Research ethics and integrity: a glossary

3Rs (replacement, reduction, refinement)

Also called: 3Rs, three Rs, the 3Rs, 3Rs principles

The 3Rs are replacement, reduction and refinement, the framework for more humane animal research first defined by Russell and Burch in The Principles of Humane Experimental Technique. They are now built into law, regulation and funders' policies, and in the UK the Home Office applies them to every licensed project that uses animals in science.

Academic integrity

Also called: academic honesty

Academic integrity is a commitment by students, teachers and researchers to honest, fair and responsible work, in which the sources and help behind a piece of work are acknowledged. The International Center for Academic Integrity names six fundamental values: honesty, trust, fairness, respect, responsibility and courage. Research integrity applies the same commitment to doing and publishing research.

Academic integrity policy wording for AI and references

AI authorship

Also called: AI as an author, ChatGPT as an author, AI co-author

AI authorship is the question of whether an artificial intelligence tool, such as a chatbot, can be named as an author. COPE and the ICMJE both say it cannot, chiefly because a tool cannot take responsibility for the work, and COPE adds that it cannot declare conflicts of interest or agree to copyright and licence terms. The human authors stay responsible for everything, including text the tool produced.

AI in peer review

Also called: generative AI in peer review, using AI to peer review, LLMs in peer review

AI in peer review is the use of artificial intelligence tools by reviewers or editors to read, summarise or draft assessments of manuscripts and grant applications. The main ethical objection is confidentiality, since pasting unpublished work into an outside tool discloses it. On that ground the US National Institutes of Health has barred its reviewers from using generative AI to write critiques since June 2023.

Amendment

Also called: protocol amendment, study amendment, modification, study modification

An amendment is a change made to a research study after it has received ethical approval, such as a new questionnaire, an added site or longer follow-up. Depending on its effect it may need the ethics committee's approval before it is made, only notification, or neither. For UK health and social care studies other than drug trials, the Health Research Authority now calls amendments modifications.

Anonymisation

Also called: anonymization, anonymising data, anonymizing data, data anonymisation, anonymised data, anonymized data

Anonymisation is the process of turning personal data into information that no longer relates to an identifiable person, by removing or altering identifiers so that re-identification is not reasonably likely. Under UK data protection law truly anonymous information falls outside the law, while pseudonymised data do not. In qualitative reports it also means changing indirect details, such as places and job titles.

Anonymity

Also called: participant anonymity, anonymous participation

Anonymity is the condition in which information in a study cannot be linked to the person who gave it, sometimes not even by the researcher, as in a survey that never records names. It differs from confidentiality, where the researcher knows who said what but undertakes not to reveal it. BERA notes that anonymity may be impossible in small communities and that some participants prefer to be named.

Assent

Also called: child assent, children's assent, participant assent

Assent is the agreement to take part in research of someone who cannot give legally valid consent, usually a child, sought alongside the permission of a parent or legal representative. US regulations define it as the child's affirmative agreement, so a mere failure to object does not count. UK guidance says a child able to weigh the information must have their wishes, including a refusal, considered.

Autonomy

Also called: respect for autonomy, principle of autonomy, patient autonomy

Autonomy, in research ethics, is the principle that people able to decide for themselves have the right to choose what happens to them, so researchers must disclose what taking part involves and accept a refusal. It is one of the four principles of biomedical ethics and the root of informed consent, confidentiality and truthfulness. The Belmont Report folds it into respect for persons.

Belmont Report

Also called: Belmont principles, Ethical Principles and Guidelines for the Protection of Human Subjects of Research

The Belmont Report is a 1979 statement of ethical principles for research with human participants, written by the US National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. It names three principles, respect for persons, beneficence and justice, and applies them in turn to informed consent, the assessment of risks and benefits, and the fair selection of participants.

Beneficence

Also called: principle of beneficence

Beneficence is the ethical principle that research should secure people's wellbeing, by maximising possible benefits and minimising possible harms. The Belmont Report treats it as an obligation, not an act of kindness, and grounds the requirement for a favourable assessment of risks and benefits in it. In the four principles of biomedical ethics, the duty to avoid harm is split off as non-maleficence.

Broad consent

Also called: broad consent for future research

Broad consent is agreement to the storage of a person's data or samples and their use in future research of a described general kind, without the specific studies being known when consent is given. It is common in biobanks. The revised US Common Rule allows it for identifiable information and biospecimens, and UK data protection guidance recognises consent to certain areas of scientific research.

Capacity to consent

Also called: mental capacity, decision-making capacity, capacity, competence to consent

Capacity to consent is a person's ability to understand the information about a study, weigh it and reach and communicate a decision about taking part. It relates to a particular decision and can change over time. UK guidance asks researchers to make the most of whatever capacity a person has, and Canada's TCPS 2 prefers judging capacity to relying on age alone.

CARE principles

Also called: CARE Principles for Indigenous Data Governance, CARE

The CARE principles are four principles for governing Indigenous data: collective benefit, authority to control, responsibility and ethics. Drafted in 2019 through the International Indigenous Data Sovereignty Interest Group of the Research Data Alliance, they complement the FAIR principles, which concern the data themselves, by focusing on people and purpose and on Indigenous peoples' control of data about them.

Certificate of Confidentiality

Also called: CoC, certificates of confidentiality

A Certificate of Confidentiality is a US protection that prohibits researchers from disclosing identifiable, sensitive research information about participants to anyone not connected with the research, except with the participant's consent or in a few specific situations. Since 2017, research funded by the National Institutes of Health that meets set criteria has been deemed to have one automatically.

CIOMS guidelines

Also called: CIOMS, CIOMS ethical guidelines, International Ethical Guidelines for Health-related Research Involving Humans

The CIOMS guidelines are the International Ethical Guidelines for Health-related Research Involving Humans, published by the Council for International Organizations of Medical Sciences, prepared in cooperation with the World Health Organization, and last revised in 2016. They show how the principles of the Declaration of Helsinki can be applied in practice, especially in low-resource settings, and cover vulnerable groups and the research use of samples and health data.

Citation cartel

Also called: citation cartels, citation ring, citation manipulation cartel

A citation cartel is a group of journals or researchers who cite one another's work to inflate each other's citation counts, whatever its relevance. COPE describes a citation manipulation cartel as two or more journals collaborating to boost their impact factors, and the European Code of Conduct separately condemns cartels of reviewers and authors who collude to review each other's papers.

Citation manipulation

Also called: citation gaming, citation padding

Citation manipulation is adding citations that do not contribute to a paper's scholarly content, to raise the citation counts of an author, a journal or someone else. COPE describes excessive self-citation, excessive citation of the publishing journal, and honorary citation or citation stacking for others, and the European Code of Conduct condemns padding a bibliography to please editors, reviewers or colleagues.

Clinical audit

Also called: audit, clinical audit project

A clinical audit is a project that measures whether a health or care service meets a predetermined, recognised standard, using only interventions already in use. The UK Health Research Authority does not class it as research, so it needs no review by a research ethics committee, although it has governance requirements of its own. A service evaluation, by contrast, measures a service without reference to a standard.

