Skip to content

GRF · Research Nexus · Foundations

Research integrity: principles, misconduct, retractions and reform

Research integrity is the practice of doing, reviewing and reporting research honestly and accountably. This foundational guide explains the principles, what counts as misconduct, what the evidence on misconduct and retractions shows, the instruments that govern integrity and how assessment reform is changing incentives.

As of
3 November 2026
Ukufunda
13 minutes
Ubufakazi
17 sources, cited in order
NEXUS TWIN · WF-1 · NETWORKSEED 6d6e5f83R01.RESEARCH-INTEGRITY.FOUNDATIONSxyzSCHEMATIC · NETWORKST-01 · 0.562 0.276ST-02 · 0.201 0.307ST-03 · 0.623 0.429
WF-1 · network Schematic wireframe twin (network, decision under uncertainty). An illustration, not data.
Misconduct
8.3%
of researchers in a 2022 Dutch survey reported fabrication or falsification in three years6
Umkhuba
51.3%
reported at least one questionable research practice frequently6
Record
10,000+
research papers were retracted in 2023, a record8

01What research integrity is

What research integrity is

Research integrity is the quality of research that is designed, carried out, reviewed and reported honestly, so that others can trust and build on it. The European Code of Conduct for Research Integrity, revised by the All European Academies (ALLEA) in June 2023, rests it on four principles: reliability in the quality of research, honesty in developing and reporting it, respect for colleagues, participants, society and ecosystems, and accountability from idea to publication 1. The World Health Organization's code for research associated with its work names a similar set: integrity, accountability, independence and impartiality, respect and professional commitment 12.

Integrity is wider than the absence of fraud. UK Research and Innovation (UKRI) describes it as the product of responsible research practices and of a culture that supports them, built on honesty, rigour, transparency and open communication, care and respect, and accountability 13. The legal floor is narrower. In United States federal policy, research misconduct means fabrication, falsification or plagiarism in proposing, performing or reviewing research, or in reporting results, and it explicitly does not include honest error or differences of opinion 2,3.

This guide uses the two layers on purpose. The floor tells institutions what they must investigate. The broader principles tell them what kind of research culture makes the floor rarely needed. Other US agencies use the same core definition; the National Science Foundation's rule, for example, defines misconduct as fabrication, falsification or plagiarism and also excludes honest error 14.

The Global Risks Forum (GRF) publishes this guide on its Research platform, where research integrity is a topic of its own with a longer overview. Risk decisions depend on evidence that is trustworthy as well as relevant. The Global Centre for Risk and Innovation (GCRI) leads on science and methods, and The Global Risks Alliance (GRA) reads evidence for finance.

02Why it matters for risk evidence

Why it matters for risk evidence

Decisions about floods, pandemics, financial stability and new technologies borrow their authority from research. When a study is fabricated, selectively reported or quietly retracted, the error travels: into guidance, models, risk registers and the next round of research that cites it.

The scale is not trivial. A meta-analysis of 18 surveys found that a pooled 1.97 per cent of scientists admitted to fabricating, falsifying or modifying data at least once, and up to 33.7 per cent admitted other questionable practices; asked about colleagues, 14.12 per cent reported falsification 5. A national survey in the Netherlands, with 6,813 respondents and a method designed to protect identities, estimated self-reported fabrication at 4.3 per cent and falsification at 4.2 per cent, 8.3 per cent for either, and found that 51.3 per cent engaged frequently in at least one questionable research practice 6.

These numbers come from self-reports and differ in method, so they are not trends and should not be compared directly. They do show that most of the integrity problem lies in grey practices, such as selective reporting, weak supervision or poor data management, rather than in rare frauds. That is where institutions, funders and journals have most room to act.

