03 · What You Need to Know
Understanding Repeated Interviews as a Longitudinal Qualitative Method
What Are Repeated Research Interviews?
Repeated interviews involve collecting interview accounts from the same participants on more than one occasion.
They are commonly used in qualitative longitudinal research, where the researcher seeks to understand experiences, processes, and interpretations across time.
For example, a researcher might interview newly appointed teachers before their first semester, midway through the academic year, and after completing their first year of teaching.
Each conversation provides an account situated within a particular stage of the participants' professional experience.
Because the interviews are linked to the same individuals, researchers can examine continuities, developments, contradictions, and reinterpretations within cases.
However, repeated interviews are not automatically longitudinal in a substantively meaningful sense merely because the same person is interviewed twice. The study should have a temporal research purpose and an analytical strategy capable of examining the relationship between observations.
What Makes Longitudinal Qualitative Research Different?
Longitudinal qualitative research is concerned with processes and experiences as they unfold across time.
It may examine how people interpret transitions, how decisions develop, how relationships change, or how earlier experiences shape later understandings.
Thomson and Holland (2003) emphasize the importance of time and temporality in qualitative longitudinal research, including the relationship between individual biographies and broader social circumstances.
Repeated interviewing is one way to generate such evidence, but longitudinal qualitative studies may also use observations, diaries, documents, or other materials collected at multiple periods.
The defining feature is not simply repeated data collection. It is the systematic investigation of temporal processes.
When Is One Interview Insufficient?
A single interview may be insufficient when the phenomenon involves developments that participants cannot fully anticipate or reconstruct retrospectively.
For example, a teacher interviewed before adopting an AI-supported assessment system may describe expectations and concerns. An interview after implementation may reveal unexpected difficulties, changed priorities, or new interpretations.
If the researcher interviews the teacher only at the end, earlier expectations may be reconstructed through knowledge of what eventually occurred.
Repeated interviews allow accounts to be collected at different stages, preserving some of the uncertainty and contextual understanding present at those times.
However, repeated interviewing does not provide unmediated access to experience. Each interview remains a situated account shaped by memory, interaction, and current circumstances.
Repeated Interviews Can Examine Change Within Individuals
One major advantage of repeated interviews is the opportunity to examine how the same participant's experiences and interpretations develop.
For example, a student may initially describe generative AI as a convenient tool for completing assignments, later express concerns about dependence, and eventually develop a more selective approach to its use.
Repeated interviews can explore how these interpretations relate to particular experiences and decisions.
Such evidence differs from interviewing three different students at separate times because the accounts can be linked within one person's trajectory.
Nevertheless, researchers should avoid assuming that every difference between interviews represents a stable change in an underlying attitude.
Participants may emphasize different aspects of an experience because the interview context, question wording, or immediate circumstances have changed.
Repeated Interviews Can Reveal Continuity as Well as Change
Longitudinal research should not be designed on the assumption that participants will necessarily change.
Continuity may be equally important.
For example, a teacher may consistently avoid generative AI despite increasing institutional encouragement. Repeated interviews could reveal the reasons for that continuity and whether the teacher's explanation remains stable.
Similarly, participants may change their practices while maintaining consistent underlying commitments.
Researchers should therefore investigate both stability and development rather than treating change as the only meaningful outcome.
How Do Repeated Interviews Compare with One-Time Interviews?
| Consideration |
One-time interviews |
Repeated interviews |
| Temporal evidence |
Accounts collected during one interview period |
Accounts collected across multiple periods |
| Individual development |
Often reconstructed retrospectively |
Can be examined through linked accounts |
| Changing interpretations |
Limited direct temporal comparison |
Can compare interpretations across occasions |
| Participant relationships |
Usually concentrated in one encounter |
May develop across repeated encounters |
| Recruitment and retention |
One principal participation occasion |
Requires sustained contact or follow-up arrangements |
| Analytical demands |
Within-case and cross-case interpretation |
Within-case temporal and cross-case analysis |
| Distinctive limitations |
Retrospective reconstruction of earlier experiences |
Attrition, repeated-contact effects, and temporal comparability |
The appropriate approach depends on whether the temporal evidence contributes something necessary to answering the research question.
When Should the First Interview Occur?
The first interview should be scheduled at a point that provides a meaningful basis for understanding the process being investigated.
In an intervention study, this may be before implementation. In research on professional transitions, it may be shortly before or after participants enter a new role.
For example, interviewing teachers before an institution introduces a new AI policy can document their expectations and existing practices.
Later interviews can explore how those expectations relate to subsequent experiences.
However, baseline interviewing is not always feasible or necessary. Researchers may begin during an ongoing process and reconstruct earlier circumstances through participant accounts or other sources.