Clinical equipoise

Also called: community equipoise

Clinical equipoise is genuine uncertainty within the expert medical community, not necessarily in the mind of the individual investigator, about which of the treatments compared in a trial is better. Benjamin Freedman proposed it in 1987 in place of the stricter demand that each investigator have no treatment preference, which he argued made most controlled trials almost impossible to start or finish.

Clinical trial registration

Also called: trial registration, registering a trial, clinical trials registration

Clinical trial registration is the public recording of a trial's key design details, such as its outcomes, size and sponsor, in an approved registry such as ClinicalTrials.gov or a WHO primary registry. The World Health Organization calls registering every interventional trial a scientific, ethical and moral responsibility, and the Declaration of Helsinki requires research to be registered before the first participant is recruited.

Coercion

Also called: coerced participation, coerced consent

Coercion, in research, is an overt threat of harm or punishment used to make someone take part or stay in a study. It destroys the voluntariness that valid consent needs. Codes frame it differently: the Belmont Report sets it apart from undue influence, which works through excessive rewards or improper pressure, while Canada's TCPS 2 calls coercion an extreme form of undue influence.

Coercive citation

Also called: citation coercion, coerced citation

Coercive citation is the practice of an editor, reviewer or journal making the citation of their own work, or of the journal's articles, a condition of accepting a manuscript regardless of relevance. COPE treats it as a form of citation manipulation. Suggesting a genuinely relevant paper is not coercive, provided acceptance does not depend on the author adding it.

Committee on Publication Ethics

Also called: COPE

The Committee on Publication Ethics, COPE, is a charity registered in England and Wales that helps journal editors, publishers, universities and others handle ethical problems in scholarly publishing, through guidance, flowcharts, discussion documents and training. It grew from an informal gathering of UK biomedical editors in 1997 and now serves all disciplines worldwide.

Common Rule

Also called: the Common Rule, Federal Policy for the Protection of Human Subjects, 45 CFR 46, 45 CFR part 46, revised Common Rule

The Common Rule is the US federal policy for protecting people who take part in research, first published in 1991, codified by more than a dozen federal departments and agencies, and found for the health department in 45 CFR 46, subpart A. It requires institutional review board approval and informed consent for most research it covers, exempting some low-risk categories. Its revised version is known as the 2018 Requirements.

Community-based participatory research ethics

Also called: CBPR ethics, ethics of community-based participatory research

Community-based participatory research ethics are the commitments and problems that arise when researchers and communities share the planning, conduct and use of research as partners. A literature review by Mikesell and colleagues found ethical CBPR described as requiring collaboration, tangible community benefit and shared control of data, and ethics committees often unfamiliar with its evolving methods.

Concordat to Support Research Integrity

Also called: the Concordat, UK Concordat, Research Integrity Concordat

The Concordat to Support Research Integrity is the UK's national framework for good research conduct, first published in 2012, reviewed in 2019 and refreshed in April 2025, and now hosted by the UK Committee on Research Integrity. It sets five commitments for researchers, research organisations and funders, founded on honesty, rigour, transparency and open communication, care and respect, and accountability, and many funders make it a condition of their grants.

Confidentiality

Also called: duty of confidentiality, research confidentiality

Confidentiality is a researcher's obligation to protect information entrusted by participants from unauthorised access, use or disclosure, and to report findings without identifying anyone who has not agreed to it. It differs from anonymity, because the researcher usually does know who said what. Its limits, such as a duty to report a risk of serious harm, should be explained before consent is given.

Conflict of interest

Also called: conflicts of interest, COI, competing interest, competing interests

A conflict of interest exists when a researcher's judgement about a primary interest, such as participants' welfare or the validity of research, may be swayed by a secondary interest, such as money, employment, a personal relationship or academic rivalry. The ICMJE treats the appearance of a conflict as mattering as much as an actual one. Journals ask authors to disclose relevant relationships, often calling them competing interests.

Consent form

Also called: informed consent form, ICF, consent document

A consent form is the document a participant signs, or otherwise marks, to record that they have been informed about a study and agree to take part, often with a statement for each point they are agreeing to. It records consent but does not replace the participant information sheet or the conversation with the researcher. UK drug trials need written consent, but many other studies may use oral consent.

Consultee

Also called: personal consultee, nominated consultee

A consultee, under the Mental Capacity Act 2005 in England and Wales, is a person asked for advice on whether an adult who lacks capacity would want to take part in research other than a drug trial. A personal consultee is someone who cares for the adult unpaid, and a nominated consultee is an independent professional. Consultees give advice, not consent, recorded on a declaration form.

Continuing review

Also called: continuing IRB review, annual IRB review, IRB renewal

Continuing review is the periodic re-examination by an institutional review board of research it has already approved, at intervals suited to the risk and at least once a year for studies approved at a convened meeting. Under the revised US Common Rule it is no longer required, unless the board decides otherwise, for studies eligible for expedited review or those that have reached data analysis only.

Contract cheating

Also called: assignment outsourcing, outsourced assessment

Contract cheating is having a third party, such as an essay mill, a tutor, a friend or a relative, complete assessed work that a student then submits as their own. The term was coined by Clarke and Lancaster in 2006. In England the Skills and Post-16 Education Act 2022 made it an offence to provide or arrange such a service commercially for students.

COPE flowcharts

Also called: COPE flowchart, COPE guidance flowcharts

COPE flowcharts are step-by-step guides from the Committee on Publication Ethics showing editors and others what to do when a particular problem is suspected, such as ghost, guest or gift authorship, an undisclosed conflict of interest, a request to remove an author, or signs of a paper mill. COPE publishes several dozen, covering authors, reviewers, editors and data publishers.

Data monitoring committee

Also called: DMC, data safety monitoring board, DSMB, data and safety monitoring board, data monitoring board, independent data monitoring committee, IDMC, data safety monitoring committee

A data monitoring committee is a group of independent experts, external to a clinical trial, that reviews its accumulating data on progress, safety and, where needed, effectiveness, and recommends to the sponsor whether the trial should continue, change or stop. Because it may review unblinded data, its independence matters. European guidance notes that not every trial needs one, though trials in life-threatening diseases usually do.

Data protection

Also called: data protection law, data privacy law

Data protection is the body of law and practice governing how personal data may be collected, used, stored and shared, which gives people rights over information about them. In the UK and EU it rests mainly on the GDPR, with the UK's Data Protection Act 2018 alongside, and it applies to research with special provisions that allow longer storage and some limits on individual rights.

Data protection impact assessment

Also called: DPIA

A data protection impact assessment is a documented process for identifying and reducing the risks to people's rights of a planned use of personal data, carried out before the processing starts. UK and EU data protection law require one when processing is likely to result in high risk, for example when it is large-scale, monitors people systematically or concerns vulnerable people.

Data retention

Also called: retention period, data retention period, research data retention

Data retention is the keeping of research data, and the decision about how long they are kept, for what purpose and when they are destroyed or archived. Periods are set by funders, institutions and law, and participants should be told about them before consenting. UK data protection law allows personal data processed only for research to be kept longer than otherwise, with safeguards.

