Interviews and focus groups in a Purdue Global application: guides, recording, consent, confidentiality
A qualitative study is the hardest thing to draw and the easiest to under-draw. The design lives in the researcher's head as a conversation; the board reads it as documents. Purdue University Global's file wants the questions as they will be asked, the recording as it will happen, the transcription as it will be handled, and a confidentiality promise the design can keep. Where those four disagree, the application comes back — not because the study is unsound, but because the drawing contradicts itself.
Short answer
File the interview guide as it will be used, name every recording and what becomes of it, attach a confidentiality agreement for any transcriber or translator, and promise only the confidentiality the design supports. In a focus group you cannot promise what other participants do.
Which pathway do interviews and focus groups belong to?
Purdue Global runs three submission pathways, and a talking study can honestly sit on two. Its exemption worksheet reproduces the category covering interactions involving educational tests, survey procedures, interview procedures or observation of public behaviour, including visual or auditory recording — and notes it is generally appropriate for surveys, focus groups or interviews. That is the exempt route, if one of the criteria is met.
The expedited worksheet holds the other route. Category seven reaches research into the characteristics or behaviour of individuals and groups, plus studies run by survey, interview, oral-history, focus-group, programme-evaluation or quality-assurance methods. Category six reaches data drawn from recordings — voice, video, digital, image — created for research, so if you are recording there is a category built for exactly that.
Purdue Global's expedited worksheet includes a comparison that saves a good deal of argument: the exempt benign-behavioural-intervention category is minimal risk, adults only, non-sensitive and often anonymous; expedited category seven is still minimal risk but higher — sensitive topics, possibly vulnerable groups, and often identifiable data. Interview studies are usually identifiable by construction, and that alone frequently decides the pathway.
One boundary is worth naming because candidates trip on it in the opposite direction. The Common Rule at 45 CFR 46.102 deems certain scholarly and journalistic activities — oral history, journalism, biography, legal research, historical scholarship — not to be research where they focus directly on the individuals the information is about. Purdue Global's expedited worksheet carries that carve-out too. It is narrower than it looks, and it does not rescue a study generalising across a set of interviews.
Why "semi-structured" is not an interview guide
Purdue Global's initial-review procedure sets out what an application must supply, and the list is concrete: participant instructions, questionnaires and assessment instruments, consent documents, recruitment materials, and data collection instruments and debriefing information. An interview guide is a data collection instrument. It goes in.
So the guide has to exist as an object, not as a methodological adjective. "Semi-structured" describes how you will use a guide; it does not replace one. The board needs the opening script, the questions in the order you intend to ask them, the follow-up prompts, and any closing. If the design allows the conversation to travel, say so and show the range — a stated feature of the instrument, not an absence of one.
Two drafting habits cause most findings here. The first is a guide asking something the protocol never mentioned — an adverse event, a manager, a diagnosis — in a file whose risk section describes none of it; the reviewer reads the guide as the true statement of what happens to a participant. The second is a guide with a different question count or topic list from the methods narrative. Same dimension, two documents, two answers.
What must the file say about recording?
Recording is not a detail of method; it is the creation of an identifiable artefact. A voice is an identifier, and a face plainly is. The file should carry each of these identically everywhere.
- What is recorded — audio, video, platform transcript, notes — and whether recording is required or optional. If optional, guide and consent must both say how somebody who declines is still interviewed.
- How consent to record is obtained, and where it is documented. A separate line to initial beats a sentence buried in a paragraph.
- Where the recording lives — device, transfer, storage — with encryption and access named rather than called "secure".
- Who hears it. Everybody with access is on the protocol or under an agreement.
- When identifiers come off — where the transcript is de-identified and pseudonyms assigned — and what happens to the link between pseudonym and person.
- When the recording is destroyed, as a rule tied to a milestone, consistent with the retention obligations that follow closure.
Transcription and translation: the agreement Purdue Global supplies
The moment a recording leaves your hands, somebody outside the protocol holds identifiable material. Purdue Global's toolkit anticipates this with a confidentiality agreement for transcription and translation services, and the file is stronger for using it than for improvising one.