Fig. 01Most integrity problems are not outright fraud

Most integrity problems are not outright fraud Bar chart of self-reported research behaviour in the Netherlands, 2022: fabrication 4.3 per cent, falsification 4.2 per cent, either 8.3 per cent and at least one frequent questionable research practice 51.3 per cent. Fabrication4.3 per cent of respondentsFalsification4.2 per cent of respondentsFabrication or falsification8.3 per cent of respondentsAt least one frequent questionablepractice51.3 per cent of respondents
Self-reported prevalence from a survey of 6,813 researchers that used a method designed to protect identities. Fabrication and falsification are rare; at least one frequent questionable practice is reported by about half 6.6Sources: Gopalakrishna G. et al., PLOS ONE, 2022. Figure CC BY 4.0. As of 3 November 2026.
Figure 1 in words

Four horizontal bars on one scale of per cent of respondents, from a national survey of 6,813 academic researchers in the Netherlands published in 2022. Fabrication in the last three years: 4.3 per cent. Falsification: 4.2 per cent. Fabrication or falsification: 8.3 per cent. At least one questionable research practice engaged in frequently: 51.3 per cent, highlighted. Self-reported estimates; not comparable across countries or over time. Source: Gopalakrishna and others, PLOS ONE, 2022.

03Misconduct, questionable practices and honest error

Misconduct, questionable practices and honest error

Clear categories help institutions respond in proportion. Fabrication is making up data or results and recording or reporting them; falsification is manipulating materials, equipment or processes, or changing or omitting data, so that the research record misrepresents the work; plagiarism is taking another person's ideas, processes, results or words without credit 2. These three are the core of misconduct in US federal policy, which the Office of Science and Technology Policy set for all federal agencies in December 2000 3.

Between misconduct and good practice lies a wide band of questionable research practices, which the surveys cited above measure separately from fabrication and falsification 5,6. Examples include misleading statistical analysis, inadequate supervision and failing to share data or code that others need to check a claim. Funders define their own breaches as well: the Canadian Tri-Agency Framework for the Responsible Conduct of Research lists, among others, mismanagement of grant funds and listing co-applicants without their agreement 15.

Honest error is not misconduct, and treating it as such harms research. A field that punishes correction teaches people to hide mistakes. Good integrity systems therefore separate three routes: investigation for misconduct, remediation and training for poor practice, and fast, unstigmatised correction for honest error.

Fig. 02Three routes for three kinds of problem

Three routes for three kinds of problem Comparison of three columns: misconduct, meaning fabrication, falsification or plagiarism; questionable practices, measured separately in surveys; and honest error, which the US definition of misconduct excludes.
Categories follow the US federal definition of misconduct and the surveys' separate measurement of questionable practices 2,5,6.2,5,6Sources: US Office of Research Integrity; PLOS ONE 2009 and 2022. Figure CC BY 4.0. As of 3 November 2026.
Figure 2 in words

Three columns. Misconduct: fabrication, falsification or plagiarism in US federal policy. Questionable practices: conduct short of misconduct, measured separately from fabrication and falsification in the meta-analysis and the Dutch survey. Honest error: excluded from the US federal definition of misconduct. Sources: US Office of Research Integrity; Fanelli 2009; Gopalakrishna and others 2022.

04Who is responsible: researchers, institutions, funders and publishers

Who is responsible: researchers, institutions, funders and publishers

Integrity is shared work. Researchers carry the first responsibility for their own conduct and for the people they supervise. Institutions set policies, train staff, receive allegations and run investigations; in the United States the 2024 final rule from the Department of Health and Human Services revised the Public Health Service regulations on research misconduct, which apply to institutions that apply for or receive PHS support, effective from 1 January 2025 4. Funders attach conditions to grants and can withdraw funding for breaches, as the Canadian framework does 15.

Publishers and editors decide what enters the record and what is corrected or retracted. Authorship rules are part of this. The International Committee of Medical Journal Editors (ICMJE) recommends four criteria for authorship: a substantial contribution to the work, drafting or critically revising it, approving the final version, and agreeing to be accountable for all aspects of it 16. National bodies coordinate across sectors; UKRI, for example, hosts the UK Committee on Research Integrity and supports the concordat to support research integrity 13.