The limitation should be acknowledged when earlier experiences were not documented contemporaneously.
How Many Interview Waves Are Needed?
There is no universally required number of interview waves.
Two interviews may be sufficient when the research question concerns experiences before and after a clearly defined event.
Three or more interviews may be useful when the researcher needs to examine intermediate developments, changing interpretations, or nonlinear trajectories.
For example, interviewing teachers before, during, and after an AI professional-development program may reveal difficulties that would not be apparent from a simple before-and-after comparison.
However, additional waves increase participant burden, fieldwork demands, and opportunities for attrition.
The number should follow the temporal structure of the phenomenon and the analytical questions, not an assumption that more waves automatically produce stronger research.
How Far Apart Should Interviews Be?
Interview spacing should reflect the expected pace of meaningful developments.
Weekly interviews may be appropriate for a rapidly changing experience, such as the early implementation of an intensive educational intervention.
Semester-based interviews may be more suitable for studying teaching practices across an academic year.
Annual interviews may be appropriate for longer professional or organizational transitions.
However, these are illustrative possibilities rather than prescribed schedules.
Closely spaced interviews may produce repetitive accounts when little has changed. Widely spaced interviews may miss important intermediate events or increase reliance on retrospective reconstruction.
Researchers should justify the interval according to the phenomenon and intended analysis.
Fixed-Interval vs. Event-Triggered Interviews
Longitudinal interviews can be scheduled according to predetermined intervals or meaningful events.
Fixed-interval interviewing
Interviews occur according to a planned schedule, such as every semester or every six months.
Event-triggered interviewing
Interviews occur around relevant developments, such as policy implementation, a professional transition, or completion of an intervention.
Fixed intervals can support consistent temporal comparisons, particularly when participants experience broadly similar timelines.
Event-triggered interviews may be more informative when the process unfolds differently across participants or when particular events are analytically important.
Some studies combine both approaches.
Researchers should document the actual timing of each interview because the interval between observations may differ from the planned schedule.
Should the Same Questions Be Asked at Every Wave?
Not necessarily.
Repeated interviews often benefit from shared core topics that allow researchers to examine developments over time.
For example, teachers might be asked at each wave how they currently use generative AI, what influences their decisions, and which institutional conditions matter.
However, later interviews can also explore events and interpretations that emerged after earlier conversations.
Researchers may revisit previous statements and ask whether participants still interpret those experiences in the same way.
The decision about using identical interview questions should therefore consider temporal comparability alongside the need for responsive inquiry.
When the study uses a flexible semi-structured interview guide, changes should be documented so that later differences are not attributed automatically to participant development.
How Can Earlier Interviews Inform Later Questions?
Repeated interviews create opportunities to revisit earlier experiences and interpretations.
For example, a participant might initially express concern about AI-generated inaccuracies. During a later interview, the researcher could ask whether any subsequent experiences affected that concern.
This can reveal continuity, revision, or more complex understandings.
However, referring to earlier statements may also influence how participants construct their current accounts.
A participant may feel expected to remain consistent or may reinterpret earlier statements after hearing them summarized by the interviewer.
Researchers should use previous material carefully and avoid presenting their own interpretation as an established description of what the participant meant.
What Are the Risks of Developing Long-Term Researcher-Participant Relationships?
Repeated encounters can create familiarity and potentially support richer accounts.
Participants may become more comfortable discussing experiences that were difficult to articulate initially.
However, sustained contact can also create expectations about personal relationships, advice, confidentiality, or continued availability.
Researchers should maintain appropriate professional boundaries and communicate the purpose and duration of participation clearly.
Repeated interviews may also change participants' understanding of the research itself, potentially influencing subsequent accounts.
These effects should be considered as part of the data-generation process rather than assumed to be entirely beneficial or harmful.
Attrition Can Change the Evidence Available Over Time
Attrition occurs when participants who contributed earlier interviews do not participate in later waves.
Reasons may include relocation, changing responsibilities, loss of interest, illness, discomfort, or difficulties maintaining contact.
Attrition matters when participants who remain differ in analytically important ways from those who withdraw.
For example, teachers who encounter the greatest difficulties during an AI implementation may become less available for follow-up interviews because of increased workload.
If their experiences disappear from later waves, the remaining accounts may provide an incomplete picture of the process.
Researchers should document attrition and consider its implications rather than treating the remaining participants as equivalent to the original sample.
How Can Participant Retention Be Planned Ethically?
Longitudinal interviewing requires an appropriate system for maintaining contact across waves.
Researchers may obtain permission to retain contact details, establish preferred communication channels, and explain the expected interview schedule.