Data sharing agreement

Also called: DSA, information sharing agreement, data use agreement, DUA, data sharing protocol

A data sharing agreement is a document agreed between organisations setting out which data will be shared, why, on what legal basis, how they will be kept secure, how long they will be kept and how people's rights will be handled. The UK Information Commissioner's Office calls it good practice rather than a legal requirement. In US health research, a data use agreement governs sharing of a limited data set.

De-identification

Also called: deidentification, de-identified data, de-identified

De-identification is the removal or masking of details that identify individuals in a dataset, such as names, addresses and record numbers. The word is used loosely: US health privacy rules give it a defined meaning, while the UK Information Commissioner's Office discourages using it as a synonym for anonymous or pseudonymised data, because de-identified data may still be re-identified.

Debriefing

Also called: debrief, participant debriefing, post-study debriefing

Debriefing is the explanation given to participants after they have taken part in a study, especially one using deception or incomplete disclosure, of its true purpose and why information was held back. It aims to correct misconceptions, deal with any distress and, where possible, let participants withdraw their data. Peer debriefing, a check on qualitative analysis, is a different practice.

Deception

Also called: deception in research, research deception, deceptive research

Deception in research is deliberately giving participants false information, or hiding a study's true purpose, because knowing it would change how they behave. UK guidance lets an ethics committee approve it only when no other method would work and important advances are likely, never to hide more than minimal risk, and with debriefing afterwards and an offer to withdraw data.

Declaration of Helsinki

Also called: Helsinki Declaration, WMA Declaration of Helsinki

The Declaration of Helsinki is the World Medical Association's statement of ethical principles for medical research involving human participants, including research on identifiable human material and data. First adopted in Helsinki in 1964 and most recently revised there in October 2024, it covers ethics committee review, informed consent, vulnerability, placebo use, post-trial provisions, registration before recruitment and publication of all results.

Disclosure of AI use

Also called: AI disclosure, AI use statement, AI declaration, declaring AI use, generative AI disclosure

Disclosure of AI use is a statement in a paper, thesis or assignment saying which artificial intelligence tools were used and for what, such as drafting, editing, analysing data or making images. COPE asks authors to describe it in the methods or a similar section, the ICMJE in the acknowledgements or methods, and the European Code of Conduct counts hiding AI use as an unacceptable practice.

How to declare AI use in an assignment

Distress protocol

Also called: distress management protocol, participant distress protocol

A distress protocol is a written plan setting out how researchers will recognise and respond to a participant becoming upset during data collection, and where to refer them for support. It is common in research on sensitive topics, and Draucker and colleagues described two, one for screening people at high risk of adverse reactions and one guiding interviewers during in-depth interviews.

Dual role

Also called: dual relationship, multiple relationship, practitioner-researcher dual role

A dual role is the position of a researcher who also holds another relationship with participants, such as a teacher studying their own pupils, a nurse their patients or a manager their staff. It can create pressure to take part and conflicts over confidentiality. BERA suggests making the researcher role explicit, involving an independent third party and reassuring people that declining is acceptable.

Duplicate publication

Also called: redundant publication, dual publication, overlapping publication

Duplicate publication is publishing a paper that overlaps substantially with one already published, without clear, visible reference to the earlier version. The ICMJE warns that it can distort the evidence through double counting of data. Openly acknowledged secondary publication, such as a translation agreed by both editors for a different readership, is acceptable.

Duty of care

Also called: researcher duty of care

A duty of care, in research, is the responsibility researchers and their institutions have to foresee and reduce harm to the people affected by a study, including participants, colleagues and researchers themselves. BERA asks researchers to think it through before collecting data, so they are ready to manage distress, and the Social Research Association stresses employers' duty towards researchers' own safety.

Dynamic consent

Also called: digital dynamic consent

Dynamic consent is an approach that uses a secure digital interface through which participants can follow how their data and samples are used, give or change consent for particular studies over time, and choose how they are contacted. Developed for biobanks and described by Kaye and colleagues in 2015, it is offered as an alternative to one-off broad consent for long-running research.

Electronic consent

Also called: eConsent, e-consent, online consent, digital consent

Electronic consent is the use of electronic methods, such as websites, tablets, videos and electronic signatures, to give participants information and to seek and record their consent. A joint statement from the Health Research Authority and the medicines regulator sets out the UK requirements, and HRA guidance treats an electronic signature as consent in writing, which matters for drug trials.

Equipoise

Also called: therapeutic equipoise

Equipoise is a state of genuine uncertainty about which of the treatments compared in a trial is better, widely held to be the ethical condition for randomly allocating patients between them. Freedman described the usual version as uncertainty in the individual investigator and argued instead for clinical equipoise, uncertainty in the expert community about the preferred treatment.

Ethical approval

Also called: ethics approval, ethics clearance, ethical clearance, favourable opinion, IRB approval, REC approval

Ethical approval is the formal decision of a research ethics committee or institutional review board that a proposed study may go ahead, often subject to conditions. In the UK health system it is called a favourable opinion. It must be in place before the research begins, and UKRIO lists proceeding without the necessary ethical approval among the examples of research misconduct.

Ethics application

Also called: ethical approval application, ethics submission, IRB application, IRB submission, REC application

An ethics application is the set of documents a researcher submits to a research ethics committee or institutional review board to seek approval for a study, usually including the protocol, participant information sheets, consent forms, recruitment material and data collection tools. In UK health and social care research it is made through the Integrated Research Application System, IRAS.

Ethics dumping

Ethics dumping is the export, by researchers from wealthy or otherwise privileged settings, of studies that would be unethical or unacceptable at home to lower-income or less-privileged settings with weaker oversight. Nature's journals now ask authors of such research questions drawn from the Global Code of Conduct for Research in Resource-Poor Settings, including whether the higher of two standards was followed.

Ethics in practice

Also called: process ethics, ethically important moments

Ethics in practice is the ethical judgement researchers exercise during a study, in ethically important moments that no approval form anticipates, such as a participant disclosing something unexpected or becoming upset. Guillemin and Gillam contrasted it with procedural ethics and linked it to reflexivity. Tracy's related scheme adds situational ethics, relational ethics and the ethics of leaving the field.

European Code of Conduct for Research Integrity

Also called: ALLEA code, ALLEA European Code of Conduct, European Code of Conduct

The European Code of Conduct for Research Integrity is a framework for good research practice in every discipline, published in 2017 and revised in 2023 by ALLEA, the federation of European academies. Its four principles are reliability, honesty, respect and accountability. It defines misconduct as fabrication, falsification and plagiarism and lists further unacceptable practices, and the European Commission treats it as the reference standard for research it funds.

Exempt research

Also called: exempt review, exempt determination, exempt category, exemption from IRB review

Exempt research, under the US Common Rule, is research with human participants that falls wholly within listed low-risk categories, such as normal educational practices, some surveys and interviews, or certain secondary uses of data, and so is excused from most of the rule's requirements, including full board review. Some categories still need a limited review, and exemption does not remove the duty to treat participants ethically.

Expedited review

Also called: expedited IRB review

Expedited review is review of a research proposal by the chair of an institutional review board, or experienced members the chair designates, instead of the full board. The US Common Rule allows it for research on an approved list that involves no more than minimal risk and for minor changes to approved research. Expedited reviewers may approve or require changes but cannot disapprove a study.

