what are semi structured interviews in research

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

• Knowledge Base

Methodology

Semi-Structured Interview | Definition, Guide & Examples

Published on January 27, 2022 by Tegan George . Revised on June 22, 2023.

A semi-structured interview is a data collection method that relies on asking questions within a predetermined thematic framework. However, the questions are not set in order or in phrasing.

In research, semi-structured interviews are often qualitative in nature. They are generally used as an exploratory tool in marketing, social science, survey methodology, and other research fields.

They are also common in field research with many interviewers, giving everyone the same theoretical framework, but allowing them to investigate different facets of the research question .

• Structured interviews : The questions are predetermined in both topic and order.
• Unstructured interviews : None of the questions are predetermined.
• Focus group interviews : The questions are presented to a group instead of one individual.

Table of contents

What is a semi-structured interview, when to use a semi-structured interview, advantages of semi-structured interviews, disadvantages of semi-structured interviews, semi-structured interview questions, how to conduct a semi-structured interview, how to analyze a semi-structured interview, presenting your results (with example), other interesting articles, frequently asked questions about semi-structured interviews.

Semi-structured interviews are a blend of structured and unstructured types of interviews.

• Unlike in an unstructured interview, the interviewer has an idea of what questions they will ask.
• Unlike in a structured interview, the phrasing and order of the questions is not set.

Semi-structured interviews are often open-ended, allowing for flexibility. Asking set questions in a set order allows for easy comparison between respondents, but it can be limiting. Having less structure can help you see patterns, while still allowing for comparisons between respondents.

Semi-structured interviews are best used when:

• You have prior interview experience. Spontaneous questions are deceptively challenging, and it’s easy to accidentally ask a leading question or make a participant uneasy.
• Your research question is exploratory in nature. Participant answers can guide future research questions and help you develop a more robust knowledge base for future research.

Just like in structured interviews, it is critical that you remain organized and develop a system for keeping track of participant responses. However, since the questions are less set than in a structured interview, the data collection and analysis become a bit more complex.

Differences between different types of interviews

Make sure to choose the type of interview that suits your research best. This table shows the most important differences between the four types.

Fixed questions
Fixed order of questions
Fixed number of questions
Option to ask additional questions

Semi-structured interviews come with many advantages.

Best of both worlds

No distractions, detail and richness.

However, semi-structured interviews also have their downsides.

Low validity

High risk of research bias, difficult to develop good semi-structured interview questions.

Since they are often open-ended in style, it can be challenging to write semi-structured interview questions that get you the information you’re looking for without biasing your responses. Here are a few tips:

• Define what areas or topics you will be focusing on prior to the interview. This will help you write a framework of questions that zero in on the information you seek.
• Write yourself a guide to refer to during the interview, so you stay focused. It can help to start with the simpler questions first, moving into the more complex ones after you have established a comfortable rapport.
• Be as clear and concise as possible, avoiding jargon and compound sentences.
• How often per week do you go to the gym? a) 1 time; b) 2 times; c) 3 times; d) 4 or more times
• If yes: What feelings does going to the gym bring out in you?
• If no: What do you prefer to do instead?
• If yes: How did this membership affect your job performance? Did you stay longer in the role than you would have if there were no membership?

Once you’ve determined that a semi-structured interview is the right fit for your research topic , you can proceed with the following steps.

Step 1: Set your goals and objectives

You can use guiding questions as you conceptualize your research question, such as:

• What are you trying to learn or achieve from a semi-structured interview?
• Why are you choosing a semi-structured interview as opposed to a different type of interview, or another research method?

If you want to proceed with a semi-structured interview, you can start designing your questions.

Step 2: Design your questions

Try to stay simple and concise, and phrase your questions clearly. If your topic is sensitive or could cause an emotional response, be mindful of your word choices.

One of the most challenging parts of a semi-structured interview is knowing when to ask follow-up or spontaneous related questions. For this reason, having a guide to refer back to is critical. Hypothesizing what other questions could arise from your participants’ answers may also be helpful.

Step 3: Assemble your participants

There are a few sampling methods you can use to recruit your interview participants, such as:

• Voluntary response sampling : For example, sending an email to a campus mailing list and sourcing participants from responses.
• Stratified sampling of a particular characteristic trait of interest to your research, such as age, race, ethnicity, or gender identity.

Step 4: Decide on your medium

It’s important to determine ahead of time how you will be conducting your interview. You should decide whether you’ll be conducting it live or with a pen-and-paper format. If conducted in real time, you also need to decide if in person, over the phone, or via videoconferencing is the best option for you.

Note that each of these methods has its own advantages and disadvantages:

• Pen-and-paper may be easier for you to organize and analyze, but you will receive more prepared answers, which may affect the reliability of your data.
• In-person interviews can lead to nervousness or interviewer effects, where the respondent feels pressured to respond in a manner they believe will please you or incentivize you to like them.

