the interview pocket
The guide, the recording, and the confidentiality no one can promise.
An interview guide is an instrument, and the board reads it the way it reads a questionnaire: question by question, as it will actually be asked. A recording is not a step on the way to a transcript — it is identifiable data from the moment the device is on. And in a focus group, confidentiality is a promise you can only keep for your own conduct. Those three facts shape every document in this pocket, and getting them into the file in plain words is most of the work.
Helen Marsh, PhD · filed 23 August 2026
File the guide verbatim, treat the recording as identifiable data with its own storage and destruction plan, and write consent that states the limits of confidentiality rather than guaranteeing it. In a focus group, you can promise your own conduct and nobody else's.
What is a reviewer actually reading when they open an interview guide?
Three things at once, and it helps to write for all three. The reviewer is reading the questions as words a real person will hear — so an "exploratory prompt" phrased for a proposal reads differently when you imagine it spoken to a colleague on a break. The reviewer is reading for what the questions could surface: distress, disclosure of unsafe practice, admission of an error, an account of harm. And the reviewer is reading for consistency with everything else in the folder, because the guide is the one document that shows what the project actually does.
Herzing's applications call for copies of every questionnaire or survey a project uses. A semi-structured guide belongs in that slot, and "semi-structured" is not a licence to file an outline. What goes in is every question you intend to ask, the probes you intend to use, and — honestly — a note that follow-up probes will pursue whatever the participant raises. Reviewers are used to qualitative work; what they cannot review is a page that keeps its questions to itself.
Write the opening and closing of the encounter into the guide as well. The first minutes carry the consent conversation and the recording permission; the last minutes carry the withdrawal offer and the contact information required by 45 CFR 46.116(b)(7). Putting them on the same page as the questions is how they survive contact with a real interview.
Does recording change the review path?
Often, yes, and in a direction that surprises people who assume qualitative work is inherently low-touch. Among the categories Herzing's Statement of Principles lists as candidates for expedited review under 46.110 is the collection of recordings — voice, video, digital or image — made for the purposes of research, sitting alongside noninvasive collection and studies of how individuals and groups behave, or what characterises them. Expedited is not a demotion: 46.110(b)(2) puts it in the hands of the Chair or an experienced designee holding every authority the board has except the power to disapprove. It is, though, a different lane from exemption, and a file that assumed exemption because "it is only interviews" has assumed the wrong lane.
The reason is that a recording is identifiable by construction. The de-identification list in 164.514(b)(2) names biometric identifiers including voice prints, together with full-face photographic images and anything comparable, among the elements that must be removed before health information counts as de-identified. However thoroughly the transcript is scrubbed of names, the audio itself identifies whoever is speaking. The folder has to face the consequence: recording and transcript are two separate objects, with two storage locations, two access lists and two end points.
The exemption route for interview procedures at 46.104(d)(2) remains open, but its conditions decide everything: responses kept in a form from which identity cannot readily be worked out; or disclosure beyond the study that would not reasonably expose someone to liability, or damage reputation, employability, educational advancement or financial standing; or identifiable responses accompanied by a limited board review. Hold those up against an audio file of a nurse describing a near-miss on their own unit and the honest answer tends to present itself. The review levels piece works through how a path is assigned; the point here is only that the recording is what moves it.
Where does confidentiality actually leak?
Among the basic elements of consent, 46.116(b)(5) asks for a description of how far — if at all — confidentiality will be kept for records that identify someone. Those two words are the regulation conceding that the answer is sometimes "less than you would like". A file promising that confidentiality "will be guaranteed", without saying by whom and against what, has not answered the question. Herzing's Research Protocol Application gives it a checklist line of its own: attach material describing how privacy is maintained and confidentiality kept.
| The leak | Who controls it | What the consent document has to say |
|---|---|---|
| Other people in the room. A focus-group participant repeats what was said | Nobody. You can request discretion; you cannot enforce it | That you ask everyone to keep the discussion private, that you cannot guarantee they will, and that participants should share only what they are content for the group to know |
| Recognition in a small setting. A quotation identifies its speaker by role, unit or turn of phrase | You, at the writing stage | How quotations will be handled: pseudonyms, roles generalised, identifying detail changed, and whether participants may review quotations attributed to them |
| Third parties in the chain. Transcription services, translators, cloud platforms, a co-investigator | You, by naming them and binding them | Who will hear the recording besides the researcher, and under what undertaking |
| Duties that override. A disclosure of abuse, of a person at immediate risk, of practice that must be reported, or a lawful demand for the data | Neither of you. The obligation sits outside the study | Named limits, stated before the recorder starts, in ordinary words — not buried as a clause |
The last row is the one most often missing, and it is the one that protects the participant most. Someone deciding whether to describe a colleague's unsafe practice is entitled to know, in advance, what you would be obliged to do with that account. Saying so does not deter participation as often as people fear; discovering it afterwards ends trust permanently.