Shared responsibility has a cost when it is unclear who acts first. Cross-border projects, multi-site trials and data shared between countries are where allegations most often stall, because each institution follows its own procedure. Agreeing in advance who investigates what is one of the simplest integrity measures a collaboration can take.

TwinSystem map of research integrity

Interactive · 3D
System map of research integrity
The causal structure of research integrity as this article describes it: drivers on the left, then hazards, exposure, impacts, responses and levers, with signed links and three feedback loops. Every number on the map names its source in the reference list. Drag to rotate and scroll to zoom; Explore opens the layers, story steps, time bar and charts.
Method and description

Article system map: nodes and signed links are the editorial desk's reading of the cited sources; link strengths are stated judgements on a 0 to 1 scale, not measurements. Node values quote the source named on hover.

The map has fifteen nodes in six kinds. Drivers: pressure to publish fast and positively; assessment by journal metrics and paper counts; generative AI and organised paper mills. Hazards: questionable research practices, reported frequently by about half of researchers in a 2022 Dutch survey; fabrication, falsification and plagiarism. Exposure: flawed papers in the published record; guidance, models and risk registers that cite it. Impacts: decisions built on unreliable evidence; loss of public trust. Responses: corrections and retractions, more than 10,000 in 2023; institutional investigations. Levers: open retraction data; codes and funder rules; assessment reform; training and data sharing. Three loops: corrections clean the record; investigations deter misconduct; failures erode trust, prompting assessment reform that lowers the pressure that started the chain.

  1. The whole system 15 nodes in six columns by kind, 18 signed links and 3 feedback loops: 3 drivers, 2 hazards, 2 exposure nodes, 2 impacts, 2 responses and 4 policy levers. The columns with the heaviest influence out are drivers (2.3) and policy levers (2.1); the heaviest dependence in falls on exposure (2.9) and hazards (2.5). The next steps add the chain one rank at a time with the sign and strength of every link.
  2. Drivers and hazards 3 drivers on the left push the system: pressures stated without a headline number. 2 hazards follow: questionable and detrimental research practices: 51.3% report a frequent questionable practice (NL, 2022) (s10); fabrication, falsification and plagiarism: 8.3% self-reported fabrication or falsification (NL, 2022) (s10). Link width is the stated strength (0 to 1); red links amplify, green links dampen; links within a column show hazards that feed each other.
  3. Exposure and impacts Hazards meet 2 exposure nodes and produce 2 impacts: . Hover a node for its value, unit and source id.
  4. Feedback loops 3 loops close the graph. B1 (correction loop): flawed record → retractions → flawed record, loop gain 0.25. B2 (deterrence loop): misconduct → investigations → misconduct, loop gain 0.2. B3 (reform loop): metric-based assessment → publication pressure → questionable practices → flawed record → downstream use → unreliable decisions → lost trust → assessment reform → metric-based assessment, loop gain 0.01. A reinforcing loop has an even number of dampening links and runs away; a balancing loop has an odd number and self-corrects.
  5. Responses and policy levers 4 levers and 2 responses on the right act back on the chain: open retraction data dampens downstream use (−0.5); assessment reform dampens metric-based assessment (−0.6); training and sharing dampens questionable practices (−0.5). The panel sums link strengths by kind: influence out of a column against dependence into it.

05What the evidence on retractions shows

What the evidence on retractions shows

Retractions are the most visible signal of problems in the published record, although a retraction can follow honest error as well as misconduct. Until 2023 the largest count sat in a database run by the Center for Scientific Integrity, the organisation behind Retraction Watch. On 12 September 2023 Crossref, the not-for-profit infrastructure that registers scholarly identifiers, acquired that database and made it a public resource: at the time Crossref held about 14,000 retraction records and the Retraction Watch database about 43,000, around 50,000 together once overlaps were removed 7.

The same year set a record. A Nature analysis counted more than 10,000 retractions issued in 2023, driven by publishers cleaning up sham papers and peer-review fraud 8. Rising counts partly reflect better detection and more active editors, so they are not a direct measure of how much misconduct exists. They do show that the published record now needs routine maintenance, and that the tools to check a paper's status are open to anyone.