Step 5: Conduct your interviews

As you conduct your interviews, keep environmental conditions as constant as you can to avoid bias. Pay attention to your body language (e.g., nodding, raising eyebrows), and moderate your tone of voice.

Relatedly, one of the biggest challenges with semi-structured interviews is ensuring that your questions remain unbiased. This can be especially challenging with any spontaneous questions or unscripted follow-ups that you ask your participants.

After you’re finished conducting your interviews, it’s time to analyze your results. First, assign each of your participants a number or pseudonym for organizational purposes.

The next step in your analysis is to transcribe the audio or video recordings. You can then conduct a content or thematic analysis to determine your categories, looking for patterns of responses that stand out to you and test your hypotheses .

Transcribing interviews

Before you get started with transcription, decide whether to conduct verbatim transcription or intelligent verbatim transcription.

• If pauses, laughter, or filler words like “umm” or “like” affect your analysis and research conclusions, conduct verbatim transcription and include them.
• If not, you can conduct intelligent verbatim transcription, which excludes fillers, fixes any grammatical issues, and is usually easier to analyze.

Transcribing presents a great opportunity for you to cleanse your data . Here, you can identify and address any inconsistencies or questions that come up as you listen.

Your supervisor might ask you to add the transcriptions to the appendix of your paper.

Coding semi-structured interviews

Next, it’s time to conduct your thematic or content analysis . This often involves “coding” words, patterns, or recurring responses, separating them into labels or categories for more robust analysis.

Due to the open-ended nature of many semi-structured interviews, you will most likely be conducting thematic analysis, rather than content analysis.

• You closely examine your data to identify common topics, ideas, or patterns. This can help you draw preliminary conclusions about your participants’ views, knowledge or experiences.
• After you have been through your responses a few times, you can collect the data into groups identified by their “code.” These codes give you a condensed overview of the main points and patterns identified by your data.
• Next, it’s time to organize these codes into themes. Themes are generally broader than codes, and you’ll often combine a few codes under one theme. After identifying your themes, make sure that these themes appropriately represent patterns in responses.

Analyzing semi-structured interviews

Once you’re confident in your themes, you can take either an inductive or a deductive approach.

• An inductive approach is more open-ended, allowing your data to determine your themes.
• A deductive approach is the opposite. It involves investigating whether your data confirm preconceived themes or ideas.

After your data analysis, the next step is to report your findings in a research paper .

• Your methodology section describes how you collected the data (in this case, describing your semi-structured interview process) and explains how you justify or conceptualize your analysis.
• Your discussion and results sections usually address each of your coded categories.
• You can then conclude with the main takeaways and avenues for further research.

Example of interview methodology for a research paper

Let’s say you are interested in vegan students on your campus. You have noticed that the number of vegan students seems to have increased since your first year, and you are curious what caused this shift.

You identify a few potential options based on literature:

• Perceptions about personal health or the perceived “healthiness” of a vegan diet
• Concerns about animal welfare and the meat industry
• Increased climate awareness, especially in regards to animal products
• Availability of more vegan options, making the lifestyle change easier

Anecdotally, you hypothesize that students are more aware of the impact of animal products on the ongoing climate crisis, and this has influenced many to go vegan. However, you cannot rule out the possibility of the other options, such as the new vegan bar in the dining hall.

Since your topic is exploratory in nature and you have a lot of experience conducting interviews in your work-study role as a research assistant, you decide to conduct semi-structured interviews.

You have a friend who is a member of a campus club for vegans and vegetarians, so you send a message to the club to ask for volunteers. You also spend some time at the campus dining hall, approaching students at the vegan bar asking if they’d like to participate.

Here are some questions you could ask:

• Do you find vegan options on campus to be: excellent; good; fair; average; poor?
• How long have you been a vegan?
• Follow-up questions can probe the strength of this decision (i.e., was it overwhelmingly one reason, or more of a mix?)

Depending on your participants’ answers to these questions, ask follow-ups as needed for clarification, further information, or elaboration.

• Do you think consuming animal products contributes to climate change? → The phrasing implies that you, the interviewer, do think so. This could bias your respondents, incentivizing them to answer affirmatively as well.
• What do you think is the biggest effect of animal product consumption? → This phrasing ensures the participant is giving their own opinion, and may even yield some surprising responses that enrich your analysis.

After conducting your interviews and transcribing your data, you can then conduct thematic analysis, coding responses into different categories. Since you began your research with several theories about campus veganism that you found equally compelling, you would use the inductive approach.

Once you’ve identified themes and patterns from your data, you can draw inferences and conclusions. Your results section usually addresses each theme or pattern you found, describing each in turn, as well as how often you came across them in your analysis. Feel free to include lots of (properly anonymized) examples from the data as evidence, too.

