciences
- Publisher
- National Academy of Sciences
- Published
- 2021-04-09
- Volume
- 118
- Issue
- 15
- ISSN
- 0027-8424 (ISSN-L)
- 1091-6490
- PMC
- PMC8054004
- OpenAlex
- W3148182339
- Cited by
- 391
- References
- 144
- Citations by year
- 2026: 38 · 2025: 97 · 2024: 108 · 2023: 77 · 2022: 51 · 2021: 18
- FWCI
- 12.18
- Type
- journal article
- Language
- en
- Update notices
- correction (2024-08-02)
- Open access
- green
- www.ncbi.nlm.nih.gov
- Locations
- Proceedings of the National Academy of Sciences (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Funders
- National Science Foundation (1735194)
- John S. and James L. Knight Foundation (G-2019-58788)
- Topics
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
- Social Sciences › Education › Academic integrity and plagiarism
- Keywords
- misinformation, science communication, scholarly publishing, publication bias, predatory publishing, hype, filter bubbles, information filtering, misinformation interventions
- MeSH terms
- Humans, Mass Media, Periodicals as Topic, Biomedical Research, Health Communication
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Misinformation in and about science. https://app.cochat.ai/papers/10.1073/pnas.1912444117
- The paper
- West, J. D., & Bergstrom, C. T. (2021). Misinformation in and about science. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.1912444117
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.