University research ethics review in Singapore protects people whose information, experience or biological material might be used in a student thesis, dissertation or academic study. Research involving surveys, interviews, experiments or existing identifiable personal data can be educationally valuable, but it can also raise questions about informed consent, privacy, undue pressure and safety. A sincere academic motive does not automatically give researchers permission to collect or publish another person’s information.
For students asking when NUS IRB approval is needed, how NTU IRB works, what an ethics exemption means and whether a university research questionnaire needs consent, the official guidance makes several distinctions. NUS IRB’s FAQ says human-subject research must receive the appropriate ethics approval, exemption or formal determination. NTU’s student-research guidance distinguishes postgraduate research-degree applications via supervisors from qualifying coursework FYP reviews through School Ethics Review Committees. Neither institution permits students to decide for themselves that a real research project is exempt simply because it appears low risk.
This guide explains the review categories, Singapore-specific 2026 training requirements, proposals and participant information, consent, existing data, minors, changes to approved methods and responsible reporting. It is an educational planning guide, not a substitute for formal ethics determination or legal advice. Always obtain direction from the university’s authorised reviewing body before starting work involving human participants.
The first decision: is this human-subject research?
A researcher should identify whether the activity is a systematic investigation intended to contribute to generalisable knowledge, and whether it involves collecting information through interaction with people or accessing identifiable private information. NUS and NTU publish these concepts to distinguish research from some routine activities.
A simple question list does not settle classification on its own. The intended use, audience, design and nature of data matter. If unsure, consult the institutional ethics office or authorised departmental committee before approaching potential participants.
A research thesis may involve people in many ways
Human participants are not limited to clinical trials. Interviews, student surveys, focus groups, behavioural tests, observations, identifiable online information and certain secondary datasets can involve human-subject research under institutional rules.
An engineering or computing project can also raise ethics questions when it tests a device or interface on identifiable users. Technical students should not assume that IRB review concerns only healthcare. The institution assesses the real activity and its risks.
NUS IRB versus departmental review
NUS has an Institutional Review Board and Departmental Ethics Review Committees (DERCs). Its DERC guidance allows certain qualifying minimal-risk student social, behavioural and educational research to be reviewed at departmental level under specified conditions.
Other studies, including those involving particular funding, human biomedical research or other excluded categories, must go through NUS IRB. A student’s school affiliation alone does not choose the reviewing body. The actual protocol and institutional screening determine the correct route.
The two NUS systems have different purposes
NUS’s DERC Review and Approval Management application, called DREAM, handles DERC submissions; NUS IRB applications use iRIMS-IRB. The NUS DERC page explicitly distinguishes these systems.
A student should obtain guidance from the departmental coordinator before submitting. Sending a protocol to the wrong administrative system can delay study without generating any valid approval. The correct application path must be established before recruitment or data collection.
NUS 2026 CITI training update
From 1 March 2026, NUS requires CITI Social and Behavioral Research Basic/Refresher certification for its staff and students submitting relevant social, behavioural and educational research to IRB or DERC review. The IRB training page describes the change and certificate submission.
This is not the same as claiming every undergraduate course requires the same CITI certificate. The obligation applies to the named research-review processes. A student should complete the required programme honestly and attach the official evidence as instructed.
NTU IRB has a student-research framework
NTU states that postgraduate students conducting human research for a research master’s or PhD submit through their supervisor to NTU IRB. Under the published system, eligible student researchers may help prepare a submission as Student PI while the institutional responsibilities remain.
By contrast, qualifying undergraduate FYP and postgraduate coursework projects may be reviewed through a School Ethics Review Committee (SERC), with other studies escalated to IRB. See the NTU student-research page for current categories.
NTU School committees have limits
NTU explains that an SERC may review certain Social, Behavioural and Educational Research proposals in the Exempt or Expedited categories where other required conditions are met. It cannot simply approve every higher-risk study or a project covered by the Human Biomedical Research Act.
A school committee can refer an application for further IRB review. Students should not pressure a lecturer to sign a form outside the approved scope. Research safety and participant rights should determine the correct authority, not the quickest apparent route.
