Ethical Considerations – Types, Examples and Writing Guide

Ethical considerations in research are the principles and safeguards used to protect participants, communities, researchers and the integrity of knowledge. They include informed and voluntary consent, minimising harm, fair participant selection, privacy, confidentiality, scientific validity, independent review, honest reporting and responsible management of data throughout the research lifecycle.

Research can produce valuable knowledge, improve services and help solve social, scientific and health problems. It can also expose people to physical, psychological, social, legal, financial or privacy-related harm. Ethical research therefore requires more than choosing an interesting topic or following a technically sound method.

This guide explains the main ethical principles, how they apply at each stage of a study, and how researchers can address them in proposals, dissertations and published papers. It also covers online data, artificial intelligence, vulnerable groups, research integrity and differences between ethical approval, legal compliance and responsible conduct.

Key takeaways

• Ethical responsibility begins when a study is designed and continues after publication or data sharing.
• A study must have sufficient social or scientific value and a valid design to justify participants’ time, burden or risk.
• Consent must be informed, understandable and voluntary; a signed form alone does not prove valid consent.
• Confidentiality, anonymity, privacy and legal data protection are related but different concepts.
• Ethical approval does not remove the researcher’s continuing responsibility to respond to new risks or protocol changes.
• Publicly accessible online data and AI tools can still create consent, privacy, bias and confidentiality concerns.

What Are Ethical Considerations in Research?

Ethical considerations in research are the moral principles, professional standards and practical safeguards that guide how a study is planned, reviewed, conducted, analysed and reported.

Their purpose is to ensure that:

• People are treated with dignity and respect.
• Participation is voluntary wherever consent is required.
• Risks are minimised and justified.
• Benefits and burdens are distributed fairly.
• Personal information is handled responsibly.
• The study is capable of producing meaningful knowledge.
• Findings are recorded and communicated honestly.
• Researchers remain accountable for their decisions.

Although ethics is often associated with studies involving human participants, ethical issues can also arise in animal research, environmental studies, laboratory work, secondary-data analysis, historical research, internet research, artificial intelligence, and studies that may affect identifiable communities.

Why Are Ethical Considerations Important?

Ethical considerations protect people and improve the trustworthiness of research. They reduce avoidable harm, preserve individual autonomy, promote fairness, support public confidence and help ensure that research resources are used for scientifically worthwhile purposes.

They protect participants

Participants may disclose sensitive experiences, undergo physical procedures, experience discomfort or allow researchers to access personal data. Ethical safeguards reduce the chance that participation will lead to distress, discrimination, financial loss, damaged relationships, legal consequences or loss of privacy.

They protect groups and communities

Removing names does not eliminate every possible harm. A study may stigmatise a cultural group, neighbourhood, occupation, patient population or online community. Researchers should therefore consider collective and reputational consequences as well as individual risks.

They improve scientific quality

An invalid or badly designed study may expose participants to inconvenience or risk without producing reliable knowledge. Scientific validity is therefore not only a methodological concern; it is an ethical requirement.

They support public trust

Research depends on cooperation among participants, researchers, institutions, funders, journals and communities. Misleading recruitment, data breaches, concealed conflicts or dishonest reporting can reduce trust in both the study and the wider research system.

They clarify responsibility

Ethical planning requires researchers to decide in advance how they will handle consent, distress, withdrawal, disclosure of harm, data retention, unexpected findings, adverse events and requests for access to results.

Ethics, Law, Research Integrity and Ethical Approval

These terms overlap, but they are not interchangeable.

Concept	Main question	Example
Research ethics	Is the study morally responsible and respectful?	Whether observing an online support group without consent is justified
Law and regulation	Is the activity legally permitted and compliant?	Whether there is a lawful basis for processing personal data
Research integrity	Is the research honest, rigorous and accountable?	Whether results were fabricated, selectively omitted or plagiarised
Ethical approval	Has an authorised review body assessed the proposed study?	Approval from an IRB or Research Ethics Committee
Professional guidance	Does the study follow discipline-specific standards?	Following psychology, medical, educational or internet-research guidance

An action can be legal but still ethically questionable. A researcher may also receive approval for a protocol but later act unethically by changing recruitment procedures, collecting additional data without permission, ignoring participant distress or sharing data in an unsafe way.