If you want to know more about statistics , methodology , or research bias , make sure to check out some of our other articles with explanations and examples.

• Student’s  t -distribution
• Normal distribution
• Null and Alternative Hypotheses
• Chi square tests
• Confidence interval
• Quartiles & Quantiles
• Cluster sampling
• Stratified sampling
• Data cleansing
• Reproducibility vs Replicability
• Peer review
• Prospective cohort study

Research bias

• Implicit bias
• Cognitive bias
• Placebo effect
• Hawthorne effect
• Hindsight bias
• Affect heuristic
• Social desirability bias

A semi-structured interview is a blend of structured and unstructured types of interviews. Semi-structured interviews are best used when:

• You have prior interview experience. Spontaneous questions are deceptively challenging, and it’s easy to accidentally ask a leading question or make a participant uncomfortable.

The four most common types of interviews are:

• Structured interviews : The questions are predetermined in both topic and order. 
• Semi-structured interviews : A few questions are predetermined, but other questions aren’t planned.

Social desirability bias is the tendency for interview participants to give responses that will be viewed favorably by the interviewer or other participants. It occurs in all types of interviews and surveys , but is most common in semi-structured interviews , unstructured interviews , and focus groups .

Social desirability bias can be mitigated by ensuring participants feel at ease and comfortable sharing their views. Make sure to pay attention to your own body language and any physical or verbal cues, such as nodding or widening your eyes.

This type of bias can also occur in observations if the participants know they’re being observed. They might alter their behavior accordingly.

The interviewer effect is a type of bias that emerges when a characteristic of an interviewer (race, age, gender identity, etc.) influences the responses given by the interviewee.

There is a risk of an interviewer effect in all types of interviews , but it can be mitigated by writing really high-quality interview questions.

Inductive reasoning is a bottom-up approach, while deductive reasoning is top-down.

Inductive reasoning takes you from the specific to the general, while in deductive reasoning, you make inferences by going from general premises to specific conclusions.

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

George, T. (2023, June 22). Semi-Structured Interview | Definition, Guide & Examples. Scribbr. Retrieved August 21, 2024, from https://www.scribbr.com/methodology/semi-structured-interview/

Is this article helpful?

Tegan George

Other students also liked, structured interview | definition, guide & examples, unstructured interview | definition, guide & examples, what is a focus group | step-by-step guide & examples, what is your plagiarism score.

• What is a semi-structured interview?

Last updated

5 February 2023

Reviewed by

Cathy Heath

Short on time? Get an AI generated summary of this article instead

When designed correctly, user interviews go much deeper than surface-level survey responses. They can provide new information about how people interact with your products and services, and shed light on the underlying reasons behind these habits.

Semi-structured user interviews are widely considered one of the most effective tools for doing this kind of qualitative research , depending on your specific goals. As the name suggests, the semi-structured format allows for a more natural, conversational flow, while still being organized enough to collect plenty of actionable data .

Analyze semi-structured interviews

Bring all your semi-structured interviews into one place to analyze and understand

A semi-structured interview is a qualitative research method used to gain an in-depth understanding of the respondent's feelings and beliefs on specific topics. As the interviewer prepares the questions ahead of time, they can adjust the order, skip any that are redundant, or create new ones. Additionally, the interviewer should be prepared to ask follow-up questions and probe for more detail.

Semi-structured interviews typically last between 30 and 60 minutes and are usually conducted either in person or via a video call. Ideally, the interviewer can observe the participant's verbal and non-verbal cues in real-time, allowing them to adjust their approach accordingly. The interviewer aims for a conversational flow that helps the participant talk openly while still focusing on the primary topics being researched.

Once the interview is over, the researcher analyzes the data in detail to draw meaningful results. This involves sorting the data into categories and looking for patterns and trends. This semi-structured interview approach provides an ideal framework for obtaining open-ended data and insights.

• When to use a semi-structured interview?

Semi-structured interviews are considered the "best of both worlds" as they tap into the strengths of structured and unstructured methods. Researchers can gather reliable data while also getting unexpected insights from in-depth user feedback.

Semi-structured interviews can be useful during any stage of the UX product-development process, including exploratory research to better understand a new market or service. Further down the line, this approach is ideal for refining existing designs and discovering areas for improvement. Semi-structured interviews can even be the first step when planning future research projects using another method of data collection.

• Advantages of semi-structured interviews

Flexibility

This style of interview is meant to be adapted according to the answers and reactions of the respondent, which gives a lot of flexibility. Semi-structured interviews encourage two-way communication, allowing themes and ideas to emerge organically.

Respondent comfort

The semi-structured format feels more natural and casual for participants than a formal interview. This can help to build rapport and more meaningful dialogue.

Semi-structured interviews are excellent for user experience research because they provide rich, qualitative data about how people really experience your products and services.