NUS and NTU can have different administrative steps
A NUS student may work through DREAM or iRIMS-IRB, while NTU’s current policy and procedure page describes research ethics applications through its NORA system with required endorsement. Procedures are not interchangeable.
A tutorial on a different university’s forms is not authoritative for another institution. Record the university, school, programme, responsible PI and current portal beside the protocol. Use official logins and do not upload research identifiers to unofficial websites.
Exempt does not mean self-approved
NTU’s Exempt Review guidance says projects meeting specified low-risk categories must still apply for formal Exempt-status determination rather than decide privately that IRB is unnecessary.
The distinction is important: exempt from full review is not necessarily exempt from submitting a protocol. Approval must precede the relevant research activity, and changes can require fresh review. Read the exact category and reviewer decision.
What minimal risk means
NTU describes minimal risk in terms of the probability and magnitude of harm or discomfort not exceeding those ordinarily encountered in daily life or routine examinations. The classification is based on the anticipated study procedures and participants, not on how harmless the researcher feels the topic is.
A seemingly simple question might still expose sensitive personal information or create pressure if asked by someone with authority over participants. Risk includes social, reputational, privacy and other harms, not only physical injury.
Expedited review is not an ethical shortcut
An Expedited review may be suitable for defined minimal-risk studies that do not qualify for Exempt classification. The study still undergoes formal ethics assessment, with review arrangements set by the institutional rules.
For example, recording identifiable interviews can create confidentiality risks even when the discussion is not physically dangerous. Researchers should provide accurate procedures and safeguards, rather than describe every project as anonymous to obtain a faster route.
Full Board review for greater risk
Certain studies with more than minimal risk, sensitive populations, invasive procedures or other complex safeguards require Full Board review. NTU gives examples and identifies when social, behavioural and human biomedical research needs that level of consideration.
A student should not redesign a risky study solely to avoid scrutiny while leaving the actual risks unchanged. Research methods need to be proportionate, approved and feasible. A slower, proper review is preferable to gathering unusable or unauthorised evidence.
Review Not Required is another classification
Some activities fall outside the institution’s formal definition of research, or involve fully de-identified information that no longer constitutes human-subject research under particular rules. NTU has a Review Not Required category for certain cases.
This classification should be established through the appropriate process. The student should not assume that every publicly visible social-media post, school record or personal dataset is unrestricted simply because it can technically be accessed.
A course feedback survey may differ from research
NTU’s student guidance explains that routine post-event feedback about an orientation programme can be non-academic service evaluation that does not require an IRB submission. Research designed to produce generalisable conclusions or publication can trigger different review requirements.
The purpose, questions and intended use must be considered honestly. Do not gather personal information under a ‘feedback only’ label and later treat it as approved thesis research. Ask the relevant adviser before changing the activity’s purpose.
Secondary data can still involve human subjects
A dataset assembled by someone else may include names, identifiers, unique combinations or linkable records. NTU’s Exempt Category 4 guidance explains when secondary research needs Exempt review and when fully irreversibly de-identified data may instead fit Review Not Required.
Existing data are not automatically harmless or freely reusable. Permission, original consent conditions, data-sharing agreements and the risk of re-identification all matter. Combining two datasets may reveal an identity that seemed hidden in either source alone.
Anonymous is not the same as confidential
Anonymous data cannot reasonably be linked back to an individual in the relevant context, while confidential identifiable data remain linked but protected through access and use restrictions. Researchers should not casually promise anonymity when they retain names, contact details or identifiers.
A signed consent record, unique demographic combination or audio voice can complicate claims of anonymity. Explain the actual design in participant information and data management plans, using the approved institutional definitions.
Consent is a process of informed choice
Participants should understand the purpose, procedures, risks, time, confidentiality and voluntary nature of participation under the approved consent procedure. They must not be misled about important aspects of the research without legitimate ethics approval for an authorised alternative.
Consent is more than collecting a signature. A participant needs meaningful information and space to ask questions. Where power relationships or sensitive data are involved, special safeguards may be required to protect genuine voluntary choice.
A lecturer recruiting their own students needs care
An academic may be able to invite participation in research, but students could feel pressured if the recruiter controls grades or learning opportunities. Ethics review can require measures to reduce coercion or undue influence.