Core Ethical Principles in Research

Ethics frameworks use different terminology, but most converge on a common group of principles.

Principle	Meaning	Practical application
Social or scientific value	The research should address a worthwhile question.	Explain who may benefit and why the knowledge is needed.
Scientific validity	The design must be capable of answering the question.	Use suitable methods, adequate sampling and a justified analysis plan.
Respect for persons	Individuals should be treated as autonomous agents.	Provide meaningful choices and protect people with reduced autonomy.
Informed consent	Participants should understand what participation involves.	Use clear information, comprehension checks and documented consent where appropriate.
Voluntary participation	Participation must be free from coercion or inappropriate pressure.	Separate recruitment from grading, employment, treatment or authority relationships.
Beneficence	Research should aim to produce benefit.	Design the project to generate useful knowledge or practical value.
Non-maleficence	Researchers should avoid or minimise harm.	Reduce unnecessary procedures, questions, exposure and data collection.
Justice	Benefits and burdens should be distributed fairly.	Recruit according to scientific reasons rather than convenience or manipulability.
Privacy and confidentiality	People should retain appropriate control over access to themselves and their information.	Limit collection, restrict access and protect identities in outputs.
Independent review	People outside the immediate research team should assess the protocol when required.	Obtain IRB, REC or institutional review before beginning covered activities.
Integrity and transparency	Researchers must describe methods and findings honestly.	Do not fabricate, falsify, plagiarise or conceal material limitations.
Accountability	Researchers remain responsible for foreseeable and emerging consequences.	Record decisions, report incidents, seek amendments and correct errors.

Respect for persons

Respect means recognising that individuals have values, preferences and the right to make decisions about matters affecting them. Researchers should not treat participants merely as sources of data.

Respect also requires added protection when a person’s ability to make an independent decision is limited by age, cognitive capacity, illness, detention, dependency, financial pressure or an unequal relationship.

Beneficence and non-maleficence

Beneficence means seeking worthwhile benefits. Non-maleficence means avoiding unnecessary harm.

Researchers should ask:

• What benefit could reasonably result?
• Who is likely to receive that benefit?
• What burdens or risks will participants carry?
• Can the same question be answered through a less intrusive method?
• Are the remaining risks proportionate to the study’s value?

A promise of general social benefit cannot justify unlimited risk.

Justice

Justice concerns fairness in recruitment and access to research benefits. Researchers should not repeatedly recruit disadvantaged populations simply because they are accessible, inexpensive or less able to refuse.

At the same time, automatically excluding children, older adults, pregnant people, disabled people or minority populations can also be unjust when the findings are intended to apply to them. Inclusion and protection must be balanced rather than treated as opposites.

Social value and scientific validity

A research question should be important enough to justify the resources, participant burden and possible risk involved.

Examples of ethically weak design include:

• Repeating a study unnecessarily without reviewing existing evidence
• Using a sample too small to answer the main question
• Choosing measures that do not represent the intended concept
• Collecting extensive personal data that will not be analysed
• Conducting interviews without an adequate analysis plan
• Exposing participants to an intervention without appropriate monitoring

Methodological weakness becomes an ethical problem when people accept burdens for knowledge that the design is unlikely to produce.

A Ten-Step Process for Ethical Research

1. Establish the value of the question

Explain the knowledge gap and why answering it matters. Review existing studies so that participants are not exposed to unnecessary duplication.

2. Identify affected people and groups

Consider participants, non-participants, communities, researchers, partner organisations, data subjects and people who may be affected by publication.

3. Map possible risks and benefits

Assess physical, psychological, social, privacy, legal, economic, cultural, environmental and reputational effects. Consider both likelihood and seriousness.

4. Select participants fairly

Use inclusion and exclusion criteria that follow from the research question. Examine whether convenience, vulnerability or authority relationships are influencing recruitment.

