Getting ethics approval without losing a term to it

Ask a group of candidates where their timeline first slipped and a striking number will name ethics. Not because the work was unethical, and not because committees are obstructive, but because the application was submitted three weeks before recruitment was due to begin, went back for revision, and caught the next meeting cycle six weeks later. Two months disappear without a single participant being approached.

This is avoidable. Ethics review is a predictable process with a small set of recurring failure modes, and knowing them in advance is most of the battle.

Why it takes longer than you planned

Three structural facts drive the timeline, and none of them are about the quality of your work.

Committees meet on a cycle. Many full committees sit monthly, some less often in vacation periods, and each has a submission deadline a week or two before the meeting. Miss a deadline by a day and you wait for the next cycle. This is the single biggest cause of ethics delay, and it is entirely calendar arithmetic — find your committee's dates before you plan anything else.

Revision is the normal outcome, not the exceptional one. A large proportion of first submissions come back with conditions or requested clarifications. Plan for at least one round. If you get through first time, you have gained weeks; if you planned for first-time approval and did not get it, you have lost them.

Some studies need more than one approval. Research in health services, schools, prisons, or with an external organisation's data typically needs both your university's approval and the host organisation's, sometimes sequentially because the second wants to see the first. That is not one timeline, it is two end to end.

What a committee is actually assessing

It helps to understand the committee's own position. Its members are largely reviewing on top of full workloads, often outside their own discipline, and their responsibility is institutional: they are asked to be satisfied that participants are protected and that the university is not exposed. They cannot ask you clarifying questions in the room. They can only read what you submitted.

That has a direct consequence for how you write. A study can be entirely sound and still be sent back because the documents did not make it legible. The recurring questions are:

• Is consent genuinely informed and genuinely voluntary? Do participants understand what they are agreeing to, in plain language, and can they decline without cost?
• Is the risk proportionate to the knowledge gained? Not "is there risk" — almost all research carries some — but is it justified and minimised.
• What happens to the data? Where stored, how secured, who has access, how long retained, and what happens at the end.
• Is anyone in this study vulnerable, or made vulnerable by their relationship to you? Your own students, your own patients, and your own colleagues all raise this.
• What happens if something goes wrong? If a participant becomes distressed, or discloses harm, what will you actually do?

Assembling the application

Requirements vary, but almost every application is built from the same components. Draft them together rather than sequentially, because the commonest inconsistencies appear between documents rather than within one.

• The protocol or application form — aims, design, recruitment, procedure, analysis. Write it for an intelligent non-specialist.
• The participant information sheet — the document that does the most work and gets the most criticism. Plain language, no jargon, and it must answer "what will actually happen to me, for how long, and what if I change my mind".
• The consent form — discrete, separately-agreed statements rather than one global tick, so a participant can consent to the study but decline, say, audio recording or future data reuse.
• Recruitment materials — every advert, email, and social post, exactly as it will appear.
• Instruments — the actual questionnaire, interview guide, or observation schedule. Not a description of it.
• A data management plan — storage, security, retention, anonymisation, and sharing.
• A risk and distress protocol — what you do if a participant becomes upset or discloses something you cannot ignore.

Read the set as a whole before submitting. If the protocol says interviews last 45 minutes and the information sheet says "about an hour", that single inconsistency can generate a revision round on its own.

The revisions that come back most often

Across institutions the same problems recur, and all of them are cheaper to fix before submission than after.

• The information sheet is written for academics. If it contains "semi-structured", "thematic analysis", or "epistemological", rewrite it. Aim for the register of a good newspaper, not a methods chapter.
• Withdrawal is vague. "You may withdraw at any time" is not enough. Until when, exactly? Once data are anonymised and pooled, withdrawal is usually impossible — say so, and give the cutoff.
• Data storage is described aspirationally. Name the actual system, the actual access controls, and the actual retention period.
• Recruitment carries implicit pressure. Recruiting your own students or patients needs an explicit buffer — someone else making the approach, or a clear statement that participation has no bearing on grades or care.
• Incentives look coercive. A payment large enough to override someone's judgement about risk is a problem; a token is not.
• The instrument does not match the protocol. The protocol promises no sensitive questions, and question 14 asks about mental health history.
• Confidentiality is promised absolutely. You cannot promise to keep a disclosure of serious harm confidential. State the limits honestly.

Amendments, and what counts as one

Approval covers the study you described. Change the study and you generally need an amendment — usually much faster than a fresh application, often reviewed by the chair rather than the full committee, but not instant.

An amendment is typically needed for a new participant group, a new site, a changed instrument, extra data collection, a new member of the research team, or an extension to the retention period. Fixing a typo in the information sheet usually is not. If you are unsure, ask the ethics administrator: the cost of asking is an email, and the cost of not asking is data you may not be permitted to use.

This matters more than it sounds. Data collected outside the terms of your approval can be challenged at examination, and in the worst case cannot be included in the thesis at all. That is a genuinely catastrophic outcome discovered very late.

Putting it on the critical path

Ethics approval is a hard dependency: recruitment cannot begin without it, so everything downstream inherits its delay. Treat it accordingly.

• Work backwards from the committee's deadline, not forwards from when your draft happens to be ready.
• Give your supervisor a full week to read the application before submission. They will catch the inconsistencies you have stopped seeing.
• Use the waiting period productively — pilot your instrument with colleagues, write the methods chapter, build your analysis plan. None of that requires approval.
• If a second approval is needed from a host organisation, start that conversation in parallel, not after.

A candidate who submits a complete, internally consistent, plainly-written application to a known deadline generally clears in one or two cycles. A candidate who writes it in a hurry to hit a date they invented usually does not.