What a board is actually reading for
Not novelty, not significance, and not how well written it is. A board reads for a short list: whether participants can be harmed, whether they genuinely consented, whether their data is protected, and whether anybody in the study is in a position to feel pressured into taking part.
That last one catches more student researchers than anything else. A nurse recruiting on their own unit, a teacher surveying their own class, a manager interviewing their own team: each of those is an ordinary research design and a coercion question, and a protocol that raises it first and explains the mitigation reads completely differently from one that leaves the board to notice.
- Risk to participants, named plainly rather than minimised
- Consent: how it is obtained, by whom, and how refusal is made genuinely safe
- Whether the researcher holds power over the participants, and what offsets it
- Data protection: where it lives, who opens it, when it is destroyed
- Vulnerable populations, flagged by you before the board flags them
- What happens to somebody who withdraws halfway through
The timeline is the cost, not the writing
A protocol takes days to write and weeks to clear, and the weeks are what damage a program. Boards meet on a schedule, they queue, and a submission that misses a meeting waits for the next one regardless of how urgent your term has become.
So the arithmetic that matters is calendar rather than effort. Find the submission deadlines for the next three meetings, work backwards from the one you intend to hit, and assume at least one round of revisions. Students who plan for a single clean pass are the ones who lose a term to this, and it is entirely predictable.
Consent forms come back more than anything else
Usually for reading level. A consent form written in the vocabulary of your literature review is not consent in any meaningful sense, and boards are increasingly explicit that it has to be understandable to the person signing it rather than to the researcher.
The other frequent cause is a missing consequence: what actually happens if somebody declines, or withdraws after starting. Both need stating in ordinary words, and both need to be true. The pattern is consistent enough to write down, which is why it is worth checking before submission rather than after.
A revision request is not a rejection
It is the ordinary outcome. Most protocols are revised at least once and that says nothing about the study, but the response is where students lose more weeks than they need to, usually by rewriting the whole document and returning it without a map.
The format that clears quickly is a table: the board's numbered comment, what changed in response, and the page where the change now sits. It takes an hour to build and it removes the reviewer's entire checking burden. Where a review has already cost you a term and the design itself now looks doubtful, that is a methods conversation rather than a paperwork one.
What stays with you
The study is yours. The question, the design decisions and the defence of both belong to the person whose name is on the application, and a protocol you cannot discuss in a meeting is worse than no protocol at all. Everything here is prepared with you and returned for your approval before it is submitted.
There is one absolute limit. Nothing is ever submitted to a board under anybody's name but the researcher's own, and no data is fabricated, adjusted or described as collected when it was not. Those are refused at enquiry, and unlike most refusals on this site they are refused for reasons that outlast the degree.
Send what your board publishes
The application template, the submission guidance and any exemplar they have posted. Boards differ enough that a protocol written to a generic standard will be revised on formatting alone, before anybody has read the science in it.
The protocol is built to that template
Written in the board's own sections and vocabulary, with the consent form, the instruments and the recruitment materials assembled as one package rather than sent piecemeal and chased across three separate rounds.
Revisions answered in a table
When it comes back, each numbered comment gets a row: what was asked, what changed, and where in the document to find it. Boards approve that format quickly because it removes the work of checking, which is the work they were dreading.
Questions the desk is asked.
How long does board approval take?
Weeks rather than days, and the variable is the meeting schedule rather than the quality of your submission. Find the deadlines for the next three meetings, work backwards from the one you intend to hit, and assume at least one round of revisions. Planning for a single clean pass is what costs people a term.
Why do consent forms get returned so often?
Reading level, most commonly. A form written in the vocabulary of a literature review is not consent anybody can meaningfully give, and boards say so increasingly plainly. The second cause is a missing consequence: what happens if somebody declines, or withdraws partway through. Both belong in ordinary words.
I am recruiting from my own workplace. Is that a problem?
Not automatically, and it is the thing boards look at hardest. The question is whether anybody could feel pressured to take part because of who you are to them. Raise it yourself, first, and describe what offsets it: recruitment by somebody else, anonymity from you, or an explicit statement that participation is unconnected to anything at work.
Do I need approval for a quality improvement project?
Often not, and it depends on how your institution draws the line between improvement and research. Ask before you start rather than after, because the distinction usually turns on whether you intend to generalise the findings beyond your setting. Getting the answer in writing protects you if anybody asks later.
What can you not do with an IRB submission?
Submit it under anybody's name but yours, or describe data as collected when it was not. Neither moves for any deadline. Everything else, meaning the drafting, the template compliance, the consent materials and the revision response, is prepared with you and approved by you before it goes to the board.