5. Design consent and recruitment procedures

Prepare clear information about the purpose, procedures, duration, risks, benefits, data use, withdrawal, compensation and contact routes.

6. Prepare a data-management plan

Specify:

• What data will be collected
• Why each item is necessary
• Where data will be stored
• Who will have access
• How identifiers will be separated
• How long data will be retained
• Whether data will be shared or reused
• How files will be securely deleted or archived

7. Obtain the required review and permissions

Submit the protocol and supporting documents to the appropriate institutional or external body. Separate permissions may be required from research sites, data controllers, schools, health services, community organisations or national authorities.

8. Train the research team

Team members should understand consent, confidentiality, distress procedures, safeguarding, data security, incident reporting and their limits of responsibility.

9. Monitor the study and manage changes

Ethics is ongoing. Record unexpected problems, adverse events, complaints, deviations and emerging privacy risks. Obtain approval for material amendments before implementing them unless immediate action is required to protect safety.

10. Analyse, report, share and close responsibly

Report findings accurately, including null or inconvenient results where relevant. Protect identities in quotations and datasets. Follow approved retention, sharing and destruction arrangements.

Informed Consent and Voluntary Participation

Informed consent is an ongoing process through which a person receives relevant information, understands it and freely agrees to participate. It is not simply a signature on a form.

What makes consent valid?

Valid consent is generally:

• Informed: Relevant information has been provided.
• Understood: The person can comprehend what participation means.
• Voluntary: The decision is free from coercion and inappropriate pressure.
• Given by a person with capacity: The participant can make the relevant decision, with support where appropriate.
• Specific enough: The intended activities and data uses are adequately described.
• Documented when required: The process is recorded in an appropriate form.
• Ongoing: The researcher remains attentive to willingness throughout participation.

What should participants be told?

Information normally includes:

1. The study’s purpose
2. Why the person was invited
3. What they will be asked to do
4. How long participation will take
5. Foreseeable risks or discomforts
6. Expected benefits, including when no direct benefit is expected
7. Whether participation is voluntary
8. Withdrawal arrangements
9. Payment, expenses or other incentives
10. What data will be collected
11. How privacy and confidentiality will be protected
12. Whether data may be shared or reused
13. Limits to confidentiality
14. Researcher and complaint contacts
15. Any relevant alternatives to participation

Consent should be understandable

Long, technical forms do not necessarily produce informed decisions. Use plain language, accessible formats, translations and opportunities for questions.

A comprehension check may be appropriate for complex or higher-risk studies. For example, a researcher might ask the participant to explain in their own words what the study involves and what they can do if they wish to stop.

Voluntariness and power relationships

A participant may technically say “yes” while feeling unable to refuse.

Risk of undue influence is higher when researchers recruit:

• Their own students
• Employees they manage
• Their patients or clients
• Prisoners or detained people
• People dependent on services
• Individuals in severe financial need
• People seeking immigration, educational or medical assistance

Safeguards may include neutral recruitment, alternative course credit, delayed access to identities, modest compensation, private decision-making and a clear statement that refusal will not affect services, grades, employment or treatment.

Payment and incentives

Payment is not automatically coercive. Researchers may reasonably reimburse expenses, compensate time or recognise inconvenience.

The ethical question is whether the offer is so influential that it could distort a person’s assessment of the risks or make economically disadvantaged people feel unable to refuse. The amount, population, burden and context should be assessed together.

Right to withdraw

Researchers should explain:

• How to withdraw
• Until what point withdrawal is possible
• Whether already collected data can be removed
• What happens after data have been irreversibly anonymised
• Whether withdrawal affects payment
• Whether safety follow-up may still be necessary

Researchers should not promise deletion after a stage at which individual data can no longer be identified.

When consent may be altered or waived

Some research may use existing records, unobtrusive observation, emergency procedures, deception or very large datasets where individual consent is impracticable.

A waiver is not a convenience decision made solely by the researcher. It generally requires justification and review under the applicable rules. Researchers should consider whether the study has important value, involves minimal or proportionate risk, could not practicably be completed otherwise, and includes alternative protections.

