How to Write a Strong Interview Guide for Qualitative Research

A good interview guide is a loose script, not a questionnaire. It gives you a clear sequence of topics and starter questions while leaving enough room to follow what the participant actually says. The most common way to wreck a qualitative dataset isn't bad analysis: it's an interview guide that boxes participants into yes/no answers or leads them toward the answer you were hoping for.

This guide is specifically about semi-structured interviews, the format most qualitative research actually runs on. It sits between a fully scripted structured interview (fixed questions, fixed order, built for comparability) and an unstructured, conversational one (no predetermined questions, closer to participant observation): enough structure to keep you on track, enough flexibility to follow what actually matters.

Start with your research question, not your questions

Before writing a single interview question, write down what you actually need to understand. Every question in the guide should trace back to that: if you can't say why a question matters to your research question, cut it. A guide with twenty-five questions "just in case" usually produces a rushed, surface-level interview; a guide with eight well-chosen questions and good follow-up habits produces a deep one.

Structure: funnel, don't interrogate

A workable structure for most semi-structured interviews:

  1. Warm-up – Easy, low-stakes questions that build rapport and get the participant talking in their own words (not yes/no)
  2. Broad opening questions – Open, general questions on the core topic, letting the participant set initial direction
  3. Funnel questions – Progressively more specific questions that narrow toward what you actually need to understand
  4. Probes – Not separate questions, but planned follow-ups ("Can you tell me more about that?" "What happened next?" "How did that feel at the time?") that you use throughout, not just at the end
  5. Closing – A chance for the participant to add anything you didn't ask about, which often surfaces the most useful material in the whole interview

Writing questions that don't lead the witness

  • Avoid double-barreled questions – "How do you feel about the pricing and the onboarding?" forces the participant to answer two things as one
  • Avoid questions that assume the answer – "What frustrates you about X?" presumes frustration exists; "How do you feel about X?" doesn't
  • Prefer "how" and "tell me about" over "why" – "why" questions can feel like they demand a justification and often get a rehearsed answer; "tell me about a time when..." tends to get a real one
  • Keep questions short – a long, multi-clause question is hard to answer coherently and easy to misremember mid-answer

Build in probes on purpose

The guide itself is often less important than what you do after the scripted question. A short list of general-purpose probes, planned in advance, does more for data quality than a longer question list: "What made you say that?" "Can you walk me through what happened?" "Was that typical, or unusual?" Novice interviewers tend to move on too quickly after the first answer: the most useful material is frequently in the second or third follow-up, not the first response.

Pilot it before you rely on it

Run the guide with one or two people who resemble your real participants before you commit to it across a full study. This surfaces the questions that confuse people, take too long, or don't actually produce the kind of answer you need, cheaply, before it costs you real fieldwork time.

What to do with the guide once fieldwork starts

Treat it as a living document, not a fixed script. It's normal, and methodologically fine, to adjust question order or add a probe after the first few interviews reveal something you didn't anticipate, as long as you're consistent about it and can describe what changed and why. What matters for later analysis is that you kept a record of the guide's evolution, since a wildly changing structure across interviews affects how comparable the resulting data is.

Once the interviews are done, that record matters again: Qualibr codes the resulting transcripts and traces every quote back to its source, so a guide that evolved partway through your fieldwork doesn't turn into an analysis you can't defend.

Start freeNo card required, try it on your own interviews

FAQ

How many questions should an interview guide have? There's no fixed number: what matters is coverage of your research question, not question count. As a rough rule of thumb, 6–12 core questions with planned probes usually cover a 45–60 minute interview better than 20+ scripted questions, which tend to crowd out probing.

Semi-structured, structured, or unstructured: which do I need? Structured interviews use a fixed question set and order, useful when you need comparability across a large number of respondents. Unstructured interviews drop the script almost entirely: you start from a topic, not a question list, and let the conversation go where the participant takes it, common in ethnographic work built on repeated, informal contact rather than one scheduled session. Semi-structured sits between the two: a guide as a flexible starting point, useful when you need depth and want to follow unexpected threads without losing your research question entirely, the more common choice for exploratory qualitative research. (The classic "focused interview," associated with Merton, is an early semi-structured format built around a stimulus every participant shared, like a film or an event.)

Should I share the interview guide with participants beforehand? Sometimes, especially for sensitive or complex topics where advance thinking time improves the quality of the answer. The trade-off is spontaneity: rehearsed answers can be more polished but less revealing than the first thing someone says when actually asked.

What do I do with the interview guide once I start analyzing transcripts? Nothing structural. The guide shaped how the interview went, but analysis (coding, thematic analysis, axial coding) works on what was actually said, not on the question list. See our guides on axial coding and thematic analysis for what comes next.

Sources

  • Kallio, H., Pietilä, A.-M., Johnson, M., & Kangasniemi, M. (2016). "Systematic methodological review: developing a framework for a qualitative semi-structured interview guide." Journal of Advanced Nursing, 72(12), 2954–2965.
  • McGrath, C., Palmgren, P. J., & Liljedahl, M. (2019). "Twelve tips for conducting qualitative research interviews." Medical Teacher, 41(9), 1002–1006.
  • ATLAS.ti – "Preparing a Research Interview" guide (consulted for competitive framing/gap-check, not quoted)