Fabrication

Also called: data fabrication, fabricating data, fabricated data

Fabrication is making up data or results and recording or reporting them as if they were real, such as inventing survey responses or participants who never existed. It is one of the three core forms of research misconduct, with falsification and plagiarism, in US federal rules, the European Code of Conduct and UK guidance. UKRIO notes it can include inventing records of patient consent.

Falsification

Also called: data falsification, falsifying data, falsified data

Falsification is manipulating research materials, equipment, images or processes, or changing, omitting or suppressing data or results without justification, so that the research record is inaccurate. It is one of the three core forms of research misconduct. It is unrelated to falsifiability, the philosophical idea that a scientific claim must be open to being proved wrong.

Four principles of biomedical ethics

Also called: four principles, four principles approach, principlism, Beauchamp and Childress principles, four principles of medical ethics

The four principles of biomedical ethics are respect for autonomy, beneficence, non-maleficence and justice, set out by Tom Beauchamp and James Childress in Principles of Biomedical Ethics and used across medical and research ethics. Each is a prima facie duty, binding unless it conflicts with another, and when two conflict the weight of each has to be judged in the particular case.

Gatekeeper consent

Also called: gatekeeper permission, gatekeeper approval, institutional permission, organisational consent

Gatekeeper consent is the permission a researcher obtains from a person or body that controls access to a setting or group, such as a head teacher, a ward manager or the moderators of an online forum, before approaching potential participants. It is usually required but never replaces each participant's own consent, and BERA says members covered by a leader's agreement should still be informed and free to opt out.

General Data Protection Regulation

Also called: GDPR, UK GDPR, EU GDPR

The General Data Protection Regulation is the European Union's data protection law, applied since 2018 and kept in UK law as the UK GDPR alongside the Data Protection Act 2018. For research it requires a lawful basis for processing personal data, an extra condition for special category data and safeguards such as data minimisation and pseudonymisation, while allowing longer storage and some exemptions from individual rights.

Ghost authorship

Also called: ghost author, ghostwriting, ghost writing

Ghost authorship is leaving off a paper's author list someone who made contributions that merited authorship or who helped write it. It hides who shaped the work and any interests behind it. In a survey of corresponding authors of articles published in 2008 in six leading medical journals, Wislar and colleagues found ghost authors in about 8 per cent.

Gift authorship

Also called: gift author, gifted authorship

Gift authorship is naming someone as an author who did not contribute enough to qualify, as a favour or in the hope that they will return it on their own papers. Writers draw the lines differently: some group it with guest authorship under honorary authorship, others treat it as a form of guest authorship. COPE has a single flowchart for ghost, guest or gift authorship.

Gillick competence

Also called: Gillick competent, Gillick-competent

Gillick competence is the standard, from English case law, under which a child under 16 who understands fully what is proposed and can weigh it up may consent to their own medical treatment. In the absence of specific law it is commonly applied to consent for research other than drug trials in England, Wales and Northern Ireland, where under-16s cannot consent to drug trials themselves.

Good Clinical Practice

Also called: GCP, ICH GCP, ICH-GCP, ICH E6

Good Clinical Practice is the international standard for designing, conducting, recording and reporting clinical trials so that participants' rights, safety and wellbeing are protected and the results are credible. It is set out in the ICH E6 guideline, whose third revision, E6(R3), took effect in the European Union in July 2025 and encourages a risk-based, proportionate approach.

Guest authorship

Also called: guest author

Guest authorship is including a senior or well-known person as an author, despite little or no contribution, in the hope that their name will make publication more likely or the paper more influential. Some writers use honorary authorship as the umbrella for all unearned authorship, while others treat honorary authorship as one form of guest authorship. Either way it breaches criteria such as the ICMJE's.

Helicopter research

Also called: helicopter science, parachute research, parachute science

Helicopter research is research in which investigators from high-income or otherwise privileged settings study lower-income settings or historically marginalised groups with little or no involvement of local researchers or communities in designing, conducting or publishing it. Nature's journals now ask such authors whether local scientists took part, whether the work is locally relevant and how its benefits will be shared.

HIPAA Privacy Rule

Also called: HIPAA, Privacy Rule, HIPAA research rules

The HIPAA Privacy Rule is the US federal rule governing how health plans, health care providers and other covered entities may use and disclose identifiable health information, including for research. Researchers usually need each person's authorisation, a waiver approved by an institutional review board or privacy board, de-identified data, or a limited data set under a data use agreement.

Honest error

Also called: honest mistake, genuine error

Honest error is a genuine mistake in research, such as a coding slip or a miscalculation, made without intent to deceive. US federal rules, UKRIO and COPE all distinguish it from misconduct, as they do reasonable differences of opinion. It still needs correcting, and COPE notes that serious honest error can justify retraction, which exists to correct the literature, not to punish.

Honorary authorship

Also called: honorary author, undeserved authorship, unjustified authorship

Honorary authorship is the listing as an author of someone who does not meet the criteria for authorship, often because of their seniority or position. Wislar and colleagues found honorary authors in about 18 per cent of articles published in 2008 in six high-impact medical journals. Usage varies: some writers make it the umbrella for gift and guest authorship, others one form of guest authorship.

Human research ethics committee

Also called: HREC, institutional ethics committee

A human research ethics committee is the Australian name for a committee that reviews research with people against the National Statement on Ethical Conduct in Human Research. Such committees were earlier called institutional ethics committees. Research carrying more than low risk must be reviewed by one, while lower-risk research may go through other institutional processes.

Human subjects research

Also called: human subject, human subjects, human participant research, research involving human participants

Human subjects research, under the US Common Rule, is research in which an investigator, professional or student, obtains information or biospecimens from living individuals through intervention or interaction, or obtains, uses or analyses identifiable private information or biospecimens about them. Whether an activity meets this definition decides whether it falls under the rule and needs institutional review board oversight.

ICMJE authorship criteria

Also called: ICMJE criteria, Vancouver authorship criteria, ICMJE authorship

The ICMJE authorship criteria are the four conditions every author must meet in the International Committee of Medical Journal Editors' recommendations: a substantial contribution to the design or to acquiring, analysing or interpreting data, drafting or critically revising the work, approving the final version, and agreeing to be accountable for all of it. Other contributors are acknowledged, and funding or supervision alone does not qualify.

Image manipulation

Also called: inappropriate image manipulation, inappropriate image duplication, figure manipulation

Image manipulation, as an integrity problem, is altering, duplicating or splicing images in a paper, such as microscope pictures or gel blots, in ways that misrepresent the original data. It is falsification when it changes what the image shows. Screening 20,621 papers from 40 journals, Bik and colleagues found problematic figures in 3.8 per cent, at least half suggesting deliberate manipulation.

Implied consent

Also called: implicit consent, tacit consent, consent by completion

Implied consent is consent inferred from what a person does, most often completing and returning an anonymous questionnaire after reading an information sheet, rather than from a signature or spoken agreement. Some ethics boards, such as the University of Victoria's, recommend it for minimal-risk anonymous surveys, while some UK universities now ask for an explicit tick box instead, so local rules should be checked.

