CoChat

How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data

Open access
Fanelli, Daniele · 2009 · PLoS ONE

The DOI matches its Crossref record; 64 of the 74 references it cites were found in a registry or an index; none of the 21 sources its claims cite is checked for support yet.

Abstract

The frequency with which scientists fabricate and falsify data, or commit other forms of scientific misconduct is a matter of controversy. Many surveys have asked scientists directly whether they have committed or know of a colleague who committed research misconduct, but their results appeared difficult to compare and synthesize. This is the first meta-analysis of these surveys. To standardize outcomes, the number of respondents who recalled at least one incident of misconduct was calculated for each question, and the analysis was limited to behaviours that distort scientific knowledge: fabrication, falsification, "cooking" of data, etc... Survey questions on plagiarism and other forms of professional misconduct were excluded. The final sample consisted of 21 surveys that were included in the systematic review, and 18 in the meta-analysis. A pooled weighted average of 1.97% (N = 7, 95%CI: 0.86-4.45) of scientists admitted to have fabricated, falsified or modified data or results at least once--a serious form of misconduct by any standard--and up to 33.7% admitted other questionable research practices. In surveys asking about the behaviour of colleagues, admission rates were 14.12% (N = 12, 95% CI: 9.91-19.72) for falsification, and up to 72% for other questionable research practices. Meta-regression showed that self reports surveys, surveys using the words "falsification" or "fabrication", and mailed surveys yielded lower percentages of misconduct. When these factors were controlled for, misconduct was reported more frequently by medical/pharmacological researchers than others. Considering that these surveys ask sensitive questions and have other limitations, it appears likely that this is a conservative estimate of the true prevalence of scientific misconduct.

Overview

CoChat Summary

TL;DR

This systematic review and meta-analysis of survey data estimated that about 2% of scientists admitted to fabricating, falsifying, or modifying data or results at least once, while 14.12% reported knowing a colleague who did so. The paper argues these are conservative estimates and that methodological choices strongly shape the observed rates.

Key findings

  • Across surveys explicitly asking whether scientists ever fabricated or falsified data or altered/modifed results, the pooled weighted estimate was 1.97% (95% CI 0.86–4.45; N = 7), with significant heterogeneity (Cochran’s Q = 61.7777, df = 6, P < 0.0001).
  • When only questions explicitly using the words “fabrication” or “falsification” were included, the pooled estimate for self-reports was 1.06% (95% CI 0.31–3.51; N = 4).
  • For reports about colleagues’ misconduct, the pooled weighted estimate was 14.12% (95% CI 9.91–19.72; N = 12); restricting to questions using “fabrication” or “falsification” gave 12.34% (95% CI 8.43–17.71; N = 11).
  • Other questionable research practices were reported more often: up to 33.7% in self-reports and up to 72% when asking about colleagues.

Methods

  • Systematic review and meta-analysis of quantitative surveys asking scientists whether they had committed or observed scientific misconduct in the past; plagiarism and other non-knowledge-distorting misconduct were excluded.
  • Searches were run in the first two weeks of August 2008 across citation databases, journals, grey literature sources, and internet search engines; the final review included 21 studies, with 18 in the meta-analysis.
  • Main analysis computed the proportion recalling at least one incident, logit-transformed and inverse-variance weighted in a random-effects meta-analysis; meta-regression tested methodological factors such as self vs. non-self reports, mailed vs. handed-out surveys, and use of the words “fabrication”/“falsification.”

Limitations

  • The paper states that using the percentage of respondents recalling at least one incident is a rough measure of misconduct frequency; this can under- or overestimate prevalence depending on whether respondents committed multiple incidents or only one.
  • The paper notes that non-response bias was recognized by most surveys as a limitation, and that the validity of self-reports may be affected by social expectations; the text also suggests self-reports likely underestimate misconduct.
  • Evident from the text: the included studies were heterogeneous in methods and wording, and some outcomes were too diverse for meaningful meta-analysis, so only crude unweighted parameters were calculated for broader questionable research practices.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0005738