For instance, an approved recruitment process might separate the consent decision from grading and avoid making participation a condition of assessment. The precise safeguards must come from the IRB or SERC rather than be improvised after students have responded.
Research involving minors needs additional safeguards
Children and other potentially vulnerable participants may need parental permission, appropriate assent, institutional access approval and special handling of identifiable information. NTU’s waiver guidance sets conditions for any proposed departure from standard parental-consent arrangements.
A secondary school student’s involvement in a university survey does not erase age-specific protections. Researchers should obtain the correct assessment before recruitment. Permission from a school alone may be insufficient when university research approval is also required.
Human biomedical research has another legal layer
Singapore’s Human Biomedical Research Act regulates defined human biomedical research and related human biological materials. University IRB guidance explains that some projects fall under those requirements and cannot be reviewed through committees limited to social and educational minimal-risk studies.
A classroom science label does not determine the law. Work involving human tissue, clinical interventions or regulated medical products deserves specialised institutional screening. Never perform invasive procedures or handle sensitive biological material without proper training and authorisation.
The protocol describes exactly what will happen
An ethics application normally includes the question, purpose, recruitment, participant population, procedures, consent method, expected risks, safeguards, data handling and research team responsibilities. Reviewers need enough detail to judge the actual study rather than a vague promise that care will be taken.
Prepare realistic questionnaires or interview guides when required. If the procedure changes, the approval may no longer cover what the researcher intends. A carefully specified protocol helps both participants and investigators understand the conditions.
Changes to an approved protocol may need review
An investigator might want to add participants, change survey questions, introduce a new dataset or move from a non-identifiable analysis to identifiable records. NTU’s Exempt Review page says amendments must be submitted for reassessment before the changed work begins.
Do not treat the old approval number as permission for any later variation. Ask the reviewing body what kind of modification requires formal amendment and keep the resulting authorisation. Research quality and compliance both depend on recording the method actually used.
Retrospective ethics approval is not offered at NTU
NTU’s journal and publisher guidance says the university cannot retrospectively approve human research already conducted, because ethical corrections cannot be applied to earlier participant interactions.
This can be especially important if students conduct a survey first and only later decide to publish it. Do not fabricate an approval number or claim the study was exempt without a valid determination. Proper review belongs before the research activities that require it.
Ethics approval and publication are connected
Many journals ask authors for evidence of ethics review, consent and appropriate data handling. A compelling result does not erase the need for prior approval where required. Reviewers may reject a paper that used unauthorised participant research.
Discuss future dissemination before data collection, including whether consent permits publication of quotes or images. A thesis-only intention today does not automatically authorise broader sharing tomorrow. Intellectual ambition must remain compatible with participant rights.
Participant information should be understandable
A participant information sheet should explain the study in language appropriate to the people being invited. It should make the time commitment, kinds of questions, potential discomfort, contact routes and data uses clear without relying on technical terms the reader has not been taught.
A university may provide an approved template, and the reviewing committee can require specific wording. The researcher should not replace essential disclosure with promises that everything is completely safe. Participants need a realistic account of both what will happen and what protection is offered.
Withdrawal conditions must be stated accurately
Voluntary consent normally allows participants to make a meaningful decision about participation and withdrawal under the approved procedure. However, once responses have been irreversibly anonymised or combined, removing an individual’s contribution may not always be feasible.
Describe the real withdrawal process and limits rather than promise unlimited deletion in all circumstances. Researchers should follow the consent terms approved for the specific protocol. A participant who asks a question about withdrawal deserves clear and respectful guidance.
Recruitment materials are part of the protocol
A flyer, online invitation or message used to recruit participants can influence who volunteers and whether they understand the research. The ethics reviewer may need to see its wording and proposed distribution method.
Avoid using pressure, misleading rewards or claims that participation will improve a student’s grade. If an incentive is offered under an approved protocol, describe it accurately. A fair invitation should make declining as straightforward as consenting.
Interview recordings require security planning
Audio and video can contain identifiable voices, faces and sensitive remarks even without a participant’s name. A research plan should identify whether recording is essential, how permission is obtained, where files are stored and who may access them.