Incidental finding

Also called: incidental findings, unexpected finding

An incidental finding is a discovery about a participant, made in the course of research, that lies outside the study's aims, such as an unexpected abnormality on a research scan. Canada's TCPS 2 requires researchers to plan for foreseeable ones and to disclose material findings, those with significant welfare implications, within the limits of what the participant agreed to receive.

Incomplete disclosure

Also called: partial disclosure, non-disclosure

Incomplete disclosure is telling participants less than the full purpose of a study, without actively misleading them, because full information would affect the results. The Belmont Report accepts it only when it is truly necessary, when no undisclosed risk is more than minimal and when there is a plan for debriefing. Canada's TCPS 2 treats partial disclosure alongside deception as an alteration of consent.

Indigenous data sovereignty

Also called: IDSov, Indigenous data rights

Indigenous data sovereignty is the assertion of Indigenous peoples' rights and interests in data about them, their territories and their ways of life, including the right to govern how such data are collected, owned and used. It draws on the UN Declaration on the Rights of Indigenous Peoples and underlies both the global CARE principles and the First Nations principles of OCAP in Canada.

Indigenous research ethics

Also called: research involving Indigenous peoples, ethics of Indigenous research

Indigenous research ethics are the principles and practices for research with or about Indigenous peoples, which add collective and community rights to the usual protections for individuals. Canada's TCPS 2, for example, requires researchers to engage the relevant community whenever research is likely to affect its welfare, and to respect its customs, codes of practice and research agreements.

Informed consent

Also called: free and informed consent, voluntary informed consent, valid consent, consent to participate

Informed consent is a person's voluntary agreement to take part in research, given after they have received and understood information about its purpose, procedures, risks, benefits and their right to withdraw, and while they have the capacity to decide. The Belmont Report analyses it into information, comprehension and voluntariness, and current guidance treats it as an ongoing process rather than a single signature.

Institutional review board

Also called: IRB, institutional review boards, independent review board

An institutional review board is the US name for a committee that reviews research with human participants to protect their rights and welfare, with the power to approve, require changes to or disapprove studies. Under the Common Rule it has at least five members of varied backgrounds, including at least one whose main concerns are not scientific. Elsewhere the same body is usually called a research ethics committee.

Integrated Research Application System

Also called: IRAS

The Integrated Research Application System, IRAS, is the single online system through which researchers apply for the permissions and approvals needed for health, social and community care research in the UK, including NHS research ethics committee review and HRA Approval. Applicants enter details of their project once, and the system filters its questions to suit the type of study and the approvals it needs.

Internet research ethics

Also called: online research ethics, digital research ethics, ethics of online research, IRE

Internet research ethics is the application of research ethics to studies of online spaces, platforms and data, such as forums, social media and digital traces. The Association of Internet Researchers' guidelines favour judgement in context over fixed rules, because expectations of privacy vary between settings, and consent from every user is often impracticable in large-scale studies.

Justice (research ethics)

Also called: justice, principle of justice, distributive justice in research

Justice, as a principle of research ethics, is the requirement that the benefits and burdens of research be shared fairly, so that no group is recruited because it is easy to reach or pressure, or left out of research that could help it. The Belmont Report applies it to choosing participants, citing the Tuskegee syphilis study, and it is one of the four principles of biomedical ethics.

Lawful basis for processing

Also called: lawful basis, legal basis for processing, Article 6 lawful basis

A lawful basis for processing is one of the six legal grounds, such as consent, public task or legitimate interests, that UK and EU data protection law requires for any use of personal data. The Health Research Authority advises that consent should not usually be the lawful basis in health research, universities and the NHS relying on public task, even though consent to take part is still needed ethically.

Legally authorised representative

Also called: legally authorized representative, LAR, legal representative, authorised third party, authorized third party

A legally authorised representative is a person or body permitted by law to consent on behalf of a prospective participant who cannot consent, such as a relative of an adult who lacks capacity. It is the US Common Rule's term. UK drug-trial law speaks of a personal or professional legal representative, and Canada's TCPS 2 of an authorised third party.

Limits of confidentiality

Also called: limits to confidentiality, breach of confidentiality, breaking confidentiality

The limits of confidentiality are the circumstances in which a researcher may have to pass on something a participant said in confidence, usually because of a risk of serious harm to the participant or others, or a legal duty to report, such as suspected abuse. ESRC and BERA guidance say participants should be told these limits before they consent, and researchers should seek advice before disclosing.

Minimal risk

Also called: no more than minimal risk, minimal-risk research

Minimal risk is a level of risk at which the probability and size of harm or discomfort expected in a study are no greater than those ordinarily met in daily life or during routine physical or psychological examinations. Under the US Common Rule it is the threshold for expedited review and for waiving consent. Canada's TCPS 2 uses a similar threshold for altering consent requirements.

National Statement on Ethical Conduct in Human Research

Also called: National Statement, NHMRC National Statement, Australian National Statement

The National Statement on Ethical Conduct in Human Research is Australia's national guideline for research with people, developed jointly by the National Health and Medical Research Council, the Australian Research Council and Universities Australia, and most recently updated in 2023. It rests on four values, research merit and integrity, justice, beneficence and respect, and scales ethics review to a study's level of risk.

Non-maleficence

Also called: nonmaleficence, do no harm, primum non nocere, principle of non-maleficence

Non-maleficence is the ethical principle of not causing harm, traced back to the Hippocratic maxim to do no harm. It is one of the four principles of biomedical ethics, where it is kept apart from beneficence, the positive duty to do good. In research it underlies the rule that participants must not be exposed to unnecessary risk and that harms must be weighed against benefits.

Nuremberg Code

Also called: Nuremberg Code of 1947

The Nuremberg Code is a ten-point set of principles for experiments on human beings, set out in 1947 in the judgement of the Doctors' Trial at Nuremberg, where German physicians and administrators were tried for crimes that included experiments on prisoners. Its first principle is that the voluntary consent of the human subject is absolutely essential, and it also gives subjects the liberty to end their participation.

OCAP principles

Also called: OCAP, ownership, control, access and possession

The OCAP principles, ownership, control, access and possession, are First Nations standards in Canada for how data and information about First Nations communities should be collected, protected, used and shared. Dating from the 1990s and now a registered trademark of the First Nations Information Governance Centre, they are cited in Canada's TCPS 2 and informed the global CARE principles.

Ongoing consent

Also called: process consent, continuing consent, consent as a process

Ongoing consent is the practice of treating consent as something maintained throughout a study, not given once at the start, by checking that participants still agree as the research develops or new information emerges. The Health Research Authority calls consent an iterative, ongoing process, Canada's TCPS 2 requires it, and ESRC guidance notes that participatory research often treats it as open-ended.

Opt-in consent

Also called: opt-in, opt in, active consent

Opt-in consent is a consent process in which people are included in research only if they actively agree, for example by returning a reply slip or signing up. BERA calls it the preferred approach. In a randomised trial, Junghans and colleagues found that opt-in recruitment gave a lower response rate and a less representative sample than opt-out, a trade-off ethics committees have to weigh.