Privacy, Confidentiality, Anonymity and Data Protection

These concepts should be described accurately.

Term	Meaning	Example
Privacy	A person’s control over access to themselves, their space or information	Conducting an interview where others cannot overhear
Confidentiality	The researcher knows or can access identity but protects it from unauthorised disclosure	Storing names separately from interview transcripts
Anonymity	Identity is not collected or cannot reasonably be linked to the data	A survey that collects no names, emails, IP addresses or unique combinations
Pseudonymisation	Identifiers are replaced by codes, but reconnection remains possible using separate information	Replacing a name with participant code P014
Data protection	Legal and organisational requirements governing personal-data processing	Defining a lawful basis, access controls and a retention schedule

Anonymity should not be promised casually

A dataset may appear anonymous but remain identifiable through:

• IP addresses
• Device information
• Exact job titles
• Rare diagnoses
• Detailed locations
• Voice recordings
• Faces in images
• Unique life histories
• Combinations of demographic variables
• Searchable quotations

In qualitative research, a distinctive quotation may be discoverable through an internet search. Researchers may need to paraphrase, mask details or obtain explicit permission for attributable quotations.

Confidentiality has limits

Researchers should disclose foreseeable limits before consent. These may include legal obligations, safeguarding procedures, serious threats of harm, court orders, audit access or authorised monitoring.

Do not promise “complete confidentiality” when exceptions are possible.

Data minimisation

Collect only the information reasonably necessary to answer the research question.

For example, a study comparing teaching methods may need participants’ year of study but not their full date of birth, home address, passport number or precise location.

Data minimisation reduces:

• Harm if a breach occurs
• Re-identification risk
• Storage and security burdens
• Unnecessary intrusion
• Legal and administrative complexity

Assessing Risks, Harms and Benefits

Ethical risk assessment considers both the probability and seriousness of possible harm, together with the value of the research and the safeguards available.

Types of research harm

Physical harm

Examples include injury, pain, fatigue, adverse reactions or unsafe field conditions.

Psychological harm

Examples include anxiety, shame, distress, traumatic recall, loss of self-esteem or fear of judgement.

Social and reputational harm

Disclosure could damage relationships, employment, education, professional standing or community reputation.

Economic harm

Participation or disclosure may result in lost income, travel expense, reduced benefits, fraud risk or financial discrimination.

Legal harm

Research data may reveal unlawful behaviour, immigration status, prohibited activity or information subject to legal demand.

Privacy and informational harm

A breach may expose health conditions, beliefs, sexuality, location, political activity, biometric information or other sensitive details.

Cultural and community harm

Research questions, categories or interpretations may disrespect cultural values, reproduce stereotypes or extract knowledge without fair recognition and benefit.

Researcher harm

Researchers may encounter traumatic material, harassment, physical danger, vicarious distress, cybersecurity threats or unsafe lone working.

Risk-reduction strategies

Possible safeguards include:

• Using less invasive methods
• Collecting fewer identifiers
• Allowing questions to be skipped
• Limiting session length
• Providing breaks
• Using trained interviewers
• Preparing a distress protocol
• Offering support information
• Encrypting devices and files
• Separating identity keys
• Restricting staff access
• Avoiding identifiable quotations
• Conducting researcher safety checks
• Monitoring adverse events
• Stopping procedures when risk becomes unacceptable

Vulnerable Participants and Unequal Relationships

A person should not automatically be labelled vulnerable in every context. Vulnerability can arise from the interaction between an individual, the research environment and the decision being requested.

Additional safeguards may be needed for:

• Children and adolescents
• Adults with impaired decision-making capacity
• People with serious illness
• Prisoners and detained populations
• Refugees and displaced people
• People experiencing poverty or dependency
• Employees recruited by managers
• Students recruited by instructors
• Survivors of violence
• Stigmatised or criminalised populations
• People with limited language or digital access

Children and young people

Depending on age, maturity, risk and local rules, research may require:

• Permission from a parent or guardian
• The child’s affirmative agreement or assent
• Age-appropriate information
• A procedure for responding to dissent
• Safeguarding arrangements
• Additional privacy protections

Parental permission does not justify ignoring a child’s clear unwillingness unless a specific, ethically justified framework permits otherwise.