What must a focus-group consent form say that a one-to-one form need not?
Four things, and they are all consequences of the room having other people in it.
- The limit of the promise. The researcher's undertaking covers the researcher's conduct and the study's records. It cannot cover what another participant repeats afterwards.
- The ground rules, as agreed rather than announced. What is said in the session stays in the session; no naming of absent colleagues or patients; anyone may pass on any question.
- What withdrawal can and cannot undo. Under 45 CFR 46.116(b)(8) participation is voluntary and may be discontinued at any time. In a group, someone can leave and can ask that their later contributions be excluded — but a contribution already made in front of others cannot be recalled from the others' memories, and if the recording is analysed as a whole, say how a withdrawn voice is handled.
- Who else knows you were there. Attendance itself is disclosure. In a small department, being seen entering the room is information, and the recruitment and scheduling plan should account for it.
Consent for one-to-one interviews is simpler in structure but not lighter in requirement. It still needs the basic elements at 46.116(b): that the activity is research, its purpose and the expected extent of participation, foreseeable risks and discomforts, benefits, the confidentiality description, whom to contact, and the voluntary-participation statement. Under 46.116(a)(5)(i) the document has to begin with a short, focused account of the key information somebody needs in order to decide — which in practice means a first paragraph saying what will be asked, that it will be recorded, and that declining costs nothing. Documentation follows 46.117: a signed form with a copy given to the person who signed it, or the short-form procedure where an oral presentation with a witness is the appropriate route.
What the interview pocket carries into the folder
- The guide, verbatim, including probes, the opening consent script and the closing.
- Consent documents version-matched to the guide, with the confidentiality limits written into them rather than appended.
- A separate recording plan. Device, whether the file leaves the device, where it rests, who can reach it, when it is destroyed — and the same four answers again for the transcript. Both Herzing applications ask for storage, security, retention and disposition for physical and electronic data.
- The third-party list. Transcription, translation, analysis software, anyone who hears the audio, and the undertaking each is under.
- The de-identification convention. What replaces names and units in the transcript, and who does it.
- Recruitment text as it will be sent — and where participants are colleagues, the voluntariness safeguards set out in the piece on surveying staff where you work.
- Site permission from whoever can authorise access to the people and the space, covered in the site permission letter piece.
- CITI certification covering each person the application names — a copy of it, per the checklists, obtainable at no cost to Herzing affiliates via Canvas.
Amendments matter more in qualitative work than anywhere else, because guides evolve. Herzing's Research Protocol Application is direct about it: anything added or altered must carry written board approval before it is put into practice. Adding a question mid-study because the first interviews were interesting is exactly the change that needs filing first.
What to do next
Read your guide aloud to someone who does not know the project, and stop at the first question that makes them wince. That question is the one the board will find, and it is easier to rewrite now than to defend in a response.
Or send it to us. The free application review takes the guide, the consent and the recording plan and reads all three the way a reviewer will, returning a written note on what would hold and what would not. From there this desk drafts the documents, matches their versions, and keeps answering the board until approval arrives. The study belongs to you throughout — so do the recordings and whatever you conclude from them — and the ruling belongs to the board. How it works opens each tab in turn; the FAQ is the same ground at a paragraph apiece.
Sources
- Herzing University Institutional Review Board — the board's page, forms and meeting calendar.
- Herzing University IRB Statement of Principles (PDF) — the expedited categories, including recordings made for research purposes.
- Herzing University IRB Research Protocol Application (PDF) — the consent and confidentiality checklist lines, and the rule on changes.
- Herzing University IRB Bylaws (PDF) — who may conduct an expedited review.
- 45 CFR 46.116 and 45 CFR 46.117 — consent elements, the key-information requirement, and documentation including the short form.
- 45 CFR 46.104(d)(2) and 45 CFR 46.110 — the interview exemption conditions and the expedited procedure.
- 45 CFR 164.514(b) — the de-identification list, including voice prints and facial images.
- CITI Program, Human Subjects Research.
Named forms, routes and prerequisites move between editions; verify every Herzing-specific item against Herzing's current IRB handbook or portal. This practice is independent and not affiliated with Herzing University or its Institutional Review Board.
the other tabs in this file