Opt-out consent

Also called: opt-out, opt out, passive consent

Opt-out consent is a process in which people are included in research unless they actively decline, for example parents who return a form only if they do not want their child to take part. BERA prefers opt-in but accepts opt-out in some circumstances, after checking local law and practice, and Junghans and colleagues proposed it as the default for low-risk studies.

Outcome switching

Also called: outcome misreporting, switched outcomes

Outcome switching is reporting a trial with primary or secondary outcomes that differ from those pre-specified in its protocol or registry entry, by dropping, adding or swapping outcomes without saying so. The COMPare project checked every trial published over six weeks in five major medical journals and wrote to the journals about each misreported one. The distortion it leaves in the evidence is outcome reporting bias.

Paper mill

Also called: paper mills, paper factory

A paper mill is a business that produces fraudulent manuscripts made to look like genuine research and sells them, or authorship on them, to researchers under pressure to publish. STM, the publishers' association, describes paper mills as covert, profit-driven operations that also manipulate peer review and increasingly use AI, and COPE's guidance allows batch retraction of papers traced to them.

Parental consent

Also called: parental permission, parent consent, guardian consent

Parental consent is the agreement of a parent or guardian for their child to take part in research, which US regulations call permission. It is usually sought alongside the child's own assent. In the UK a competent young person may consent for themselves to research other than drug trials, and US rules allow parental permission to be waived in some cases, such as research on neglected or abused children.

Participant information sheet

Also called: PIS, information sheet, patient information sheet, participant information leaflet, information letter, plain language statement

A participant information sheet is a plain-language document telling potential participants what a study is about, why they have been invited, what taking part involves, its possible risks and benefits, how their data will be handled and that they may decline or withdraw. It is given with time to consider before consent is sought, and the Health Research Authority advises testing it with members of the public.

Patient and public involvement

Also called: PPI, public involvement, PPIE, patient and public involvement and engagement, service user involvement

Patient and public involvement is research carried out with or by members of the public, including patients, carers and service users, rather than to, about or for them, for example by bringing lived experience to judgements about what matters and whether plans will work. The NIHR distinguishes it from public engagement, which means communicating and sharing information about research with the public.

Personal data

Also called: personal information, personally identifiable information, PII

Personal data are any information relating to an identified or identifiable living person, who can be identified directly, for example by name, or indirectly, for example through a combination of location, an online identifier or physical, economic or cultural characteristics. Under UK and EU data protection law they include pseudonymised data but not truly anonymous data or information about people who have died.

Placebo ethics

Also called: ethics of placebo use, ethics of placebo controls, placebo-controlled trial ethics

Placebo ethics is the question of when it is acceptable to give trial participants a placebo, or no treatment, instead of an active one. The 2024 Declaration of Helsinki allows it when no proven intervention exists, or for compelling and scientifically sound methodological reasons, provided participants face no added risk of serious or irreversible harm from missing the best proven treatment.

Plagiarism

Also called: plagiarising, plagiarizing, research plagiarism, academic plagiarism

Plagiarism is using another person's ideas, processes, results, words or other work without giving appropriate credit, whether by copying text, paraphrasing closely without citation or taking ideas from a manuscript one is reviewing. It is one of the three core forms of research misconduct and a breach of academic integrity in student work. US federal rules exclude self-plagiarism and authorship disputes from it.

Check plagiarism

Post-trial access

Also called: post-trial provisions, post-trial obligations, continued access

Post-trial access is the continued provision, after a clinical trial ends, of an intervention found beneficial and reasonably safe to participants who still need it. The 2024 Declaration of Helsinki requires sponsors and researchers to arrange it before the trial, whether they provide it themselves or through health systems or governments, and to tell participants about it during consent.

Predatory publishing

Also called: predatory journal, predatory journals, predatory publisher, predatory conference

Predatory publishing is the activity of journals, publishers and conferences that put self-interest before scholarship, marked by false or misleading information, departure from good editorial practice, lack of transparency or aggressive, indiscriminate solicitation, in a consensus definition agreed in 2019. The European Code of Conduct counts knowingly setting up, supporting or using such outlets as an unacceptable research practice.

Privacy

Also called: participant privacy, right to privacy

Privacy, in research ethics, is a person's right to be free from intrusion and to control information about themselves, their body, their communications and the spaces they occupy. It differs from confidentiality, the researcher's duty to safeguard information once it has been shared, and from security, the measures used to protect it. Canada's TCPS 2 separates the three in this way.

Procedural ethics

Also called: ethics approval process

Procedural ethics is the formal side of research ethics: applying to an ethics committee, gaining approval and following the agreed protocol and safeguards. Guillemin and Gillam contrasted it with ethics in practice, the judgements needed as situations arise during a study, and Reid and colleagues note that approval procedures can overshadow those later ethical tensions.

Proportionate review

Also called: proportionate review service, PRS

Proportionate review is the UK Health Research Authority's faster route through ethics review for studies that raise no material ethical issues, in which a sub-committee of research ethics committee members reviews the application by email, teleconference or a smaller meeting instead of a full committee meeting. It is the nearest UK counterpart to expedited review in the United States.

Prospective registration

Also called: prospective trial registration, registration before enrolment, prior registration

Prospective registration is registering a study, with its planned outcomes and methods, before the first participant is enrolled. The ICMJE makes it a condition of publication for trials and rejects late registration, because only registration in advance guards against selective reporting and duplicated research. A trial registered after enrolment began is retrospectively registered.

Proxy consent

Also called: surrogate consent, third-party consent, substitute consent, consent by a representative

Proxy consent is consent to take part in research given on behalf of someone who cannot consent for themselves, such as a young child or an adult who lacks capacity, by a parent, relative or other legally recognised person. Who may act, and whether they consent or only advise, depends on the law. In England and Wales, consultees for most research other than drug trials give advice, not consent.

Pseudonym

Also called: participant pseudonym, false name

A pseudonym is an invented name used in place of a participant's real name, or of a place or organisation, in research reports, especially qualitative studies that quote people. It is one step in anonymising a report but is rarely enough on its own, because other details can still identify someone, and some participants ask to be named instead.

Pseudonymisation

Also called: pseudonymization, pseudonymised data, pseudonymized data, coded data, key-coded data

Pseudonymisation is the replacement of direct identifiers, such as names, with codes or pseudonyms, while the key that links them back is kept separately and securely. It reduces risk, but under UK and EU data protection law pseudonymised data remain personal data for anyone who holds the key, unlike truly anonymous data. Canada's TCPS 2 calls such data coded information.

Public online posts

Also called: publicly available online data, public social media posts, public posts

Public online posts are messages, comments and images anyone can see without logging in, such as open forum threads or public social media posts, and using them in research raises the question of whether visible means fair to use. The Association of Internet Researchers notes that users often see such spaces as private and rarely write for researchers, and that a quoted post can be traced by searching its words.

Questionable research practices

Also called: QRPs, QRP, detrimental research practices

Questionable research practices are behaviours that fall short of good research practice without reaching the definition of misconduct, such as selective reporting, inaccurate citing, misuse of statistics or withholding data without reason. UKRIO notes that the UK Concordat's definition includes avoidable errors. Statistical forms such as p-hacking and HARKing are defined in the statistics section.