Open-ended questions allow the respondent to provide nuanced answers, with the potential for more valuable insights than other forms of data collection, like structured interviews , surveys , or questionnaires.

• Disadvantages of semi-structured interviews

Can be unpredictable

Less structure brings less control, especially if the respondent goes off tangent or doesn't provide useful information. If the conversation derails, it can take a lot of effort to bring the focus back to the relevant topics.

Lack of standardization

Every semi-structured interview is unique, including potentially different questions, so the responses collected are very subjective. This can make it difficult to draw meaningful conclusions from the data unless your team invests the time in a comprehensive analysis.

Compared to other research methods, unstructured interviews are not as consistent or "ready to use."

• Best practices when preparing for a semi-structured interview

While semi-structured interviews provide a lot of flexibility, they still require thoughtful planning. Maximizing the potential of this research method will depend on having clear goals that help you narrow the focus of the interviews and keep each session on track.

After taking the time to specify these parameters, create an interview guide to serve as a framework for each conversation. This involves crafting a range of questions that can explore the necessary themes and steer the conversation in the right direction. Everything in your interview guide is optional (that's the beauty of being "semi" structured), but it's still an essential tool to help the conversation flow and collect useful data.

Best practices to consider while designing your interview questions include:

Prioritize open-ended questions

Promote a more interactive, meaningful dialogue by avoiding questions that can be answered with a simple yes or no, otherwise known as close-ended questions.

Stick with "what," "when," "who," "where," "why," and "how" questions, which allow the participant to go beyond the superficial to express their ideas and opinions. This approach also helps avoid jargon and needless complexity in your questions.

Open-ended questions help the interviewer uncover richer, qualitative details, which they can build on to get even more valuable insights.

Plan some follow-up questions

When preparing questions for the interview guide, consider the responses you're likely to get and pair them up with some effective, relevant follow-up questions. Factual questions should be followed by ones that ask an opinion.

Planning potential follow-up questions will help you to get the most out of a semi-structured interview. They allow you to delve deeper into the participant's responses or hone in on the most important themes of your research focus.

Follow-up questions are also invaluable when the interviewer feels stuck and needs a meaningful prompt to continue the conversation.

Avoid leading questions

Leading questions are framed toward a predetermined answer. This makes them likely to result in data that is biased, inaccurate, or otherwise unreliable.

For example, asking "Why do you think our services are a good solution?" or "How satisfied have you been with our services?" will leave the interviewee feeling pressured to agree with some baseline assumptions.

Interviewers must take the time to evaluate their questions and make a conscious effort to remove any potential bias that could get in the way of authentic feedback.

Asking neutral questions is key to encouraging honest responses in a semi-structured interview. For example, "What do you consider to be the advantages of using our services?" or simply "What has been your experience with using our services?"

Neutral questions are effective in capturing a broader range of opinions than closed questions, which is ultimately one of the biggest benefits of using semi-structured interviews for research.

Use the critical incident method

The critical incident method is an approach to interviewing that focuses on the past behavior of respondents, as opposed to hypothetical scenarios. One of the challenges of all interview research methods is that people are not great at accurately recalling past experiences, or answering future-facing, abstract questions.

The critical incident method helps avoid these limitations by asking participants to recall extreme situations or 'critical incidents' which stand out in their memory as either particularly positive or negative. Extreme situations are more vivid so they can be recalled more accurately, potentially providing more meaningful insights into the interviewee’s experience with your products or services.

• Best practices while conducting semi-structured interviews

Encouraging interaction is the key to collecting more specific data than is typically possible during a formal interview. Facilitating an effective semi-structured interview is a balancing act between asking prepared questions and creating the space for organic conversation. Here are some guidelines for striking the right tone.

Beginning the interview

Make participants feel comfortable by introducing yourself and your role at the organization and displaying appropriate body language.

Outline the purpose of the interview to give them an idea of what to expect. For example, explain that you want to learn more about how people use your product or service.

It's also important to thank them for their time in advance and emphasize there are no right or wrong answers.

Practice active listening

Build trust and rapport throughout the interview with active listening techniques, focusing on being present and demonstrating that you're paying attention by responding thoughtfully. Engage with the participant by making eye contact, nodding, and giving verbal cues like "Okay, I see," "I understand," and "M-hm."

Avoid the temptation to rush to fill any silences while they're in the middle of responding, even if it feels awkward. Give them time to finish their train of thought before interrupting with feedback or another prompt. Embracing these silences is essential for active listening because it's a sign of a productive interview with meaningful, candid responses.

Practicing these techniques will ensure the respondent feels heard and respected, which is critical for gathering high-quality information.

Ask clarifying questions in real time

In a semi-structured interview, the researcher should always be on the lookout for opportunities to probe into the participant's thoughts and opinions.

Along with preparing follow-up questions, get in the habit of asking clarifying questions whenever possible. Clarifying questions are especially important for user interviews because people often provide vague responses when discussing how they interact with products and services.