Capacity and supported decision-making

Capacity should be assessed for the particular decision rather than assumed from a diagnosis. Researchers should consider whether simplified information, communication support, extra time or a trusted supporter can help the person decide.

Cultural and community engagement

Individual consent remains important, but some studies also require respectful engagement with community leaders, patient groups, Indigenous governance bodies or local organisations.

Community permission should not be used to replace an individual’s voluntary choice where individual consent is required.

Ethical Considerations by Research Method

Surveys and questionnaires

Important issues include:

• Whether questions are sensitive or intrusive
• Whether survey software collects IP addresses or metadata
• How anonymity is described
• Whether incentives are proportionate
• Whether participants can skip questions
• Whether the recruitment list reveals participation
• How incomplete or withdrawn responses will be treated

An “anonymous” online survey is not anonymous if the platform records identifiable login information, emails or IP addresses accessible to the researcher.

Interviews and focus groups

Researchers should consider:

• Emotional distress
• Private interview settings
• Audio or video consent
• Secure transcription
• Identifiable stories
• Third-party information
• Limits of focus-group confidentiality
• Whether quotations can reveal identity

A researcher can promise to protect focus-group records but cannot guarantee that other participants will keep everything confidential. This limitation should be explained.

Experiments

Ethical issues may include:

• Physical or psychological risk
• Withholding normal treatment
• Random allocation
• Placebo use
• Deception
• Debriefing
• Stopping rules
• Adverse-event monitoring
• Access to effective interventions after the study

Observation and ethnography

Key questions include:

• Is the setting genuinely public?
• Would people reasonably expect to be observed for research?
• Could disclosure alter behaviour or create danger?
• Is covert observation necessary?
• How will bystanders be handled?
• Could detailed fieldnotes reveal identity?
• Does the researcher hold a dual role in the community?

Deception research

Deception may involve withholding the full purpose or providing incomplete information so that behaviour is not distorted.

It requires strong justification. Researchers should examine whether:

• The question has sufficient value.
• A non-deceptive alternative is unavailable.
• The deception conceals significant risk.
• Participants are debriefed appropriately.
• Participants can ask questions.
• The debrief itself could cause harm.
• Data can be withdrawn after disclosure where applicable.

Qualitative research

Ethics in qualitative research is relational and often develops during fieldwork.

Researchers may need to decide:

• How to respond when participants disclose unexpected harm
• Whether participants should review quotations
• How researcher identity affects the interaction
• How to represent contradictory accounts
• Whether removing context misrepresents participants
• How reflexivity and power should be documented

Secondary-data research

Using existing data does not automatically remove ethical responsibilities.

Researchers should assess:

• The original consent
• The proposed new purpose
• Identifiability
• Sensitivity
• Access conditions
• Data-sharing agreements
• Group harms
• Security requirements
• Whether additional review is required

Systematic reviews and literature research

These projects may not involve direct participant contact, but integrity issues remain. Researchers should use transparent inclusion criteria, avoid plagiarism, represent sources accurately, disclose conflicts and avoid selectively reporting only convenient evidence.

Animal research

Animal research should follow applicable laws and institutional review requirements. A common framework is the 3Rs:

• Replacement: Use non-animal alternatives where possible.
• Reduction: Use no more animals than scientifically necessary.
• Refinement: Reduce pain, distress and adverse effects.

Researchers must also justify scientific value, species selection, procedures, housing, humane endpoints and staff competence.

Digital, Social-Media and Internet Research

Data being technically accessible does not automatically make every research use ethically acceptable. Researchers should evaluate users’ expectations, sensitivity, platform context, identifiability, potential harm and whether publication would move information into a more visible or permanent setting.

Public versus private is not a complete test

A post may be publicly searchable but written for a small support community. Quoting it in a paper could make it discoverable outside its original context.