Re-identification

Also called: reidentification, re-identification risk, risk of re-identification, de-anonymisation, de-anonymization

Re-identification is the linking of supposedly anonymous or de-identified data back to the individuals they describe, for example by matching them with other datasets or by spotting someone with rare characteristics in a small area. The risk grows as datasets are linked, which is why Canada's TCPS 2 counts data as anonymised only when that risk is low or very low.

Reduction (animal research)

Also called: reduction principle

Reduction, in animal research, is the use of methods that minimise the number of animals used in each experiment or study while still meeting its scientific aims, for example through sound experimental design and analysis, imaging that allows repeated measures in the same animal, and sharing data and resources. It is one of the 3Rs.

Refinement (animal research)

Also called: refinement principle

Refinement, in animal research, is the use of methods that minimise pain, suffering, distress or lasting harm to research animals and improve their welfare, from housing and husbandry to the procedures themselves, such as appropriate anaesthesia and pain relief. It is one of the 3Rs, and the NC3Rs notes that pain and suffering can also distort experimental results.

Replacement (animal research)

Also called: replacement principle, non-animal methods, full replacement, partial replacement

Replacement, in animal research, is the use of approaches that avoid or replace animals where they would otherwise have been used, such as human volunteers, cells and tissues, and computer models. The NC3Rs distinguishes full replacement from partial replacement, which uses animals thought unable to suffer, such as some invertebrates, or cells from animals killed only for that purpose. It is one of the 3Rs.

Research ethics

Also called: ethics in research, research ethics principles

Research ethics is the field that sets out how research should treat the people, animals, communities and environments it involves, covering consent, risk and harm, privacy, fairness and the review of studies before they begin. It overlaps with research integrity, which concerns honesty and rigour in doing and reporting research, and the two are often governed by the same codes and committees.

Research ethics committee

Also called: REC, ethics committee, research ethics board, REB, ethical review board, ethics review board, independent ethics committee, IEC

A research ethics committee is an independent group of expert and lay members that reviews proposed research involving people, their tissue or their data, to safeguard participants' rights, safety, dignity and wellbeing, and gives or withholds a favourable opinion. It is the usual UK name, and the NHS alone has more than 80 such committees. The US equivalent is an institutional review board, the Canadian a research ethics board.

Research incentive

Also called: participant incentive, incentive payment, participant payment, payment for participation, paying research participants

A research incentive is anything offered to encourage people to take part in a study, such as money, vouchers or entry to a prize draw. Canada's TCPS 2 separates incentives from reimbursement of expenses and compensation for injury, neither recommends nor discourages them, and warns against offers large enough to override judgement about risk. BERA asks that their use be reported.

Research integrity

Also called: integrity in research, scientific integrity

Research integrity is the practice of proposing, doing, reviewing and reporting research according to values such as honesty, rigour, transparency, care and accountability, so that others can trust and build on it. Frameworks such as the UK Concordat and the European Code of Conduct set out what it asks of researchers, institutions and funders, and research misconduct is its most serious breach.

Research misconduct

Also called: scientific misconduct, FFP, research fraud, scientific fraud

Research misconduct is fabrication, falsification or plagiarism in proposing, performing, reviewing or reporting research, the core definition in US federal rules and European and UK codes. It excludes honest error and differences of opinion. UK guidance also counts misrepresentation, research without required approvals and undeclared conflicts of interest, and a US finding needs intent, knowledge or recklessness.

Research misconduct investigation

Also called: misconduct investigation, misconduct inquiry, research integrity investigation

A research misconduct investigation is the formal process by which an institution examines an allegation of misconduct against one of its researchers. It usually starts with a preliminary assessment, called an inquiry in US rules, of whether there is enough evidence, then a full investigation by a panel that establishes the facts and recommends action, including correcting the published record.

Research sponsor

Also called: sponsor, study sponsor, trial sponsor

A research sponsor is the organisation or partnership that accepts overall responsibility for making sure a study is properly set up, run and reported, which in UK health research is a formal role under the policy framework. In clinical trials the sponsor also receives a data monitoring committee's recommendations about continuing, changing or stopping the trial, and must arrange any post-trial provisions.

Research with children

Also called: children in research, research involving children, paediatric research ethics, pediatric research ethics

Research with children is research in which people below the legal age of consent take part, and it usually needs a parent's or guardian's permission together with the child's own assent where the child can give it. US regulations cap the risk allowed according to the prospect of direct benefit, and UK and BERA guidance stress information children can understand and respect for their refusal.

Research with prisoners

Also called: prisoner research, prisoners as research participants

Research with prisoners is research in which people held in prison or other detention take part, and it carries extra protections because confinement may prevent a truly voluntary choice. Under subpart C of the US human subjects regulations it is limited to certain categories, a prisoner or prisoner representative must sit on the reviewing board, and advantages offered must not distort judgement of the risks.

Research without prior consent

Also called: deferred consent, RWPC, emergency research without prior consent

Research without prior consent, often called deferred consent, is enrolment of people in emergency research before consent can be sought, because treatment is urgent and neither they nor a representative can be asked in time. In the UK the procedure must be approved in advance by a research ethics committee, and consent, or a consultee's advice, must be sought as soon as possible afterwards.

Researcher safety

Also called: researcher wellbeing, fieldwork safety, safety of researchers

Researcher safety is the protection of researchers from physical and psychological harm while collecting data, for instance when interviewing alone in people's homes or hearing distressing accounts. The Social Research Association's code identifies risks from physical threat to psychological trauma, and asks employers and researchers to plan for them through risk assessment, keeping in contact during fieldwork and support afterwards.

Respect for persons

Also called: principle of respect for persons

Respect for persons is the first principle of the Belmont Report, joining two convictions: that individuals should be treated as autonomous agents able to make their own choices, and that people with diminished autonomy, such as young children or some people who are ill, are entitled to protection. It is the basis of informed consent, and Canada's TCPS 2 uses the same name for one of its core principles.

Retraction for misconduct

Also called: misconduct retraction, retracted for misconduct

A retraction for misconduct is the withdrawal of a published article because fabrication, falsification, plagiarism, fictitious authorship or other misconduct means its findings cannot be relied upon. COPE stresses that retraction exists to correct the literature, not to punish, that honest error can also justify it, and that the notice should give the reason. Papers from paper mills may be retracted in batches.

What a retraction is and how to check your reference list for one

Right to withdraw

Also called: withdrawal from research, withdrawing consent, withdrawal of consent, right of withdrawal

The right to withdraw is a participant's right to stop taking part in a study at any time, without giving a reason and without penalty or loss of benefits they are otherwise entitled to. Information sheets should say what will happen to data already collected, since removal may become impossible once data are merged or anonymised, and Canadian guidance says payment already earned should not be withheld.

Risk-benefit assessment

Also called: risk-benefit analysis, risk/benefit assessment, benefit-risk assessment, risk-benefit ratio

A risk-benefit assessment is the systematic weighing of the probability and size of a study's possible harms against its expected benefits, to participants and to knowledge. Researchers use it to test a design and ethics committees to judge whether the risks are justified. The Belmont Report derives it from beneficence and asks that risks be cut to those needed to reach the research objective.

