You need to submit what the IRB/Ethics Committee requests. Most set out what they want, but if it's not clear organisations like UKRIO have checklists that describe what's needed https://ukrio.org/ukrio-resources/publications/researcher-checklist-of-ethics-applications

Many ethics forms have a limited space for providing summaries of information but you should attach additional details and examples of tools/materials used in research. Such as interview schedules, participant information sheets etc (noting too that many ethics committees will provide templates for these for adaptation per your own study).

I wasn't really clear what you're asking here. Is it that you feel your research would be exempt from approval, therefore you don't need to submit? Or that you need to submit but are not sure how much detail to go into?

Best advice is to contact the ethics committee direct with a brief summary of what you want to do, they can indicate what steps should follow and, if it's an urgent piece of vital work, might be able to arrange an expedited review.

Otherwise, you should provide full, clear details of all you intend to do because that's exactly how ethics committees/IRBs can assess work and give approvals. Approvals based on partial information may well cause issues for you, the committee, participants, communities, and your employer and funder later. If in doubt, provide all they ask, when they ask, in the form they ask, noting this is one part of the approvals process and they can only go by how open and honest you are with them.

Petra Boynton Thanks for the response, and sorry for the lack of clarity. What I mean is, if you have little reason to believe that a project will require IRB review at all as it only very indirectly involves people at all -- for example, you're analyzing an already-collected corpus of transcripts of speeches from 50 years ago -- do you still choose to submit that project to the IRB? There is obviously a point at which I suspect most researchers say, "No, there's no reason to bother the IRB with this," and most of those would be very clear (e.g., a physics experiment dealing only with rocks), but what really is the cut-off between "needs to be submitted for the IRB to decide" and "is so clearly not human subject related that it doesn't even need to be submitted"?

If you're unsure I'd ask the IRB. They're used to this and will be happy to help. It's not just about human participants, it'll be about other factors such as identifiable information, researcher safety, community risk/wellbeing etc.

So if the corpus of transcripts and speeches contained information that might identify people or their descendants and potentially put them at risk or cause other distress or nuisance, that would be an ethical issue. See also if analysis of or access to the transcripts might be sensitive politically, religiously or for some other reason. It doesn't mean the work cannot proceed, but caution and expertise would be needed to do the work respectfully and accurately. Of course that doesn't need an ethics committee to enforce as researchers should be doing it anyway, but the additional scrutiny from a committee would be invaluable. (There are also some data regulations on particular datasets that mean work cannot be published for a set time period. Historians can advise more and also note ethical pathways for this kind of research as some of it can be quite complex, particularly things like diaries and letters).

Alternatively, with the corpus example, it might be on a sensitive or challenging topic, something that those analysing it might find difficult or distressing. So again responsible research teams would account for this and there may be no issue with using the corpus for research, but there could be additional ethical scrutiny required to ensure all involved in the research would be supported and not traumatised through it (noting secondary trauma isn't unusual from external data sources). A good IRB would be wanting evidence here about costings to a project for debriefing, managing adverse events, potential counselling for researchers, or specific training for those working with the data.

In terms of the physics experiment with rocks example, that could still require health and safety oversight, more so if it involved fieldwork, and particularly if said rocks had significant cultural value (or the land they're on/in has). It's researchers deciding there's no ethical issue and proceeding because they want to look at the rocks or take them for observation that leads to huge issues for communities who may have an entirely different relationship with the earth and where extractive practices by researchers (including taking from the land) while ignoring local knowledge can and has been enormously damaging. An IRB should be checking that researchers aren't proceeding in selfish, extractive, abusive, or damaging ways (although it's fair to say not all IRBs/Committees have done this and the volume of awful research of this kind is significant).

Ultimately this is about considering the research from a 360 perspective so the immediate study and data may not necessarily be an area of ethical concern but other aspects could have significant ethical issues attached and that would all need addressing and a committee reassured work would proceed carefully, respectfully, safely and with necessary purpose. Sometimes as researchers we're so eager to proceed and fixed in our world view we don't consider other areas of risk, need, responsibility, or care. A good IRB would be interrogating all areas of the research, the researcher(s), and the planned impact and long term sustainability and legacy of the work and how it compares to past research.

TL;DR if you aren't sure ask the IRB. Don't decide it's fine without checking (and remember with or without approval needed the responsibility for ethical conduct begins and ends with us). You can make an enquiry via email, briefly setting out what you want to do. If the IRB says you don't need to submit an application that's fine. But if you're in any way unsure email and find out first, then you can proceed without approval if instructed or put in a full application if needed.

Hope this helps - this kind of thing is what makes ethics so interesting for me, but I appreciate it isn't what everyone is taught nor find particularly engaging.