For risk evidence, the practical lesson is to check the status of key sources before relying on them, and to record that check. A finding cited in a policy brief or a risk model should be traced to its current version, its corrections and any expression of concern.

Fig. 03Retractions became open data in 2023

Retractions became open data in 2023 Bar chart of retraction records in September 2023: about 14,000 at Crossref, about 43,000 in the Retraction Watch database and around 50,000 combined after removing overlaps. Crossrefabout 14,000 retraction recordsRetraction Watch databaseabout 43,000 retraction recordsCombined, without overlapsaround 50,000 retraction records
Reported counts announced when Crossref acquired the Retraction Watch database and opened it. The combined total removes records held in both 7.7Sources: Crossref, 12 September 2023. Figure CC BY 4.0. As of 3 November 2026.
Figure 3 in words

Three horizontal bars on one scale of retraction records, as announced on 12 September 2023. Crossref retraction records: about 14,000. Retraction Watch database: about 43,000. Combined, after removing records held in both: around 50,000, highlighted. Rising counts reflect better detection as well as problems in the record. Source: Crossref, 12 September 2023.

06Instruments that govern research integrity

Instruments that govern research integrity

Research integrity is governed by a mix of law, funder conditions, professional codes and voluntary declarations. The table lists the main instruments cited in this guide, with their status as of 3 November 2026. None of them is a treaty; their force comes from adoption by institutions, funders and publishers.

Two features stand out. First, the binding rules are national: the US federal policy and its agency rules 3,4, the Canadian Tri-Agency framework 15 and UK funder policies 13. Second, the international and European instruments are not national law: the European code, which the European Commission recognises as the primary standard for upholding research integrity across the projects the EU funds 1, and the declarations on research assessment 9,11.

IthuluziIssuer and yearEkwenzayo
Federal Policy on Research MisconductUS Office of Science and Technology Policy, 2000 3Defines misconduct as fabrication, falsification and plagiarism for all federal agencies
Public Health Service Policies on Research MisconductUS Department of Health and Human Services, final rule 2024, effective 2025 4Revises the Public Health Service regulations on research misconduct for institutions that apply for or receive PHS support
San Francisco Declaration on Research AssessmentDORA, 2012 9Asks assessors not to use journal-based metrics as a proxy for quality
Hong Kong Principles6th World Conference on Research Integrity, published 2020 10Five principles for assessing researchers on responsible practice
Tri-Agency Framework: Responsible Conduct of ResearchCIHR, NSERC and SSHRC, 2021 15Sets researcher responsibilities and breaches for Canadian federal funding
European Code of Conduct for Research IntegrityALLEA, revised 2023 1Four principles and good practices for European research
Guidelines on the responsible use of generative AI in researchEuropean Research Area Forum, 2024 17Guidance for researchers, organisations and funders on generative AI
Agreement on Reforming Research AssessmentCoARA 11Commitments by signatories to reform how research and researchers are assessed
Selected instruments on research integrity and research assessment, as of 3 November 2026. Status as stated by each issuer.

Fig. 04A timeline of research integrity instruments

A timeline of research integrity instruments Timeline of seven instruments: the US federal misconduct policy in December 2000, DORA in December 2012, the Hong Kong Principles in 2020, the Canadian Tri-Agency framework in 2021, the revised European code in June 2023, EU generative AI guidelines in March 2024 and the US rule effective January 2025. DEC 2000US federal misconductpolicyDEC 2012DORA2020Hong Kong Principles2021Canadian Tri-AgencyframeworkJUN 2023European code revisedMAR 2024EU guidelines ongenerative AIJAN 2025US PHS rule in force
Dates as reported by each issuer 3,9,10,15,1,17,4. Most international instruments are voluntary; national rules bind funded institutions.3,9,10,15,1,17,4Sources: OSTP; DORA; PLOS Biology; Government of Canada; ALLEA; European Commission; Federal Register. Figure CC BY 4.0. As of 3 November 2026.
Figure 4 in words

A horizontal timeline. December 2000: the US Office of Science and Technology Policy publishes the Federal Policy on Research Misconduct. 16 December 2012: DORA is drafted in San Francisco. 2020: the Hong Kong Principles, developed for the 6th World Conference on Research Integrity, are published. 2021: the Canadian Tri-Agency Framework for the Responsible Conduct of Research is issued. 23 June 2023: ALLEA publishes the revised European Code of Conduct for Research Integrity. 20 March 2024: European Research Area guidelines on generative AI in research. 1 January 2025: the revised US Public Health Service misconduct rule takes effect.