Researchers should consider:

• Platform access restrictions
• Group size
• Sensitivity of the topic
• User expectations
• Whether usernames are persistent
• Searchability of quotations
• Vulnerability of the community
• Whether consent is practicable
• Whether paraphrasing reduces risk
• Platform terms and legal requirements

Web scraping

Before scraping data, researchers should assess:

• Whether collection is permitted
• Whether personal or sensitive information will be captured
• Whether deleted posts will remain in the dataset
• Whether bots burden the service
• Whether data can be re-identified
• Whether users could reasonably expect this reuse
• How data and code will be shared
• Whether an ethics body should review the project

Digital security

Research plans should address:

• Encryption
• Multi-factor authentication
• Access permissions
• Device security
• Secure transfer
• Backups
• Data-location restrictions
• Vendor contracts
• Incident response
• Secure deletion

Ethical Use of Artificial Intelligence in Research

AI tools may support literature discovery, coding, transcription, translation, programming, drafting or data analysis. Their use introduces additional ethical and integrity risks.

Do not upload confidential data without authorisation

Public or consumer AI systems may process information outside the researcher’s approved environment. Researchers should not upload:

• Identifiable participant data
• Confidential transcripts
• Unpublished manuscripts received for review
• Proprietary datasets
• Sensitive fieldnotes
• Restricted health or administrative records

Use an institutionally approved system and verify its privacy, retention and contractual conditions.

Maintain human responsibility

AI cannot take ethical or scholarly responsibility. Researchers remain accountable for:

• Accuracy
• Citations
• Interpretation
• Bias
• Originality
• Confidentiality
• Compliance
• Final conclusions

AI-generated text, references, code and classifications must be checked rather than accepted because they appear plausible.

Disclose material AI use

Follow institutional, funder and journal policies. A disclosure may identify:

• The tool and version
• The task it performed
• The stage of research
• Whether confidential data were entered
• How outputs were verified
• What human oversight was applied

AI tools should not be listed as authors because they cannot approve the final work, disclose conflicts or accept responsibility.

Assess bias and unequal performance

An AI system may perform differently across languages, demographic groups, dialects or cultural settings. Validation should reflect the intended population rather than relying only on general vendor claims.

Avoid automated ethical outsourcing

An AI tool may help identify questions, but it should not decide whether a participant has capacity, whether risk is acceptable or whether a protocol is ethical. These decisions require accountable human judgement and, where required, independent review.

Research Integrity and Publication Ethics

Ethical research includes responsible analysis and communication.

Fabrication, falsification and plagiarism

• Fabrication means inventing data or results.
• Falsification means manipulating materials, processes or results so that the research record is misleading.
• Plagiarism means using another person’s words, ideas, methods or results without appropriate credit.

Honest mistakes and legitimate differences of interpretation are not automatically misconduct, but discovered errors should be corrected transparently.

Selective reporting

Researchers should not hide inconvenient outcomes, change hypotheses after seeing results without disclosure, omit relevant exclusions or present exploratory analyses as if they were planned in advance.

Preregistration, registered reports, analysis plans and transparent supplementary materials can reduce ambiguity, although none substitutes for good judgement.

Authorship and contribution

Authorship should reflect substantial intellectual contribution and responsibility. Gift authorship, ghost authorship and excluding deserving contributors are unethical.

Agree early on:

• Expected roles
• Authorship criteria
• Author order
• Data ownership
• Corresponding-author duties
• How disagreements will be resolved

A contributor statement can clarify who designed the study, collected data, performed analysis, wrote the manuscript and supervised the work.

Conflicts of interest

Financial, personal, institutional or professional relationships may influence—or appear to influence—research decisions.

Conflicts do not always prohibit research, but they should be disclosed and managed. Measures may include independent analysis, oversight, restricted decision-making or transparent reporting.

Ethical Approval in the United States, United Kingdom and International Research

United States

US human-participant research may be reviewed by an Institutional Review Board. The applicable requirements depend on factors such as funding, institution, research setting, data identifiability and whether the activity meets relevant regulatory definitions.