Contact information should be stored securely and, where appropriate, separately from interview transcripts.
Reminder procedures should be proportionate and consistent with participants' consent.
Researchers should not pressure participants to remain in the study or assume that agreement to the first interview establishes unconditional consent for all future interviews.
Where necessary, consent should be revisited as the study progresses, particularly when the topics or circumstances change.
How Should Changes in Participant Circumstances Be Handled?
Participants' circumstances may change substantially during longitudinal research.
A teacher may change institutions, assume a leadership role, stop using AI tools, or leave the profession.
Such developments may be analytically important rather than reasons to exclude the participant.
However, eligibility and follow-up procedures should be defined in the research protocol.
If the study specifically concerns teachers employed at one institution, a participant's departure may require a decision about whether subsequent interviews remain within scope.
Researchers should distinguish changes that are part of the phenomenon from changes that alter the population or case definition.
Repeated Interviews Require Temporal Analysis
Collecting several interviews from the same participant does not automatically produce a longitudinal analysis.
Researchers need an analytical strategy that examines the relationship between accounts across time.
Possible approaches include within-case chronological analysis, longitudinal thematic analysis, trajectory analysis, and comparison of critical events or turning points.
For example, a researcher might construct a chronological account of how one teacher's AI practices developed across three semesters.
The analysis could then compare trajectories across participants, identifying similarities and differences in the processes involved.
Neale's work on qualitative longitudinal research emphasizes the importance of temporal design and the analytical treatment of change, continuity, and process.
The analytical approach should be specified in relation to the research question rather than selected solely because multiple interviews are available.
Distinguish Actual Change from Changed Narratives
Participants may describe the same past event differently at successive interviews.
This can occur because their interpretation has changed, new information has become available, or the interview context encourages different emphases.
For example, a teacher may initially describe an AI implementation as unsuccessful but later interpret it as an important learning experience.
The later account does not necessarily invalidate the earlier one.
Both may provide evidence about how the participant understands the experience at different times.
Researchers should distinguish changes in reported behavior or circumstances from changes in meaning, memory, and narrative interpretation.
These distinctions are central to rigorous qualitative longitudinal analysis.
Can Repeated Interviews Establish Causation?
Repeated interviews can provide valuable evidence about perceived influences, temporal sequences, and processes through which participants interpret events.
However, they do not automatically establish that a particular event caused a later outcome.
For example, teachers may report greater confidence after completing AI training. Their accounts can illuminate how they believe the training contributed to their development.
Yet other experiences, institutional changes, or independent learning may also have influenced confidence.
Causal conclusions require a defensible research design and attention to alternative explanations.
Repeated qualitative interviews may contribute to causal process analysis in appropriate methodological traditions, but their temporal structure alone is insufficient to establish causal effects.
Should Interviews Be Combined with Other Longitudinal Data?
Repeated interviews may be combined with diaries, observations, documents, assessments, or administrative records when these sources contribute evidence relevant to the research question.
For example, a study of teachers' AI adoption might combine interviews with teaching reflections, policy documents, and records of professional-development participation.
These sources can illuminate different aspects of the process.
However, combining methods does not automatically validate every account or eliminate contradictions.
Differences between interview narratives and documentary evidence may reflect distinct perspectives, measurement limitations, or changes in circumstances.
Researchers should explain how the sources will be interpreted together rather than assuming that agreement is the only meaningful outcome.
When Are Repeated Interviews Unnecessary?
Repeated interviews may be unnecessary when the research question concerns experiences or interpretations that can be adequately investigated through one well-designed conversation.
For example, a study seeking to understand teachers' current perceptions of a recently introduced policy may not require repeated interviews if temporal development is not part of the intended analysis.
Similarly, repeated contact may be impractical or inappropriate when participants face substantial burdens, privacy risks, or rapidly changing availability.
Researchers should not introduce a longitudinal design merely because multiple interviews appear more comprehensive.
The choice should follow the evidentiary needs of the study.
How Should Repeated Interviews Be Reported?
The methods section should explain the temporal research purpose, participant selection, number and timing of interview waves, and procedures for maintaining contact.
Researchers should distinguish the number of participants from the total number of interviews.
For example, 15 participants interviewed three times may contribute up to 45 interviews, but this is not equivalent to interviewing 45 independent participants.
Report missed waves, withdrawals, changes in interview guides, and the analytical treatment of incomplete longitudinal cases.
The rationale for planning the number of research interviews should account for both participant coverage and repeated data collection.
Where saturation is used as a sample-adequacy criterion, researchers should explain how it relates to the longitudinal design rather than assuming that recurring topics across waves establish sufficient temporal understanding.