Do not upload unapproved recordings to a public transcription or AI service simply to save time. Use institutionally authorised tools and data locations, and follow the approved retention and deletion process. Researchers must not promise anonymity when they retain directly identifiable recordings.
Survey platforms can collect identifiers indirectly
An online survey may capture email addresses, account details, timestamps, device or network information depending on its settings and the provider. An investigator should verify what is collected before describing a questionnaire as anonymous.
A participant’s answer could also reveal an identity through a rare combination of details. The team should assess these risks in the protocol and implement approved safeguards. Technical convenience does not remove the responsibility to understand the data generated.
Consent for publication of photographs or quotations
A person may agree to participate in a study without agreeing to have their face, full name or identifiable quotation published. Consent for data collection and permission for public dissemination should be clear within the authorised research design.
Check whether a report can use an appropriately anonymised quotation or requires more explicit permission. An attractive photograph should not be published as evidence simply because the participant attended a session. Protecting dignity and expectations is part of valid research communication.
A public dataset is not always unrestricted
A dataset may appear downloadable or visible online but have licensing terms, participant restrictions, copyright conditions or risks associated with re-identification. NTU’s secondary-data guidance distinguishes fully anonymised records from personally identifiable existing data.
Researchers should examine lawful access, permitted use and what the university’s ethics office determines is necessary. A technical ability to scrape or join records is not a grant of research permission. Data provenance and safeguards must accompany methodological choices.
Data protection is not merely password selection
Sensitive research data can require institutionally approved storage, access control, encryption, retention schedules and incident response. Passwords alone cannot resolve uncertainty about ownership or whether a third-party service has permission to process participant information.
Consult the university’s research data policies and official security team when needed. Separate contact lists from analytical datasets where the approved protocol requires it. A secure plan should explain who needs access and why, not place every file in a personal folder by habit.
Consent waivers belong to the reviewer
NTU’s consent waiver guidance describes defined conditions under which the IRB may approve a change to ordinary consent requirements. Researchers must justify their particular protocol against the applicable criteria.
This is not permission to skip consent because asking participants would be inconvenient. The reviewing body decides whether the proposed alternative sufficiently protects rights and welfare. A student should never self-certify a waiver before collecting data.
How to identify avoidable pressure
Participants may feel obliged to comply when approached by a teacher, employer, lecturer or authority figure. Even a neutral-sounding request can become coercive when the person worries that refusing could affect grades or work.
A research design might reduce pressure through independent recruitment, clear separation from assessment and anonymous responses, subject to ethics approval. The right safeguards depend on the actual setting and participants. Protecting voluntary choice is a research-design requirement, not an afterthought.
An educational research example
Imagine a fictional student wants to test whether two types of worksheet explanations affect comprehension. If the study is intended to contribute to broader educational knowledge and involves real classmates, it should be submitted for the appropriate ethics determination before recruitment.
The investigator must consider whether participants are adults, who assigns their grades, what information is gathered and whether any intervention changes normal learning. A routine teaching improvement exercise may be classified differently, but the student does not decide that classification alone.
An interview-based social research example
A fictional master’s candidate proposes interviewing working adults about career decisions. The topic might appear low risk, yet personal experiences, employment history and recorded voices can make the data identifiable or sensitive.
A credible proposal describes who will be invited, how informed consent works and how recordings or quotations will be handled. An approved design may permit the research. The ethical quality lies in realistic protection of participants, not merely in presenting the study as friendly conversation.
An app usability example
An undergraduate builds a demonstration website and asks volunteers to perform specified tasks while screen recording. If the activity is a research investigation of human behaviour, participant interaction and recordings can trigger ethics review.
The researcher should examine what screen information, account names or personal identifiers may be captured and whether a fictional test environment can reduce risk. Real-world user data should not be collected before the responsible university committee has determined the applicable review route.
A secondary-dataset example
A student receives records of service use from an organisation. Names are removed, but rare combinations of age, location and activity could identify some people. Calling the table anonymous would be unwarranted without examining re-identification risk.