Being asked to go deeper will encourage them to give more detail and show them you’re taking their opinions seriously and are genuinely interested in understanding their experiences.

Some clarifying questions that can be asked in real-time include:

"That's interesting. Could you give me some examples of X?"

"What do you mean when you say "X"?"

"Why is that?"

"It sounds like you're saying [rephrase their response], is that correct?"

Minimize note-taking

In a wide-ranging conversation, it's easy to miss out on potentially valuable insights by not staying focused on the user. This is why semi-structured interviews are generally recorded (audio or video), and it's common to have a second researcher present to take notes.

The person conducting the interview should avoid taking notes because it's a distraction from:

Keeping track of the conversation

Engaging with the user

Asking thought-provoking questions

Watching you take notes can also have the unintended effect of making the participant feel pressured to give shallower, shorter responses—the opposite of what you want.

Concluding the interview

Semi-structured interviews don't come with a set number of questions, so it can be tricky to bring them to an end. Give the participant a sense of closure by asking whether they have anything to add before wrapping up, or if they want to ask you any questions, and then give sincere thanks for providing honest feedback.

Don't stop abruptly once all the relevant topics have been discussed or you're nearing the end of the time that was set aside. Make them feel appreciated!

• Analyzing the data from semi-structured interviews

In some ways, the real work of semi-structured interviews begins after all the conversations are over, and it's time to analyze the data you've collected. This process will focus on sorting and coding each interview to identify patterns, often using a mix of qualitative and quantitative methods.

Some of the strategies for making sense of semi-structured interviews include:

Thematic analysis : focuses on the content of the interviews and identifying common themes

Discourse analysis : looks at how people express feelings about themes such as those involving politics, culture, and power

Qualitative data mapping: a visual way to map out the correlations between different elements of the data

Narrative analysis : uses stories and language to unlock perspectives on an issue

Grounded theory : can be applied when there is no existing theory that could explain a new phenomenon

Should you be using a customer insights hub?

Do you want to discover previous research faster?

Do you share your research findings with others?

Do you analyze research data?

Start for free today, add your research, and get to key insights faster

Editor’s picks

Last updated: 18 April 2023

Last updated: 27 February 2023

Last updated: 5 February 2023

Last updated: 16 April 2023

Last updated: 16 August 2024

Last updated: 9 March 2023

Last updated: 30 April 2024

Last updated: 12 December 2023

Last updated: 11 March 2024

Last updated: 4 July 2024

Last updated: 6 March 2024

Last updated: 5 March 2024

Last updated: 13 May 2024

Latest articles

Related topics, .css-je19u9{-webkit-align-items:flex-end;-webkit-box-align:flex-end;-ms-flex-align:flex-end;align-items:flex-end;display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-flex-direction:row;-ms-flex-direction:row;flex-direction:row;-webkit-box-flex-wrap:wrap;-webkit-flex-wrap:wrap;-ms-flex-wrap:wrap;flex-wrap:wrap;-webkit-box-pack:center;-ms-flex-pack:center;-webkit-justify-content:center;justify-content:center;row-gap:0;text-align:center;max-width:671px;}@media (max-width: 1079px){.css-je19u9{max-width:400px;}.css-je19u9>span{white-space:pre;}}@media (max-width: 799px){.css-je19u9{max-width:400px;}.css-je19u9>span{white-space:pre;}} decide what to .css-1kiodld{max-height:56px;display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}@media (max-width: 1079px){.css-1kiodld{display:none;}} build next, decide what to build next, log in or sign up.

Get started for free

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Fam Med Community Health
• v.7(2); 2019

Semistructured interviewing in primary care research: a balance of relationship and rigour

Melissa dejonckheere.

1 Department of Family Medicine, University of Michigan, Ann Arbor, Michigan, USA

Lisa M Vaughn

2 Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio, USA

3 Division of Emergency Medicine, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, USA

Associated Data

fmch-2018-000057supp001.pdf

Semistructured in-depth interviews are commonly used in qualitative research and are the most frequent qualitative data source in health services research. This method typically consists of a dialogue between researcher and participant, guided by a flexible interview protocol and supplemented by follow-up questions, probes and comments. The method allows the researcher to collect open-ended data, to explore participant thoughts, feelings and beliefs about a particular topic and to delve deeply into personal and sometimes sensitive issues. The purpose of this article was to identify and describe the essential skills to designing and conducting semistructured interviews in family medicine and primary care research settings. We reviewed the literature on semistructured interviewing to identify key skills and components for using this method in family medicine and primary care research settings. Overall, semistructured interviewing requires both a relational focus and practice in the skills of facilitation. Skills include: (1) determining the purpose and scope of the study; (2) identifying participants; (3) considering ethical issues; (4) planning logistical aspects; (5) developing the interview guide; (6) establishing trust and rapport; (7) conducting the interview; (8) memoing and reflection; (9) analysing the data; (10) demonstrating the trustworthiness of the research; and (11) presenting findings in a paper or report. Semistructured interviews provide an effective and feasible research method for family physicians to conduct in primary care research settings. Researchers using semistructured interviews for data collection should take on a relational focus and consider the skills of interviewing to ensure quality. Semistructured interviewing can be a powerful tool for family physicians, primary care providers and other health services researchers to use to understand the thoughts, beliefs and experiences of individuals. Despite the utility, semistructured interviews can be intimidating and challenging for researchers not familiar with qualitative approaches. In order to elucidate this method, we provide practical guidance for researchers, including novice researchers and those with few resources, to use semistructured interviewing as a data collection strategy. We provide recommendations for the essential steps to follow in order to best implement semistructured interviews in family medicine and primary care research settings.