The template carries the protocol number and binds the transcriber or translator, individually and on behalf of any business, to hold in strictest confidence the identity of anybody revealed during transcription; not to disclose or use the material outside the approved research; not to make copies, downloads, screenshots or backups without written authorisation; to store everything on password-protected devices with encrypted storage; to return or securely transfer all materials on completion; and to permanently delete every file from devices, cloud services and backups when the work is done.
Three consequences follow. An automated transcription service is a third party: if the platform retains audio, that retention belongs in the protocol and must square with the agreement. A colleague who "helps with the typing" is a third party too. And where interviews run in another language, the toolkit's guidance on non-English materials applies alongside the requirement that a translated consent document be congruent in substance and intent with the English version, with the translation's accuracy documented.
What can confidentiality honestly promise?
The consent procedure calls for a statement of how far, if at all, records that identify a participant will be held confidential. The operative words are "how far". A promise wider than the design supports is a finding — and an ethical problem, not only a drafting one.
| Dimension | Individual interview | Focus group |
|---|---|---|
| Who else hears it | Only the research team and anybody under a confidentiality agreement | Every other participant, none of whom the researcher controls |
| What may be promised | That the team will not disclose identity, within stated limits | The same, plus that others are asked but cannot be guaranteed to comply |
| Withdrawal | Data can usually be removed before analysis | Contributions are already in the group transcript; say what can be pulled back |
| Recruitment risk | Participation is invisible to others | Attendance itself discloses participation to the room |
| Reporting | Pseudonyms, plus care with identifying detail in quotations | The same, plus care that a role description does not identify its only holder |
Two further limits belong in the document rather than in your intentions. Where a legal or professional obligation to report would override confidentiality — a disclosure of harm, a duty attaching to a licence you hold — say so before somebody speaks. And the consent procedure requires a statement that study records may be inspected by the board or its designees, so a promise that nobody but you will ever see anything is inaccurate on its own terms.
Which participant document, and what goes on it?
Purdue Global pairs each pathway with its own participant-facing document: the exempt protocol with an Exempt Research Information Sheet, the expedited protocol with an Expedited Informed Consent Form. The consent procedure confirms that exempt protocols use a written statement such as an information sheet in lieu of a consent document.
For a non-exempt interview study the elements include a statement that the activity is research and its purpose, the expected commitment, the procedures, foreseeable risks, benefits, the extent of confidentiality, whom to contact, and that participation is voluntary and may stop at any point without penalty. The procedure also directs that consent begin with a concise presentation of the key information most likely to help somebody decide, prefers the second person, bars exculpatory language, and requires the person conducting the process to sign as the researcher. After approval, the protocol number and expiration date go into the header or footer of every approved document.
One structural point: consent is a process, not a signature. The procedure describes securing and maintaining it as ongoing, beginning with recruitment and continuing to the end of somebody's involvement. In an interview study that is literal — the moment to reconfirm that recording is still acceptable is at the start of the recording, and the guide should show you doing it.
What to do next
Qualitative files are returned less often for their methods than for a guide that asks more than the protocol admits, a recording plan that stops at "recorded and transcribed", or a confidentiality sentence a focus group cannot honour. Send us the guide, the draft participant document, the recruitment text and the protocol; a consultant reads them against each other the way Purdue Global's reviewer will and returns a written findings list. Hand it over and we carry the lot: pathway, drafting, filing, and each reply reviewers call for until the letter lands. Your study, your transcripts and your findings never move, and the verdict stays where it belongs — with the board. Start with the free application review, see what an end-to-end engagement covers, or browse the questions candidates ask most.
Sources
- PG IRB, SOP 302: Initial Review — docs.google.com
- PG IRB, SOP 320: Informed Consent — docs.google.com
- PG IRB, IRB-312: Exemption Determination — docs.google.com
- PG IRB, IRB-313: Expedited Review — docs.google.com
- PG IRB, 8_Confidentiality Agreement, Transcription/Translation — docs.google.com
- PG IRB, IRB Toolkit — purdueglobalwriting.center
- PG IRB, Investigator Guide — purdueglobalwriting.center
- 45 CFR 46.102, Definitions — ecfr.gov
- 45 CFR 46.104, Exempt research — ecfr.gov
- 45 CFR 46.116, Informed consent — ecfr.gov