The team should ask the responsible data owner and ethics office about access, existing consent, permitted analysis and safeguards. Only a formally acceptable arrangement should be used. A strong academic result obtained from unauthorised data remains a serious problem.
Risk reduction can improve the science
A reviewer may recommend reducing unnecessary personal questions, collecting less identifiable information or improving the consent process. These changes can also make data cleaner and the study’s question more focused.
Ethics review is not simply an obstacle to producing a thesis. It can expose unclear methods, unrealistic recruitment and avoidable participant burden. The researcher should treat requested revisions as opportunities to make the project more defensible.
How to respond to reviewer feedback
A review may ask for clarification about recruitment, compensation, methods or confidentiality. Respond to the actual question with a revised protocol and explanation rather than making unrelated cosmetic edits.
Keep track of which version is approved and when a change becomes effective. If the student disagrees with a request, seek clarification through the proper institutional channel. The aim is a clear, documented decision about the authorised research, not an argument over paperwork.
Plan review time before the thesis calendar
Ethics classification and review can take time, and some submissions require additional training, committee consideration or multiple revisions. A project that starts the approval process only when its data analysis is due risks missing essential academic milestones.
Build the review step into the proposal before participant recruitment. Work on legitimate literature and methods preparation while awaiting determination. Do not begin restricted research merely because the student’s final-year report deadline feels urgent.
Unexpected problems may require reporting
A participant may experience distress or a data-access error may expose information unexpectedly. Institutions publish procedures for incident reporting and deviations from approved protocols, with responsibilities varying by study category.
Follow the actual reporting and safety channels promptly rather than conceal a problem to protect the publication timeline. The right response can include suspending affected activities, safeguarding people and obtaining formal advice. Reporting factual concerns is part of responsible investigation.
Ethics review is not proof of scientific correctness
A formal ethics approval addresses participant protection and applicable governance. It does not guarantee that the research question is important, that a statistical test is valid or that every interpretation is justified.
Students still need academic supervision and careful methods. A thesis can be ethically authorised yet scientifically weak, or methodologically clever but ethically impermissible. Strong research requires both dimensions and transparent reporting of limitations.
Publication approval and research ethics are distinct
After an approved study, a university, employer or collaborator might have separate permission requirements for releasing data, images or manuscript material. Ethics approval alone does not transfer intellectual property or permit unrestricted publication.
Check the approved consent, data agreements and publisher requirements. A study can support a thesis while particular records remain confidential. Researchers should disclose methods without revealing protected details, and seek institutional guidance before sharing research materials publicly.
A six-step ethics preparation workflow
First define the academic question. Second identify whether real people or identifiable data are involved. Third ask the authorised university office about classification. Fourth complete necessary training and prepare the protocol and consent material. Fifth obtain the formal determination before the covered activity begins. Sixth monitor amendments and reporting obligations.
This is a planning sequence, not an institutional application form. Different programmes and risk types can require additional steps or approvals. Never use a generic internet checklist to override instructions from the actual university’s IRB or School committee.
Frequently asked questions
Does every student survey need IRB approval? Not every routine feedback exercise is research, but human-subject research needs the appropriate institutional determination. Can low-risk research skip an application? Do not assume so; an Exempt determination may still require submission.
Can approval be obtained after interviewing participants? NTU does not offer retrospective IRB approval. Is all publicly available data unrestricted? No; identification and use conditions matter. Do NUS and NTU use the same portal? No. Check the specific institution’s process.
Official sources and eduKate connections
Read NUS IRB FAQs, NUS DERC, NUS CITI Training 2026, NTU Student Research Review and NTU Exempt Review.
Related eduKate reading: Choosing a Research Supervisor, University Thesis and Capstone and Academic Research Writing.
Final thought: research should respect the people behind the data
Good human-subject research asks worthwhile questions while protecting participants’ dignity, safety, privacy and ability to choose. Ethics approval helps establish safeguards before an investigator collects evidence that may affect real people.
Begin with the actual protocol and university review route, be open about uncertainty and follow authorised decisions. The strength of a thesis rests not only on what it discovers but also on how responsibly the knowledge was obtained. Reviewed 8 October 2026; institutional ethics and legal requirements can change.