Introduction

Semistructured interviews can be used by family medicine researchers in clinical settings or academic settings even with few resources. In contrast to large-scale epidemiological studies, or even surveys, a family medicine researcher can conduct a highly meaningful project with interviews with as few as 8–12 participants. For example, Chang and her colleagues, all family physicians, conducted semistructured interviews with 10 providers to understand their perspectives on weight gain in pregnant patients. 1 The interviewers asked questions about providers’ overall perceptions on weight gain, their clinical approach to weight gain during pregnancy and challenges when managing weight gain among pregnant patients. Additional examples conducted by or with family physicians or in primary care settings are summarised in table 1 . 1–6

Examples of research articles using semistructured interviews in primary care research

Article	Study purpose	Context/setting	Use of interviews
Chang T, Llanes M, Gold KJ, . Perspectives about and approaches to weight gain in pregnancy: a qualitative study of physicians and nurse midwives. 2013;13:47.	To explore prenatal care providers’ perspectives on patient weight gain during pregnancy	University hospital in the USA	10 semistructured interviews with prenatal care providers (family physicians, obstetricians, nurse midwives); thematic analysis
Croxson CH, Ashdown HF, Hobbs FR. GPs’ perceptions of workload in England: a qualitative interview study. 2017.	To understand perceptions of provider workload	NHS in England	34 semistructured interviews with general practitioners; thematic analysis
DeJonckheere M, Robinson CH, Evans L, Designing for clinical change: creating an intervention to implement new statin guidelines in a primary care clinic. 2018;5 .	To elicit provider perspectives of their uptake of new statin guidelines. To tailor a local quality improvement intervention to improve statin prescribing.	Veterans Affairs Medical Center in the USA	15 semistructured interviews with providers (primary care physicians and clinical pharmacists); deductive thematic analysis
Griffiths F, Lowe P, Boardman F, . Becoming pregnant: exploring the perspectives of women living with diabetes. 2008;58:184–90 .	To explore women's accounts of their journeys to becoming pregnant while living with type 1 diabetes	Four UK specialist diabetes antenatal clinics	15 semistructured interviews with women with pregestational type 1 diabetes; thematic analysis
Saigal P, Takemura Y, Nishiue T, Factors considered by medical students when formulating their specialty preferences in Japan: findings from a qualitative study 7:31, 2007.	To understand factors considered by Japanese medical students when choosing their specialty	Medical school in Japan	25 semistructured interviews with medical students, informal interviews with academic faculty, field notes; thematic analysis
Schoenborn NL, Lee K, Pollack CE, . Older adults’ preferences for when and how to discuss life expectancy in primary care. 2017;30:813–5.	To elucidate perspectives on how and when to discuss life expectancy with older adults	Four clinical programmes affiliated with an urban academic medical centre	40 semistructured interviews with community-dwelling older adults; qualitative content analysis

From our perspective as seasoned qualitative researchers, conducting effective semistructured interviews requires: (1) a relational focus, including active engagement and curiosity, and (2) practice in the skills of interviewing. First, a relational focus emphasises the unique relationship between interviewer and interviewee. To obtain quality data, interviews should not be conducted with a transactional question-answer approach but rather should be unfolding, iterative interactions between the interviewer and interviewee. Second, interview skills can be learnt. Some of us will naturally be more comfortable and skilful at conducting interviews but all aspects of interviews are learnable and through practice and feedback will improve. Throughout this article, we highlight strategies to balance relationship and rigour when conducting semistructured interviews in primary care and the healthcare setting.

Qualitative research interviews are ‘attempts to understand the world from the subjects’ point of view, to unfold the meaning of peoples’ experiences, to uncover their lived world prior to scientific explanations’ (p 1). 7 Qualitative research interviews unfold as an interviewer asks questions of the interviewee in order to gather subjective information about a particular topic or experience. Though the definitions and purposes of qualitative research interviews vary slightly in the literature, there is common emphasis on the experiences of interviewees and the ways in which the interviewee perceives the world (see table 2 for summary of definitions from seminal texts).

