Semi-Structured Interview — Guided Qualitative Interviewing
- Introduced by
- Robert K. Merton and Patricia Kendall (focused interview, 1946); further systematised by Steinar Kvale
- Year
- 1946 (Merton & Kendall); codified as a standard method through the 1980s–1990s
- Type
- Qualitative research method
- Family
- Interview Methods
What does it do?
The semi-structured interview is a qualitative data-collection method in which the researcher prepares a set of key questions or topic areas in advance but remains free to probe, follow up, and reorder as the conversation evolves.
Unlike structured interviews — which fix every question and sequence — or unstructured interviews — which are entirely open — the semi-structured format balances comparability across participants with the flexibility needed to capture the depth and nuance of individual perspectives. It is the most widely used interview format in social science, health, and education research.
The intuition behind the method
Think of a semi-structured interview as a guided conversation with a prepared map but no fixed road. The interviewer arrives with a topic guide listing the core areas to cover — for instance, how nurses experience handover communication — but the actual exchange follows the participant's logic, looping back to earlier topics, pursuing unexpected but revealing threads, and probing with follow-up questions ('Can you tell me more about that?', 'What did you mean by...?'). This flexibility lets the researcher stay consistent enough to compare across multiple participants while remaining open to insights that a rigid questionnaire would have missed entirely.
History
The focused interview was first articulated by Robert K. Merton and Patricia Kendall in their 1946 article in the American Journal of Sociology, developed initially to study audience responses to wartime media. The method was formalised as a distinct qualitative technique through the work of Hy Ruckdeschel and others during the 1970s and codified most influentially by Steinar Kvale, whose 1996 book InterViews established the epistemological and technical foundations still widely taught today. By the 1990s the semi-structured interview had become the dominant data-collection tool in qualitative health, education, and social science research.
When to use it
Semi-structured interviews are appropriate when the research question requires detailed, first-person accounts of experiences, perspectives, meanings, or processes that cannot be captured adequately by surveys or observation alone. They work best when the researcher needs to compare responses across participants while remaining open to unexpected depth. The method is particularly well-suited to exploratory and explanatory research in social science, health, education, and organisational studies. It is not appropriate when population-level prevalence or causal inference is required (use surveys or experiments), when the phenomenon is best understood by observing behaviour in situ (use ethnography or observation), or when the researcher cannot access participants willing to speak candidly in a one-on-one setting.
Limitations and considerations
- Findings are not statistically representative; results cannot be generalised to a population without additional evidence.
- Data quality depends heavily on the interviewer's skill — inexperienced interviewers may lead participants, fail to probe adequately, or miss important cues.
- Transcription and analysis are time-intensive; a study with 20 interviews may produce hundreds of pages of text to code.
- Social desirability and recall bias can distort accounts, particularly for sensitive or past experiences.
- Participation requires willingness to speak at length and in depth, which may exclude certain groups or cultures where direct disclosure is uncommon.
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- Balances consistency across participants with the flexibility to follow unexpected but analytically important threads.
- Produces rich, contextualised, first-person data that reveal meaning and process rather than just frequency.
- The topic guide makes the method transparent and replicable without sacrificing conversational depth.
- Adaptable to a wide range of epistemological traditions — from constructivism to critical realism — and analytic approaches.
- Allows real-time clarification and probing, reducing misunderstanding compared to written surveys.
- Relatively efficient for reaching saturation: 10–30 focused interviews often suffice.
Applications
- Exploring patients' experiences of diagnosis, treatment decisions, or chronic illness management in health research.
- Investigating teachers', students', or parents' perspectives on educational policies, classroom practices, or school transitions.
- Understanding how employees experience organisational change, leadership, or workplace culture in management research.
- Examining how community members perceive risk, displacement, or environmental change in social policy and geography studies.
- Gathering expert knowledge and tacit professional reasoning in policy evaluation and programme assessment.
- Capturing immigrant or minority group experiences of identity, discrimination, or integration in sociology and anthropology.
How it works
How it works
The researcher develops a topic guide — typically 6–12 open-ended questions or thematic areas — grounded in the study's research questions and relevant literature. The guide is piloted with one or two participants outside the target sample and revised for clarity, flow, and coverage. Probes and follow-up prompts are prepared for each main question so the interviewer can deepen responses without leading the participant.
Participants are selected purposively rather than randomly: the goal is to include people who have direct, relevant experience of the phenomenon under study. Typical sample sizes range from 10 to 30, with recruitment continuing until thematic saturation is reached — the point at which additional interviews yield no substantial new themes. Informed consent is obtained before each interview, and confidentiality procedures are established.