07Assessment and incentives: why reform focuses on careers

Assessment and incentives: why reform focuses on careers

Many integrity problems start with incentives. When hiring, promotion and funding reward the number of papers or the prestige of the journal, researchers face pressure to publish fast and positively. The San Francisco Declaration on Research Assessment (DORA), drafted by editors and publishers at a meeting in San Francisco on 16 December 2012, recommends evaluating research on its own merits rather than by journal-based metrics such as the Journal Impact Factor, which was designed to help librarians buy journals 9.

The Hong Kong Principles, developed for the 6th World Conference on Research Integrity, turn this into guidance for assessing people: assess researchers on responsible practices from conception to delivery, and value the accurate and transparent reporting of all research regardless of the results, among five principles 10. The CoARA agreement asks its signatory organisations to commit to a common vision in which assessment recognises the diverse outputs, practices and activities that maximise the quality and impact of research 11.

For institutions in the risk field, the implication is concrete: reward data sharing, replication, negative results, careful methods and service to review panels, not only first-author papers. These are the practices that make evidence usable in a crisis.

08What institutions can do now

What institutions can do now

A credible integrity system can be built with known parts. Publish a policy that uses the common definitions and states the routes for misconduct, poor practice and honest error 2,1. Name a confidential contact for concerns and protect people who raise them in good faith. Train researchers and supervisors in data management, authorship and statistics, not only in rules.

Make research checkable. Require data and code to be shared where law and ethics allow, register protocols for confirmatory studies, and check the retraction status of key sources before they enter guidance or models 7. Agree authorship at the start of a project using the ICMJE criteria or an equivalent 16.

Align incentives. Sign and implement the assessment declarations, and review hiring and funding criteria against them 9,11. In collaborations, agree in writing which institution leads an investigation and how findings are shared. For researchers using generative tools, follow the European Research Area guidance, which builds on the European code 17.

09How allegations are handled

How allegations are handled

Most national systems follow the same broad sequence, even when the details differ. A concern is raised, usually with the institution where the research was done. The institution first runs an inquiry, a preliminary look at whether the allegation has substance. If it does, a formal investigation follows, with the right to respond and a written record. The National Science Foundation rule describes this two-step sequence directly, and says NSF will find research misconduct only after careful inquiry and investigation 14. In that rule, an inquiry is preliminary information-gathering and fact-finding to decide whether an allegation has substance, and an investigation is a formal development, examination and evaluation of a factual record 14.

A finding of misconduct usually needs more than proof that the record is wrong. US federal policy requires that the conduct be a significant departure from accepted practice and be committed intentionally, knowingly or recklessly, proved by a preponderance of the evidence 3. This is why many retractions never lead to a finding: the record can be wrong without anyone having acted in bad faith.

Outcomes range from correction of the record and supervision to the end of funding or employment. Journals act on the paper, while institutions and funders act on the person. Because the two tracks run separately, a retraction notice and an institutional finding can appear years apart. Readers of risk evidence should therefore treat a retraction as a signal about the paper, not as a verdict on its authors.

10What is changing

What is changing

Generative AI. Text and image generators raise new questions of authorship, disclosure and fabrication. The European Commission and the European Research Area countries issued guidelines on 20 March 2024 that frame responsible use on the principles of the European code and on trustworthy AI 17. Expect journals and funders to keep tightening disclosure rules.

Paper mills and peer-review fraud. The record number of retractions in 2023 was driven by publishers cleaning up sham papers and peer-review fraud 8. Detection tools and shared publisher screening are growing in response.