Definitions of qualitative interviews

Authors	Definition	Purpose
DiCicco-Bloom and Crabtree	Semistructured interviews are ‘organized around a set of predetermined open-ended questions, with other questions emerging from the dialogue between interviewer and interviewee/s’ (2006, p 315)	‘To contribute to a body of knowledge that is conceptual and theoretical and is based on the meanings that life experiences hold for the interviewees’ (2006, p 314)
Hatch	‘special kinds of conversations or speech events that are used by researchers to explore informants’ experiences and interpretations’ (2002, p. 91)	‘To uncover the meaning structures that participants use to organize their experiences and make sense of their worlds’ (2002, p 91)
Kvale	‘attempts to understand the world from the subjects' point of view, to unfold the meaning of peoples' experiences, to uncover their lived world prior to scientific explanations’ (1996, p 1)	‘To gather descriptions of the life-world of the interviewee with respect to interpretation of the meaning of the described phenomena’ (1983, p 174)
Josselson	‘a shared product of two people—one the interviewer, the other the interviewee—talk about and they talk together’ (2013, p 1)	‘To enter the world of the participant and try to understand how it looks and feels from the participant’s point of view’ (2013, p 80)

The most common type of interview used in qualitative research and the healthcare context is semistructured interview. 8 Figure 1 highlights the key features of this data collection method, which is guided by a list of topics or questions with follow-up questions, probes and comments. Typically, the sequencing and wording of the questions are modified by the interviewer to best fit the interviewee and interview context. Semistructured interviews can be conducted in multiple ways (ie, face to face, telephone, text/email, individual, group, brief, in-depth), each of which have advantages and disadvantages. We will focus on the most common form of semistructured interviews within qualitative research—individual, face-to-face, in-depth interviews.

Key characteristics of semistructured interviews.

Purpose of semistructured interviews

The overall purpose of using semistructured interviews for data collection is to gather information from key informants who have personal experiences, attitudes, perceptions and beliefs related to the topic of interest. Researchers can use semistructured interviews to collect new, exploratory data related to a research topic, triangulate other data sources or validate findings through member checking (respondent feedback about research results). 9 If using a mixed methods approach, semistructured interviews can also be used in a qualitative phase to explore new concepts to generate hypotheses or explain results from a quantitative phase that tests hypotheses. Semistructured interviews are an effective method for data collection when the researcher wants: (1) to collect qualitative, open-ended data; (2) to explore participant thoughts, feelings and beliefs about a particular topic; and (3) to delve deeply into personal and sometimes sensitive issues.

Designing and conducting semistructured interviews

In the following section, we provide recommendations for the steps required to carefully design and conduct semistructured interviews with emphasis on applications in family medicine and primary care research (see table 3 ).

Steps to designing and conducting semistructured interviews

Step	Task
1	Determining the purpose and scope of the study
2	Identifying participants
3	Considering ethical issues
4	Planning logistical aspects
5	Developing the interview guide
6	Establishing trust and rapport
7	Conducting the interview
8	Memoing and reflection
9	Analysing the data
10	Demonstrating the trustworthiness of the research
11	Presenting findings in a paper or report

Steps for designing and conducting semistructured interviews

Step 1: determining the purpose and scope of the study.

The purpose of the study is the primary objective of your project and may be based on an anecdotal experience, a review of the literature or previous research finding. The purpose is developed in response to an identified gap or problem that needs to be addressed.

Research questions are the driving force of a study because they are associated with every other aspect of the design. They should be succinct and clearly indicate that you are using a qualitative approach. Qualitative research questions typically start with ‘What’, ‘How’ or ‘Why’ and focus on the exploration of a single concept based on participant perspectives. 10

Step 2: identifying participants

After deciding on the purpose of the study and research question(s), the next step is to determine who will provide the best information to answer the research question. Good interviewees are those who are available, willing to be interviewed and have lived experiences and knowledge about the topic of interest. 11 12 Working with gatekeepers or informants to get access to potential participants can be extremely helpful as they are trusted sources that control access to the target sample.

Sampling strategies are influenced by the research question and the purpose of the study. Unlike quantitative studies, statistical representativeness is not the goal of qualitative research. There is no calculation of statistical power and the goal is not a large sample size. Instead, qualitative approaches seek an in-depth and detailed understanding and typically use purposeful sampling. See the study of Hatch for a summary of various types of purposeful sampling that can be used for interview studies. 12

‘How many participants are needed?’ The most common answer is, ‘it depends’—it depends on the purpose of the study, what kind of study is planned and what questions the study is trying to answer. 12–14 One common standard in qualitative sample sizes is reaching thematic saturation, which refers to the point at which no new thematic information is gathered from participants. Malterud and colleagues discuss the concept of information power , or a qualitative equivalent to statistical power, to determine how many interviews should be collected in a study. They suggest that the size of a sample should depend on the aim, homogeneity of the sample, theory, interview quality and analytic strategy. 14

