Structured, Semi-Structured & Unstructured Interviews

Interviews are one of the most popular ways to collect qualitative data, and the format you choose shapes everything that follows. Pick the wrong one and you can end up with answers too shallow to analyze, or a pile of transcripts too scattered to compare.

Having worked one-on-one with thousands of students, we’ve found the choice is easier once you can see the three options side by side. In this post, I’ll explain how structured, semi-structured and unstructured interviews differ, show you example questions for each, and walk through how a real thesis justified its choice.

---

The three approaches at a glance

The three interviewing approaches sit on a spectrum, from fully scripted to fully open. Here’s how they compare, and what each one costs you:

Approach	How it works	The trade-off
Structured	The same fixed questions, in the same order, with few or no follow-ups	Easy to compare across many people, but answers stay shallow
Semi-structured	A guide of core questions, plus planned probes and a flexible order	A comparable core plus depth, but harder to pace
Unstructured	One or two opening prompts, then the participant steers	Rich and surprising, but scattered and slow to analyze

The rest of this post takes each approach in turn, then shows the same topic run through all three.

---

Structured interviews

Structured interviews are like a well-rehearsed play. Every question is written in advance, and every participant hears the same questions in the same order, much like a survey read aloud. That rigidity gives you consistency: answers are easy to compare, and interviews move quickly, so you can talk to more people.

The trade-off is depth. Because you can’t follow up on an interesting answer, you’ll miss anything you didn’t think to ask about, and responses tend to stay on the surface. In short, the structured approach sacrifices depth for consistency and speed.

So when does it make sense? Structured interviews suit projects that need specific, comparable data from a relatively large number of participants, such as checking whether a pattern holds across several groups. Because everyone answers the same questions, the data is easy to analyze side by side: closed answers can be summarized with descriptive statistics, and short open answers coded with content analysis. As with every methodological choice, the deciding factor is what your research aims and questions need.

---

Unstructured interviews

At the opposite end of the spectrum, unstructured interviews throw out the script. The conversation is guided by the participant’s responses, more jazz improvisation than classical concert. You open with a broad prompt and then follow wherever the participant takes you, letting their priorities and viewpoints steer the discussion.

That freedom can produce rich, surprising insights you’d never have thought to ask about. The cost is time: conversations wander, you may need a second round of interviews to fill gaps, and you’re left with a large, varied dataset that takes real effort to analyze.

Unstructured interviews are best suited to exploratory studies, where little is known about the topic and you want participants to define what matters. Research into personal experiences of trauma, or the realities of starting a business, are typical examples. Because each conversation takes its own shape, unstructured data suits analysis methods that work with whole stories, such as narrative analysis, or thematic analysis when you’re looking for patterns across participants.

---

Semi-structured interviews

Semi-structured interviews sit in the middle. You start with a guide of core questions, but you’re free to change the order, probe deeper and follow themes as they emerge. That flexibility lets you gather comparable data on your core questions while still capturing depth.

Sounds like the perfect blend? The challenge lies in your skill as an interviewer. You need to keep the core questions covered while responding to the flow of conversation, and keep an eye on the clock. It’s easy to get sidetracked, especially with a talkative participant. That’s why DeJonckheere and Vaughn’s guide to semi-structured interviewing recommends pilot testing your interview guide, which “helps you to pace the interview (and not run out of time)”.

Semi-structured interviews work well when you want both comparability and depth, for example when studying how a policy affects different groups of people. That balance is why they’re a popular choice for qualitative dissertations. The shared core of questions also makes semi-structured data a natural fit for thematic analysis, where you look for patterns across participants’ answers to the same questions.

Example questions for each approach

To see the difference in practice, here’s one topic run through all three approaches: how remote workers manage the boundary between work and home life. These are illustrative questions, written for this post rather than taken from a real study.

Every participant gets exactly these four questions, in this order, and most answers come from a fixed set of options. That makes the answers easy to count and compare, but you won’t learn why anyone answered the way they did.

Here the core questions are fixed, but each comes with probes you can use or skip depending on the answer. The probes are where the depth comes from, so plan them in advance rather than inventing them on the spot.

There’s no set list of questions, just an opening prompt and, for many researchers, a short list of topics to keep in mind. You decide the topic, and the participant decides what matters within it, which is exactly why unstructured interviews suit exploratory research.

---

How to choose an approach

Choosing comes down to what your research aims and questions need. Two questions, asked in order, will usually point you to the right approach:

Use this image

Whatever you choose, you’ll need to justify it in your methodology chapter. Here’s how one real Master’s thesis in development sociology, based on interviews with smallholder farmers in Malawi, explained its choice of semi-structured interviews. The “questions” in the first line are a set of topics the team’s earlier survey had barely covered, from seed access to market conditions. Watch how it gives a reason for the flexibility and a reason for the structure.

One note before we break it down: the […] marks a few words that are missing from the source sentence itself. Two sentences, and they make three moves:

The move	Where it happens
Ties the choice to the research questions	“deemed the best approach to answering these questions”. The format follows from what the study needs to find out.
Gives a reason for the flexibility, with a source	Probing lets the researcher follow up “when answers hint at a relevant story, uncover an unspoken logic, or need further clarification (Maxwell, 2005)”.
Gives a reason for the structure	“the breadth of issues… required a structured approach… within a reasonable timeframe (approximately an hour per interview)”. The study had a lot to cover in a limited time.

That third move is the one most students skip. Justify both halves of a semi-structured design: why you need room to probe, and why you still need a fixed core.

Once you’ve settled on an approach, the next step is turning it into a written interview guide and testing it in a pilot interview or two. Get those two things right, and the interviews themselves become far easier to run.

Still have questions?

How many interviews do I need for a dissertation?

There’s no fixed number. Qualitative studies usually keep interviewing until new interviews stop adding new insights, a point called saturation. In one well-known study of 25 in-depth interviews, Hennink, Kaiser and Marconi found the full range of themes had appeared by nine interviews, but it took 16 to 24 to understand those themes in depth. Your program and advisor will have their own expectations, so check those before you plan your sample.

How many questions should a semi-structured interview have?

There’s no set number, but most guides work with a small set of core questions, each with a few planned probes, rather than a long list. Writing those questions down as your interview guide (often called an interview protocol) also helps you explain your approach in your methodology chapter. Fewer, broader questions leave room for the follow-ups that give semi-structured interviews their depth. Pilot your guide to check it fits the time you have with each participant.

Can I combine structured and semi-structured questions?

Yes. Many interview guides open with a few short, fixed questions, such as background details, and then move into open questions with probes. Just explain the mix in your methodology chapter, including why each part is there. Whether every participant needs exactly the same questions is a separate choice, and it depends on your design.

Are structured interviews qualitative or quantitative?

They can be either. If the questions have fixed answer options, such as yes or no or a rating scale, you’re collecting quantitative data. If the questions are open but asked in the same order every time, the answers are qualitative. Semi-structured and unstructured interviews collect qualitative data.

How long should a research interview be?

It depends on your approach and how much you need to cover. Structured interviews are often shorter, while semi-structured and unstructured interviews commonly run for up to an hour or more. Pilot your guide to see how long it really takes, and tell participants how long to expect when you invite them. Our guide to running qualitative interviews covers the practical side of planning one, from building rapport to recording.

Can’t find your answer here? Ask a Grad Coach directly – the initial chat is free.

Speak with a friendly coach →