Open retraction data. With the Retraction Watch database now a public Crossref resource, Crossref says the collaboration helps the downstream services that rely on high-quality, open data about retractions 7.

Assessment reform. DORA and the CoARA agreement set out commitments to change how research and researchers are assessed 9,11.

Updated national rules. The US Department of Health and Human Services' revised misconduct rule took effect on 1 January 2025 4.

11Three faces of research integrity

Three faces of research integrity

Research integrity looks different from each of the consortium's three faces. For the science and technical side, led by GCRI, integrity is about methods, data, code and reproducible results; its Labs work on methods in the open. The GCRI Registry keeps public records of what a component or dataset is and where it came from, which is part of keeping evidence traceable under the provenance and evidence standards.

For governance, the face of GRF, integrity is about trust in the evidence that ministries, regulators and civil society use, and about clear procedures when that evidence fails. For finance, the face of GRA, it is about not building decisions on evidence that has been corrected or withdrawn. These are descriptions of how the paradigm's institutions approach the topic, not services or certifications.

What this article does not claim: it is an educational summary of public sources as of 3 November 2026, not legal advice, an assessment of any researcher, institution or journal, or a finding of misconduct. Allegations are handled by the institutions, funders and authorities named in the relevant policies. GRF does not investigate, adjudicate or certify research; GCRI does not regulate, certify, license or approve; GRA does not lend, underwrite, rate or invest. Naming the WHO, the European Commission, UKRI, ALLEA, DORA, CoARA, Crossref or any other body is not a claim of partnership, endorsement or alignment.

Imibuzo

Imibuzo nezimpendulo

What is research integrity?

It is the practice of designing, conducting, reviewing and reporting research honestly and accountably so that others can trust it. The European code names four principles: reliability, honesty, respect and accountability 1.

What counts as research misconduct?

In US federal policy, misconduct is fabrication, falsification or plagiarism in proposing, performing, reviewing or reporting research. It does not include honest error or differences of opinion 2,3.

How common is research misconduct?

Surveys differ. A meta-analysis found 1.97 per cent of scientists admitted fabricating or falsifying data at least once, and a 2022 Dutch survey estimated 8.3 per cent for either. Questionable practices were much more common 5,6.

What is a questionable research practice?

It is conduct short of misconduct that still damages research, such as selective reporting, misleading statistics or poor supervision.

How many papers have been retracted?

Crossref and the Retraction Watch database together held around 50,000 retractions in September 2023, and more than 10,000 papers were retracted in 2023 alone 7,8.

What is DORA?

The San Francisco Declaration on Research Assessment, drafted in December 2012, asks funders, institutions and publishers to assess research on its own merits rather than by journal-based metrics such as the Journal Impact Factor 9.

Imigomo

Isichazamazwi

Ubuqotho bocwaningo
Doing, reviewing and reporting research honestly and accountably, so that others can rely on it.
Fabrication
Making up data or results and recording or reporting them.
Falsification
Manipulating research materials, equipment or processes, or changing or omitting data, so that the record misrepresents the work.
Plagiarism
Using another person's ideas, processes, results or words without giving appropriate credit.
Questionable research practice
Conduct short of misconduct that weakens research, such as selective reporting or misleading analysis.
Retraction
A formal notice that a published article should no longer be relied on, issued by the journal or publisher.
Research assessment
How funders and institutions judge research and researchers for hiring, promotion and funding.