Each interview is conducted individually, face-to-face or via audio/video call, and audio-recorded with consent. The interviewer opens with rapport-building before moving to substantive questions, follows the participant's lead in sequencing while ensuring all guide topics are addressed, and uses active listening, silences, and neutral probes to elicit rich, detailed accounts. Typical duration is 45–90 minutes.
Audio recordings are transcribed verbatim, including significant pauses, laughter, and overlapping speech where relevant to analysis. Transcripts are checked against the recording for accuracy and, where relevant, returned to participants for member checking. Field notes made immediately after each interview document the researcher's impressions and any contextual details not captured in the audio.
Transcripts are subjected to systematic qualitative analysis — most commonly thematic analysis, grounded theory coding, or interpretive phenomenological analysis — depending on the study's epistemological stance. The researcher reads each transcript closely, applies codes to meaningful segments, compares codes across transcripts, and organises codes into broader themes or categories that address the research questions.
Emergent themes are interpreted in relation to the original research questions and the existing literature. The researcher checks interpretations against raw data excerpts (negative-case analysis), discusses rival explanations, and situates the findings within a broader theoretical or practical context. Trustworthiness is addressed through strategies such as reflexive journalling, audit trails, and peer debriefing.
Common pitfalls
- Leading questions — framing questions in ways that suggest the expected answer ('Don't you think that...?') produces data that reflects the researcher's assumptions rather than participants' actual views; use neutral, open-ended wording.
- Under-probing — accepting brief or vague responses without asking 'Can you give me an example?' or 'What did that mean for you?' produces surface-level data insufficient for rigorous analysis.
- Ignoring saturation — recruiting a fixed number of participants regardless of whether new themes are still emerging can either under-sample (missing important variation) or over-sample (wasting resources after saturation).
- Conflating the topic guide with a questionnaire — reading questions in rigid sequence and refusing to follow the participant's logic defeats the purpose of the semi-structured format.
- Neglecting reflexivity — failing to document and examine how the researcher's identity, assumptions, and relationship to the topic may have shaped the data and its interpretation.
Frequently asked questions
How is a semi-structured interview different from a structured interview?
A structured interview fixes every question, its wording, and its sequence in advance and is administered identically to every participant — essentially a spoken survey. A semi-structured interview provides a topic guide of key areas but allows the interviewer to probe, reorder, and adapt in response to what participants say. The result is richer, more contextualised data at the cost of strict comparability.
How many participants do I need?
A common heuristic is 10–30 participants for a standalone qualitative study, but the real criterion is thematic saturation — the point at which additional interviews are not adding new themes or meaningful variation. Some well-focused studies reach saturation at 12 interviews; others with diverse populations may require more. Plan to analyse incrementally and monitor saturation rather than fixing the number in advance.
Can semi-structured interviews be conducted online?
Yes. Video calls (Zoom, Teams, etc.) are now widely accepted as equivalent to face-to-face interviews for most research purposes and substantially expand geographic reach. Telephone interviews are also used where video is not feasible. The main practical considerations are ensuring a quiet, private setting for both parties, obtaining clear consent for recording, and acknowledging that some non-verbal cues are lost in audio-only formats.
Which analytic method should I use with semi-structured interview data?
There is no single required analysis. The most common choices are thematic analysis (Braun & Clarke), grounded theory (Charmaz or Strauss & Corbin), interpretive phenomenological analysis (Smith, Flowers & Larkin), and framework analysis (Ritchie & Spencer). The choice should be driven by the research question and epistemological stance, not by convenience. Thematic analysis is the most flexible and widely used; IPA is best when lived experience is the focus; GT when the goal is theory generation.
How do I demonstrate the trustworthiness of my findings?
Lincoln and Guba's criteria are standard: credibility (member checking, prolonged engagement, peer debriefing), transferability (thick description enabling readers to judge applicability), dependability (audit trail of analytic decisions), and confirmability (reflexive journalling documenting researcher positionality). Negative-case analysis — actively seeking data that challenges emerging interpretations — is also important.
Related methods
Sources
Kvale, S., & Brinkmann, S. (2009). InterViews: Learning the Craft of Qualitative Research Interviewing (2nd ed.). Sage.
ISBN: 978-0761925422Copied.Patton, M. Q. (2002). Qualitative Research and Evaluation Methods (3rd ed.). Sage.
Open publication referenceCopied.