Interview protocol: how to write semi-structured interview questions

An interview protocol is the document you prepare for a qualitative interview: what you say at the start and the end, the questions you plan to ask and the probes you can use to follow them up. In a semi-structured interview it guides the conversation without fixing it, so every participant is asked about the same topics while you stay free to follow what they raise. It is also called an interview guide or an interview schedule.

What is an interview protocol?

The term covers two things. The first is the instrument: the list of questions and probes you take into the interview. The second is the procedure for conducting it: how you introduce the study, check consent, record, close the conversation and what you do afterwards. A good protocol covers both, so that every interview in your study is run in a comparable way.

You will need it early. Ethics applications often ask for the questions or topics you plan to cover, and your supervisor will want to see them before you recruit anyone. In a semi-structured interview, the protocol is a plan, not a script you read aloud: you can change the order, skip questions that have already been answered and ask follow-ups that are not written down.

Castillo-Montoya’s interview protocol refinement framework

Milagros Castillo-Montoya’s article Preparing for interview research: The interview protocol refinement framework (2016), in The Qualitative Report, sets out four phases for developing a protocol. She describes it as most suitable for structured and semi-structured interviews.

• Phase 1: Ensuring interview questions align with research questions. You map each interview question onto your research questions in a matrix, to confirm that every question is needed and that no research question is left uncovered.
• Phase 2: Constructing an inquiry-based conversation. You write questions that differ from your research questions, order them by the social rules of ordinary conversation, use a variety of question types and draft a script with likely follow-up and prompt questions.
• Phase 3: Receiving feedback on interview protocols. Others read the protocol closely, and volunteers similar to your participants answer the questions while thinking aloud.
• Phase 4: Piloting the interview protocol. You run the interview under real conditions with a few people who resemble your sample.

Castillo-Montoya notes that not everyone has the time, money or access for a pilot. In that case, she argues, the feedback phase becomes even more important.

Aligning interview questions with research questions

Research questions name what your study sets out to find; interview questions are what you actually say to participants. Brinkmann and Kvale (2015) put the difference this way: research questions are usually formulated in theoretical language, while interview questions should be in the everyday language of the interviewees. One research question may need several interview questions, and one interview question can feed into several research questions.

Castillo-Montoya’s matrix makes this visible. List your interview questions in rows and your research questions in columns, and mark each cell where a question could produce relevant material. Gaps and overloaded columns show up before you collect any data, and you can see where in the interview each topic comes.

The structure of a semi-structured interview protocol

Most protocols follow the same arc. Castillo-Montoya’s example protocol opens with a script and moves from introductory and transition questions to key questions and closing questions.

1. 1

   Introduction script

   Thank the participant, explain the purpose of the study in a few plain sentences, how long the interview will take and what you will ask about. Brinkmann and Kvale call this the briefing.
2. 2

   Consent check

   Confirm that the participant has read the information sheet and signed the consent form, ask again whether you may record, and remind them that they can skip questions or stop. Plan what you will do if they decline recording, such as taking notes.
3. 3

   Opening question

   Start with an easy, non-threatening question that asks for a description, such as how they came to their programme. It helps the participant get used to talking at length.
4. 4

   Key questions

   Put the questions most closely linked to your research questions in the middle, after you have built some rapport. Write a short transition sentence between topics, so the participant knows where the conversation is going.
5. 5

   Probes

   Under each key question, note probes you may need: ‘Can you give me an example?’, ‘What happened next?’ or ‘You mentioned your supervisor. Can you tell me more about that?’
6. 6

   Closing

   End with an easier, reflective question and a chance to add anything: ‘Is there anything we haven’t talked about that you think is important?’ Then thank the participant and explain what happens next, which Brinkmann and Kvale call the debriefing.

Straight after the interview, write a few notes on the setting, the mood and anything that surprised you. They help when you transcribe and when you reflect on your own role, for example in a positionality statement.

Good and poor semi-structured interview questions

Good questions are open, short, one at a time and in the participant’s language. These rewrites follow common textbook advice.

Probes, follow-ups and how many questions to ask

Probes are the questions and prompts you use to get more depth or detail from an answer. Brinkmann and Kvale (2015) distinguish nine types of interview question, several of which work as probes:

• Follow-up: a nod, ‘mm’ or repeating a key word invites the participant to go on.
• Probing: ‘Could you say something more about that?’
• Specifying: ‘What did you actually do then?’
• Interpreting: ‘So you mean that the deadline decided who did what?’
• Silence: a pause gives the participant time to reflect and continue.

The others are introducing, direct, indirect and structuring questions. You cannot script every probe, but listing likely ones under each key question helps you ask them gently and in your participant’s words.

There is no fixed number of questions for a semi-structured interview. What matters is that each question earns its place, that key questions make up most of the protocol and that you can cover them in the time you have promised participants without rushing. A long list of short questions turns the interview into a spoken questionnaire. A pilot interview shows whether your protocol fits.

Getting feedback and piloting the protocol

For phase 3, Castillo-Montoya suggests a close reading: a fellow student or your supervisor reads the questions aloud and checks the structure, length, wording and how easy they are to understand. Then ask a few volunteers who resemble your participants, but will not take part, to answer the questions while thinking aloud. Ask them how difficult a question was to answer, or what they thought of when you used a particular word.

A pilot interview in phase 4 simulates the real thing: introduction, consent, recording, setting and timing. Afterwards, note which questions confused the participant, which produced short answers and whether you ran over time. Listen to yourself as well: did you interrupt, or slip into leading questions? Revise the protocol and keep a note of the changes for your methodology chapter. Ask your supervisor whether your ethics approval covers pilot interviews and whether pilot data may be used in the analysis.

Recording and transcribing the interviews

Follow your university’s rules on consent and data. Many universities decide which apps and equipment are approved for interviews with personal data, and some approve neither phones’ built-in recorders nor third-party apps. Find out what your university allows before you record, and make sure your participant information sheet names every service that will process the recordings.

Check your university’s rules before you upload recordings anywhere: a voice identifies the speaker, so a recording is personal data even without names. Spirelight offers no data processing agreement; your university, not you, would sign one. Transcription runs on EU servers, files are stored with Google Firebase in the US, and AI features run on OpenAI in the US, which your ethics application and participant information sheet must state.

Where your university allows it, you can record in the browser with the online voice recorder, which keeps the recording in your browser until you download it. You can then upload the file to Spirelight for a draft with speaker labels and timestamps, which you correct while you listen. The guide on how to transcribe an interview takes you through the checking, anonymising and formatting.

Using the protocol in the interview and in your dissertation

In the interview, the protocol is a map rather than a script. Castillo-Montoya notes that you probably won’t read your script word for word, but that writing it prepares you to keep the interview conversational, and that a well-prepared protocol leaves you with a map of what you still want to ask when you follow the participant somewhere new. Tick off topics as they come up, even when the participant raises them before you ask.

In the methodology chapter, describe how you developed the protocol: how the questions relate to your research questions, who gave feedback, what the pilot changed and whether you adjusted questions between interviews. Put the final protocol in an appendix. For how interviewing differs across narrative research, phenomenology, grounded theory, ethnography and case study, Creswell and Poth’s Qualitative Inquiry and Research Design (4th edition 2018; 5th edition 2024) is a useful companion.