Nobody is evaluating whether your study is interesting
This surprises students who arrive defensive, expecting their research question to be judged. It is not. Scientific merit belongs to your committee. What a review board is checking is narrower and more mechanical: whether participants are protected, whether consent is genuinely informed, whether data is handled responsibly, and whether the documents agree with one another.
That reframe is worth having before you write, because it changes what deserves your effort. A brilliantly argued significance section will not speed you up. A consent form that matches your protocol will.
It also means a return is not a rejection of your work. It is a list of places where the packet did not yet demonstrate one of those four things.
The questions, in the order we ask them
Different boards use different templates, but the reading order is remarkably consistent, because it follows the participant's experience rather than your document structure.
- Who exactly is being recruited, and how will they hear about this?
- What happens to a participant, step by step, and how long does it take?
- What could go wrong for them, and what have you done to reduce it?
- How is consent obtained, from whom, and could someone feel pressured to agree?
- What data is collected, is any of it identifying, and who can open it?
- Where does it live, who has access, and when is it destroyed?
- Does every document in this packet describe the same study?
Contradiction is what actually gets caught
The last question on that list produces more returns than all the others combined, and it is entirely preventable. Your protocol says interviews last thirty minutes; your consent form says up to an hour. Your recruitment email promises responses will be anonymous; your survey asks for a work email address so you can send the incentive.
Neither of those is dishonesty. They are the ordinary consequence of documents written on different days and edited separately. But a board reads them together, in one sitting, and contradictions are the easiest thing in the world to spot.
So before you submit, read the entire packet through as though you had never seen it, in one go, looking only for places where two documents disagree. That hour catches most of what would otherwise cost you a meeting cycle.
Consent is written for the reader, not for you
The most common substantive comment I have written on student protocols concerns reading level. Consent forms drafted by graduate students read like graduate students wrote them, which means participants sign something they have not fully understood, which defeats the entire purpose.
Write it at the level your population actually reads at, which is usually considerably plainer than you expect. Short sentences. No methodological vocabulary. Say what will happen to them, what might go wrong, that they can stop at any time without consequence, and who to contact if something concerns them.
Also check the voluntariness question honestly. If you are recruiting your own students, your own patients or people you supervise, there is an inherent pressure to agree, and a board will ask what you have done about it. Having somebody else recruit, or removing yourself from the process, is usually the answer and is much better raised by you than by us.
The calendar is the thing you actually control
Boards meet on fixed dates with submission deadlines ahead of them. A perfect packet arriving one day after a deadline waits for the next cycle, which can be a month. That single fact causes more delay in student research than any writing problem.
So work backwards. Find your board's meeting dates and submission deadlines for the next two cycles, pick your target, and build the writing schedule from that rather than submitting whenever a draft happens to be ready. Assume one revision cycle, because most packets get one.
And answer revisions fast when they come, in the board's own numbering, quoting the changed wording so nobody has to hunt for it. Arguing that the original was adequate costs a cycle even when you are right. If this is running alongside chapters and a full course load, the packet is a reasonable thing to hand over while you keep writing.
Questions people actually ask
Will the board judge my research question?
Not usually, and that's not their role. Scientific merit belongs to your committee and your chair. A review board is checking participant protection, informed consent, data handling and internal consistency. Comments that look like criticism of your design are nearly always really about risk or clarity rather than about whether the study is worth doing.
How much detail does the procedure section need?
Enough that a reader could picture a participant's whole experience without asking a question. Where they're recruited, what they're told, what they do, how long it takes, and what happens afterwards. Vagueness reads as unfinished thinking, and it's the section that most often triggers a request for clarification rather than a substantive objection.
Can I recruit my own patients or students?
Sometimes, but expect the question and answer it before it's asked. Any relationship where you have authority creates pressure to agree, and a board will want to see that addressed: someone else handling recruitment, consent taken by a third party, or a clear statement that participation is invisible to you. Raising it yourself reads far better.
What if I need to change something after approval?
You file an amendment before making the change, not afterwards. Deviating from an approved protocol is a genuinely serious problem, more so than most students realise, and it can invalidate the data you've already collected. Amendments are usually quick for minor changes, so the cost of doing it properly is small.