The Belmont Report provides three foundational principles:

1. Respect for persons
2. Beneficence
3. Justice

Researchers should not independently assume that a project is exempt merely because it is low risk, anonymous or conducted for a student dissertation. Institutional procedures should be followed.

United Kingdom

UK universities commonly use departmental or institutional Research Ethics Committees. Certain health and social-care studies require review through Health Research Authority or related systems.

Researchers must also consider current data-protection requirements. Ethical consent to participate should not be confused with the legal basis for processing personal data.

Because UK data law and research-governance procedures can change, researchers should consult current institutional, HRA and Information Commissioner’s Office guidance rather than relying on an old dissertation template.

International and multi-country research

International studies may need review from:

• The sponsoring institution
• Local institutions
• National authorities
• Health or regulatory bodies
• Community or Indigenous governance structures
• Each country in which research activities occur

Approval in one country does not automatically replace local review.

Researchers should avoid “ethics dumping,” in which activities considered unacceptable in a well-resourced setting are moved to a location with weaker oversight. Partnerships should address local relevance, fair credit, capacity building, benefit sharing and access to findings.

For medical research involving human participants, the current 2024 Declaration of Helsinki should be used rather than superseded versions.

Worked Examples of Ethical Considerations

Example 1: Anonymous student survey

Project: A lecturer wants to survey their own students about stress and teaching quality.

Main ethical issues:

• Students may fear that refusal will affect grades.
• Login-based software may identify respondents.
• Small demographic groups may be recognisable.
• Questions about mental health may cause discomfort.

Possible safeguards:

• Use a neutral person to recruit participants.
• Do not provide the lecturer with identifiable response records.
• Make participation optional and provide an equivalent alternative.
• Avoid collecting unnecessary identifiers.
• Allow questions to be skipped.
• Report only aggregated results.
• Provide relevant support contacts.

Example 2: Interviews with survivors of violence

Project: A researcher conducts interviews about experiences of domestic abuse.

Main ethical issues:

• Emotional distress
• Risk if an abusive person discovers participation
• Safe contact procedures
• Limits of confidentiality
• Researcher vicarious trauma
• Identifiable life histories

Possible safeguards:

• Agree on safe communication methods.
• Avoid revealing study details in messages.
• Use trained interviewers and a distress protocol.
• Allow pauses, skipped questions and withdrawal.
• Store contact information separately.
• Remove distinctive details from quotations.
• Prepare researcher supervision and wellbeing support.

Example 3: Scraping public social-media posts

Project: A team collects posts about a rare health condition.

Main ethical issues:

• Users may not expect research reuse.
• Exact quotations may reveal identities.
• The condition may be highly sensitive.
• Deleted content may remain in the dataset.
• Group-level conclusions may stigmatise the community.

Possible safeguards:

• Seek ethics review despite public accessibility.
• Assess platform and community expectations.
• Collect only necessary fields.
• Remove usernames and links.
• Paraphrase quotations where scientifically acceptable.
• Avoid exposing small subgroups.
• Consider community consultation.
• Restrict access to the raw dataset.

Example 4: AI-assisted interview coding

Project: A postgraduate researcher wants an AI chatbot to identify themes in interview transcripts.

Main ethical issues:

• Transcripts contain personal and confidential information.
• The system may retain or reuse inputs.
• Coding may contain cultural or linguistic bias.
• AI-generated themes may be inaccurate.
• Participants may not have consented to third-party AI processing.

Possible safeguards:

• Do not upload raw transcripts to an unapproved public system.
• Use an approved secure environment or de-identified extracts.
• Check consent and data-processing arrangements.
• Validate coding through human review.
• Record prompts, settings and verification procedures.
• Disclose material AI assistance in the methodology.

How to Write Ethical Considerations in a Research Proposal

An ethical-considerations section should connect general principles to the actual study. Avoid writing only that “all ethical rules will be followed.”