Safeguarding

Also called: safeguarding in research, preventing harm in research

Safeguarding, in research, is taking all reasonable steps to prevent harm, particularly exploitation, abuse and harassment, to anyone involved in or affected by a study, including participants, communities and researchers. UK Research and Innovation's policy applies to all the work it funds, in the UK and overseas, and researchers working with children may have duties to report concerns.

Salami slicing

Also called: salami publication, salami publishing, fragmented publication, least publishable unit, piecemeal publication

Salami slicing is splitting the results of one study into several separate papers, each reporting a thin slice, to increase the number of publications. The European Code of Conduct lists it as an unacceptable practice, and the Text Recycling Research Project calls it fragmented publication, warning that it exaggerates the novelty of the work.

Secure data storage

Also called: data security, secure storage, research data security

Secure data storage is keeping research data in ways that prevent unauthorised access, loss or alteration, through physical, administrative and technical safeguards. BERA lists current practice as secure networks, locked storage for paper records, password protection or encryption for electronic files, secure transfer, and data-sharing agreements with anyone else who uses the data.

Self-plagiarism

Also called: self plagiarism, auto-plagiarism

Self-plagiarism is presenting one's own earlier work as new, such as republishing substantial parts of a paper, including in translation, without citing it, or reusing work already assessed, which Edinburgh's rules, for example, count as plagiarism. The European Code of Conduct counts it as an unacceptable practice, US federal rules exclude it from research misconduct, and some writers prefer the neutral term text recycling for disclosed reuse.

What is self-plagiarism, and can I reuse my own work?

Service evaluation

Also called: service improvement project, service development project

A service evaluation is a project designed to define or judge a current health or care service, answering the question of what standard it achieves, without reference to a set standard and using only interventions already in use. The UK Health Research Authority does not class it as research, so it needs no research ethics committee review, although it has governance requirements of its own.

Singapore Statement on Research Integrity

Also called: Singapore Statement

The Singapore Statement on Research Integrity is a short global guide to the responsible conduct of research, agreed at the second World Conference on Research Integrity in Singapore in July 2010. It sets out four principles, honesty, accountability, professional courtesy and fairness, and good stewardship, with fourteen responsibilities for researchers and institutions, and it states that it is not a regulatory document.

Special category data

Also called: special categories of personal data, sensitive personal data

Special category data are personal data that UK and EU law protects more strictly: racial or ethnic origin, political opinions, religious or philosophical beliefs, trade union membership, genetic and biometric data, health, and sex life or sexual orientation. Processing them needs a lawful basis plus a further condition, such as the one for archiving, research and statistics. Criminal offence data have separate rules.

Substantial amendment

Also called: substantial modification

A substantial amendment is a change to an approved study that is likely to have a substantial impact on the safety or rights of participants, or on the reliability or robustness of its data, and so needs the ethics committee's favourable opinion before it is made. In the Health Research Authority's current terms for non-drug studies it is a substantial modification, and smaller changes are minor modifications.

TCPS 2

Also called: Tri-Council Policy Statement, TCPS, Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans

TCPS 2, the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans, is Canada's joint research ethics policy, whose current edition dates from 2022. It expresses respect for human dignity through three core principles, respect for persons, concern for welfare and justice, requires free, informed and ongoing consent, and has a chapter on research with First Nations, Inuit and Métis peoples.

Text recycling

Also called: recycled text, reuse of own text

Text recycling is the reuse of material, such as prose, figures or equations, from an earlier document in a new one, identical or nearly so and not presented as a quotation, where at least one author wrote both. The Text Recycling Research Project holds that it can be appropriate if the material is accurate, allowed by copyright and publisher policy, and disclosed.

What is self-plagiarism, and can I reuse my own work?

Therapeutic misconception

The therapeutic misconception is a participant's failure to understand that the defining purpose of a clinical trial is to produce generalisable knowledge rather than individually tailored care, for example believing their treatment group was chosen for their own benefit. Appelbaum and colleagues named it in 1982, and it differs from overestimating the chance of benefit, sometimes called therapeutic misestimation.

Tuskegee syphilis study

Also called: Tuskegee study, Tuskegee experiment, Untreated Syphilis Study at Tuskegee, USPHS Untreated Syphilis Study at Tuskegee

The Tuskegee syphilis study was a US Public Health Service study, run from 1932 to 1972, that observed the untreated course of syphilis in Black men in Macon County, Alabama, without their informed consent and without offering treatment even once it was easily available. Its exposure changed standard research practice, and the Belmont Report cites it as an example of injustice in choosing participants.

Undue inducement

Also called: excessive inducement

Undue inducement is an offer, usually of money or another benefit, so large or attractive that it could lead people to take part in research against their better judgement or to accept risks they would otherwise refuse. The Belmont Report notes that an inducement acceptable for most people may become undue for someone especially vulnerable, which is why committees judge payments against the people being recruited.

Undue influence

Also called: undue pressure

Undue influence is pressure that undermines the voluntariness of a decision to take part in research without an outright threat, whether through an excessive reward or through the sway of someone with authority, such as a teacher, employer, doctor or prison officer. Canada's TCPS 2 says it should be judged from the participant's point of view, since even a caring relationship of dependency can create it.

Verbal consent

Also called: oral consent, recorded verbal consent, audio-recorded consent

Verbal consent is agreement to take part in research given aloud rather than in writing, usually after the information has been read out or explained, and ideally audio-recorded. ESRC guidance says it can be ethically preferable where a signature would itself put people at risk, as in some research with drug users or political activists, though UK drug trials require consent in writing.

Voluntariness

Also called: voluntary participation, voluntary consent

Voluntariness is the condition that a decision to take part in research, or to carry on, is made freely, without coercion, undue influence or unjustifiable pressure. It is one of the three elements of informed consent in the Belmont Report, with information and comprehension, and it is most at risk where recruiters hold power over potential participants, as teachers, employers and doctors may.

Vulnerable groups

Also called: vulnerable populations, vulnerable participants, vulnerable people, vulnerability in research

Vulnerable groups, in research ethics, are people at greater risk of being wronged or harmed in research, or less able to protect their interests, such as children, prisoners and people with impaired decision-making capacity. The 2024 Declaration of Helsinki treats vulnerability as fixed or contextual, asks for specific protections, and says the harms of excluding such groups must be weighed against those of including them.

Waiver of consent

Also called: waiver of informed consent, consent waiver, alteration of consent

A waiver of consent is an ethics committee's permission to carry out research without obtaining informed consent from participants, typically for secondary use of records where asking everyone would make the study impossible. Under the US Common Rule the research must involve no more than minimal risk, be impracticable without the waiver and not harm participants' rights and welfare. A waiver of documentation only removes the signature.

Whistleblowing

Also called: whistle-blowing, whistleblower, whistle-blower, raising concerns

Whistleblowing, in research, is raising a concern in good faith about possible misconduct or other wrongdoing with an institution, a funder or another body. The European Code of Conduct asks institutions to protect genuine whistle-blowers so their careers are not harmed, US rules require protection of good-faith complainants from retaliation, and a malicious accusation is itself an unacceptable practice.