Cites

64 of 64 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Study selection flow diagram.
Figure 1 Study selection flow diagram.CC BY · PubMed Central
Figure 2 Forrest plot of admission rates of data fabrication, falsification and alteration in self reports. Area of squares represents sample size, horizontal lines are 95% confidence interval, diamond and vertical dotted line show the pooled weighted estimate.
Figure 2 Forrest plot of admission rates of data fabrication, falsification and alteration in self reports. Area of squares represents sample size, horizontal lines are 95% confidence interval, diamond and vertical dotted line show the pooled weighted estimate.CC BY · PubMed Central
Figure 3 Admission rates of Questionable Research Practices (QRP) in self- and non-self-reports. N indicates the number of survey questions. Boxplots show median and interquartiles.
Figure 3 Admission rates of Questionable Research Practices (QRP) in self- and non-self-reports. N indicates the number of survey questions. Boxplots show median and interquartiles.CC BY · PubMed Central
Figure 4 Forrest plot of admission rates of data fabrication, falsification and alteration in non-self reports. Area of squares represents sample size, horizontal lines are 95% confidence interval, diamond and vertical dotted line show the pooled weighted estimate.
Figure 4 Forrest plot of admission rates of data fabrication, falsification and alteration in non-self reports. Area of squares represents sample size, horizontal lines are 95% confidence interval, diamond and vertical dotted line show the pooled weighted estimate.CC BY · PubMed Central
Figure 5 Sensitivity analysis of admission rates of data fabrication, falsification and alteration in self reports. Plots show the weighted pooled estimate and 95% confidence interval obtained when the corresponding study was left out of the analysis.
Figure 5 Sensitivity analysis of admission rates of data fabrication, falsification and alteration in self reports. Plots show the weighted pooled estimate and 95% confidence interval obtained when the corresponding study was left out of the analysis.CC BY · PubMed Central
Figure 6 Sensitivity analysis of admission rates of data fabrication, falsification and alteration in non-self reports. Plots show the weighted pooled estimate obtained when the corresponding study was left out of the analysis.
Figure 6 Sensitivity analysis of admission rates of data fabrication, falsification and alteration in non-self reports. Plots show the weighted pooled estimate obtained when the corresponding study was left out of the analysis.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 0 Not valid · 14 Not checked · 61 Own results and interpretations

31 more claims

Create a free account to see all 75 claims and to verify the 4 references still open.

See all claims — free account
  • It may state two separate claims as one. 24
  • Probably not a substantial claim. 25

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

6 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Citations

Cites

Show all 64 references

Research works printed but not linked: 1

  1. S Eastwood, 1996, Ethical issues in biomedical research: perceptions and practices of postdoctoral research fellows responding to a survey., Science and Engineering Ethics, 2, 89, 10.1007/bf02639320 · DOI

Cited by

See all 1,924

Metadata

Authors
Fanelli, D.
Affiliations
University of Edinburgh (GB)
Corresponding author
Fanelli, Daniele
Venue
PLoS ONE
Published
2009-05-29
Volume
4
Issue
5
Pages
e5738
ISSN
1932-6203 (ISSN-L)
OpenAlex
W2152243616
Cited by
1,911
References
74
Citations by year
2026: 77 · 2025: 110 · 2024: 128 · 2023: 136 · 2022: 157 · 2021: 145 · 2020: 163 · 2019: 166 · 2018: 125 · 2017: 125 · 2016: 119 · 2015: 106 · 2014: 82 · 2013: 93 · 2012: 84
FWCI
19.42
Type
journal article
Language
en
Open access
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Edinburgh Research Explorer (University of Edinburgh) (acceptedVersion, public-domain)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
  • DigitalGeorgetown (Georgetown University Library) (submittedVersion)
Funders
  • European Commission (221441)
  • Economic and Social Research Council (ES/F028180/1)
Topics
  • Social Sciences › Education › Academic integrity and plagiarism
  • Physical Sciences › Mathematics › Statistics and Probability › Benford’s Law and Fraud Detection
  • Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
Keywords
research misconduct, scientific fraud, data fabrication, data falsification, questionable research practices, self-report questionnaire
MeSH terms
Data Collection, Regression Analysis, Research Personnel, Scientific Misconduct, Ethics, Research

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.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

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: How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data. https://app.cochat.ai/papers/10.1371/journal.pone.0005738
The paper
Fanelli, D. (2009). How Many Scientists Fabricate and Falsify Research? A Systematic Review and Meta-Analysis of Survey Data. PLoS ONE. https://doi.org/10.1371/journal.pone.0005738

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.