Include:

1. The relevant review or approval route
2. Recruitment and voluntariness
3. Consent procedures
4. Risks and mitigation
5. Privacy and confidentiality
6. Data collection and storage
7. Withdrawal arrangements
8. Vulnerable-group safeguards
9. Researcher safety
10. Reporting, sharing and retention
11. Conflicts of interest
12. Any method-specific issues

Example ethical-considerations paragraph

This study will recruit adult university students through a general departmental announcement rather than direct invitations from their instructors. Participation will be voluntary and will not affect grades, services or academic standing. Before beginning the survey, participants will receive plain-language information about the study’s purpose, procedures, possible discomfort and data use. The survey will not request names, student numbers or exact dates of birth. Participants may skip any question or exit before submitting their responses. Data will be stored in an access-controlled institutional system and reported only in aggregate form. Small demographic categories will be combined or suppressed where disclosure may create identification risk. Data collection will begin only after the required institutional ethics decision has been received.

This example must be adapted to the real protocol. It should not be copied into a study that uses different recruitment, data or risk procedures.

Ethical Considerations Checklist

Before data collection, confirm that:

• The study has a clear social or scientific value.
• The design can answer the research question.
• Existing evidence has been reviewed.
• Participants are selected fairly.
• Recruitment avoids coercion and inappropriate pressure.
• Information is clear and accessible.
• Consent procedures fit the population and method.
• Withdrawal arrangements are realistic.
• Physical, psychological, social, legal and privacy risks have been assessed.
• Vulnerable groups have appropriate safeguards.
• Only necessary data will be collected.
• Confidentiality and anonymity are described accurately.
• Storage, access, retention and deletion are documented.
• Digital platforms and third-party tools have been assessed.
• AI use complies with institutional and publication policies.
• Researcher safety and wellbeing are addressed.
• Conflicts of interest are disclosed and managed.
• The required ethics approval or determination has been obtained.
• A procedure exists for incidents and amendments.
• Findings will be reported honestly and responsibly.

Benefits and Limitations of Ethical Review

Benefits

Independent ethics review can:

• Identify risks overlooked by the research team
• Improve consent information
• Strengthen recruitment fairness
• Require clearer data protections
• Help manage conflicts of interest
• Increase accountability
• Protect participants and researchers
• Improve the study’s acceptability

Limitations

Ethics review does not:

• Guarantee that no harm will occur
• Replace legal advice
• Correct a scientifically weak design automatically
• Predict every fieldwork dilemma
• Remove researcher responsibility
• Make approved procedures suitable after the protocol changes
• Resolve every cultural or community concern
• Prevent misconduct by itself

Ethical practice requires continuing judgement, documentation, consultation and willingness to change or stop activities when necessary.

Common Ethical Mistakes

Treating ethics as a final formality

Ethical planning should shape the question, sampling, method and data plan from the beginning.

Beginning data collection before approval

Retrospective approval may not be available. Data gathered prematurely may be unusable.

Promising complete anonymity

Researchers should first examine metadata, indirect identifiers, quotations, recordings and platform settings.

Assuming a signed form proves consent

Consent is invalid when participants do not understand the information or feel unable to refuse.

Collecting unnecessary personal information

“Useful later” is not a sufficient justification for intrusive data collection.

Ignoring unequal power

Students, employees, patients and service users may experience invitations as pressure.

Assuming public online data is ethically unrestricted

Context, sensitivity, expectations and amplification risks still matter.

Uploading data to unapproved AI services

Removing names may not remove all identifying or confidential information.

Reporting only favourable results

Selective reporting distorts evidence and wastes participant contribution.

Confusing approval with continuing compliance

Material changes, new risks and unexpected events may require consultation or formal amendment.

Conclusion

Ethical considerations in research are practical responsibilities that influence every stage of a study. Researchers must justify the value and validity of their work, respect autonomy, obtain meaningful consent where required, reduce harm, recruit fairly, protect information and report findings honestly.

The correct safeguards depend on the method, population, setting, data and jurisdiction. Ethical approval is important, but responsible research also requires ongoing reflection, transparent decision-making and action when circumstances change.