Imithombo

References

17 sources, numbered in the order they are first cited
  1. ALLEA, All European Academies (2023-06-23). The European Code of Conduct for Research Integrity, revised edition 2023. allea.org. Accessed 2026-11-03.
  2. US Office of Research Integrity (ORI), Department of Health and Human Services (current web page (42 CFR Part 93)). Definition of research misconduct. ori.hhs.gov. Accessed 2026-11-03.
  3. Office of Science and Technology Policy, Executive Office of the President (2000-12-06). Federal Policy on Research Misconduct (65 FR 76260). ori.hhs.gov. Accessed 2026-11-03.
  4. US Department of Health and Human Services, Federal Register (2024-09-17). Public Health Service Policies on Research Misconduct, final rule. www.federalregister.gov. Accessed 2026-11-03.
  5. Fanelli D., PLOS ONE 4(5): e5738 (2009-05-29). How many scientists fabricate and falsify research? A systematic review and meta-analysis of survey data. doi.org. Accessed 2026-11-03.
  6. Gopalakrishna G. et al., PLOS ONE 17(2): e0263023 (2022-02-16). Prevalence of questionable research practices, research misconduct and their potential explanatory factors: A survey among academic researchers in The Netherlands. doi.org. Accessed 2026-11-03.
  7. Crossref (2023-09-12). News: Crossref acquires Retraction Watch data and opens it for the scientific community. www.crossref.org. Accessed 2026-11-03.
  8. Van Noorden R., Nature 624, 479–481 (2023-12-12). More than 10,000 research papers were retracted in 2023: a new record. doi.org. Accessed 2026-11-03.
  9. DORA (2012-12-16). San Francisco Declaration on Research Assessment (DORA). sfdora.org. Accessed 2026-11-03.
  10. Moher D. et al., PLOS Biology 18(7): e3000737 (2020-07-16). The Hong Kong Principles for assessing researchers: Fostering research integrity. doi.org. Accessed 2026-11-03.
  11. Coalition for Advancing Research Assessment (CoARA) (agreement text as published by CoARA). Agreement on Reforming Research Assessment, full text. coara.eu. Accessed 2026-11-03.
  12. World Health Organization (WHO) (current web page). Code of Conduct for responsible research. www.who.int. Accessed 2026-11-03.
  13. UK Research and Innovation (UKRI) (current web page). Research integrity: good research resource hub. www.ukri.org. Accessed 2026-11-03.
  14. US National Science Foundation, Electronic Code of Federal Regulations (45 CFR Ch. VI (10-1-23 edition), part 689 (source 67 FR 11937, 18 March 2002)). 45 CFR Part 689, Research misconduct. www.govinfo.gov. Accessed 2026-11-03.
  15. Government of Canada, Panel on Responsible Conduct of Research (CIHR, NSERC, SSHRC) (2021-12-01). Tri-Agency Framework: Responsible Conduct of Research (2021). rcr.ethics.gc.ca. Accessed 2026-11-03.
  16. International Committee of Medical Journal Editors (ICMJE) (current recommendations). Defining the role of authors and contributors. www.icmje.org. Accessed 2026-11-03.
  17. European Commission, Directorate-General for Research and Innovation (2024-03-20). Guidelines on the responsible use of generative AI in research developed by the European Research Area Forum. research-and-innovation.ec.europa.eu. Accessed 2026-11-03.

Record

Editorial record

Facts and policies are stated as of 3 November 2026. This article is reviewed at least every 180 days and whenever a cited policy, code or dataset changes.

Published by the GRF policy desk on Research Nexus under the editorial integrity policy. To report an error or request a correction, use the Research Nexus corrections route; corrections are dated and listed with the article.

Text, figures and data tables are published under the Creative Commons Attribution 4.0 licence (CC BY 4.0): reuse them with credit to GRF and a link to this page. Third-party material keeps its own terms, credited where it appears.

Cite this article

GRF policy desk (2026, November 3). Research integrity: principles, misconduct and reform. Research Nexus. https://globalriskforum.com/research/research-integrity/

Republish

Republish or translate this article free of charge, in print or online, under CC BY 4.0: keep the byline and Research Nexus, link to the original and note any change. Figures, maps and twins made by the institutions carry the same licence; third-party data keeps its own credit line.

Attribution to paste
<p>Research integrity: principles, misconduct, retractions and reform — GRF policy desk, Research Nexus. First published 3 November 2026 at <a href="https://globalriskforum.com/research/research-integrity/">https://globalriskforum.com/research/research-integrity/</a>. Republished under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/).</p>

Related across the Consortium

Share

EmailLinkedInBlueskyMastodon