Step 3: considering ethical issues

An ethical attitude should be present from the very beginning of the research project even before you decide who to interview. 15 This ethical attitude should incorporate respect, sensitivity and tact towards participants throughout the research process. Because semistructured interviewing often requires the participant to reveal sensitive and personal information directly to the interviewer, it is important to consider the power imbalance between the researcher and the participant. In healthcare settings, the interviewer or researcher may be a part of the patient’s healthcare team or have contact with the healthcare team. The researchers should ensure the interviewee that their participation and answers will not influence the care they receive or their relationship with their providers. Other issues to consider include: reducing the risk of harm; protecting the interviewee’s information; adequately informing interviewees about the study purpose and format; and reducing the risk of exploitation. 10

Step 4: planning logistical aspects

Careful planning particularly around the technical aspects of interviews can be the difference between a great interview and a not so great interview. During the preparation phase, the researcher will need to plan and make decisions about the best ways to contact potential interviewees, obtain informed consent, arrange interview times and locations convenient for both participant and researcher, and test recording equipment. Although many experienced researchers have found themselves conducting interviews in less than ideal locations, the interview location should avoid (or at least minimise) interruptions and be appropriate for the interview (quiet, private and able to get a clear recording). 16 For some research projects, the participants’ homes may make sense as the best interview location. 16

Initial contacts can be made through telephone or email and followed up with more details so the individual can make an informed decision about whether they wish to be interviewed. Potential participants should know what to expect in terms of length of time, purpose of the study, why they have been selected and who will be there. In addition, participants should be informed that they can refuse to answer questions or can withdraw from the study at any time, including during the interview itself.

Audio recording the interview is recommended so that the interviewer can concentrate on the interview and build rapport rather than being distracted with extensive note taking 16 (see table 4 for audio-recording tips). Participants should be informed that audio recording is used for data collection and that they can refuse to be audio recorded should they prefer.

Suggestions for successful audio recording of interviews

Component	Suggestions
Clarity	Audio-recording equipment should clearly capture the interview so that both interviewer’s and interviewee’s voices are easily heard for transcription. Many interviewers use small battery-powered recorders but sometimes the microphones do not work well.
Reliable	Audio-recording equipment needs to be reliable and easy to use. Increasingly, researchers are using their smartphones to record interviews.
Familiarity	Whatever kind of recording equipment is used, the researcher needs to be familiar with it and should test it at the interview location before starting the actual interview—you do not want to be fumbling with technology during the interview.
Backup	If you are the sole interviewer and do not have an additional person taking notes, we recommend having two recording devices for each interview in case one device fails or runs out of batteries. Make sure to bring extra batteries.
Note-taking	Some researchers recommend taking notes or having a partner take notes during the interviews in addition to the audio recording. Taking notes can ensure that all interview questions have been answered, guide follow-up questions so that the interview can flow from the interviewee’s lead and serve as a backup in the case of malfunctioning recorders.

Most researchers will want to have interviews transcribed verbatim from the audio recording. This allows you to refer to the exact words of participants during the analysis. Although it is possible to conduct analyses from the audio recordings themselves or from notes, it is not ideal. However, transcription can be extremely time consuming and, if not done yourself, can be costly.

In the planning phase of research, you will want to consider whether qualitative research software (eg, NVivo, ATLAS.ti, MAXQDA, Dedoose, and so on) will be used to assist with organising, managing and analysis. While these tools are helpful in the management of qualitative data, it is important to consider your research budget, the cost of the software and the learning curve associated with using a new system.

Step 5: developing the interview guide

Semistructured interviews include a short list of ‘guiding’ questions that are supplemented by follow-up and probing questions that are dependent on the interviewee’s responses. 8 17 All questions should be open ended, neutral, clear and avoid leading language. In addition, questions should use familiar language and avoid jargon.

Most interviews will start with an easy, context-setting question before moving to more difficult or in-depth questions. 17 Table 5 gives details of the types of guiding questions including ‘grand tour’ questions, 18 core questions and planned and unplanned follow-up questions.

Questions and prompts in semistructured interviewing

Type of question	Definition	Purpose	Example
Grand tour	General question related to the content of the overall research question, which participant knows a lot about
Core questions	Five to 10 questions that directly relate to the information the researcher wants to know
Planned follow-up questions	Specific questions that ask for more details about particular aspects of the core questions
Unplanned follow-up questions	Questions that arise during the interview based on participant responses

To illustrate, online supplementary appendix A presents a sample interview guide from our study of weight gain during pregnancy among young women. We start with the prompt, ‘Tell me about how your pregnancy has been so far’ to initiate conversation about their thoughts and feelings during pregnancy. The subsequent questions will elicit responses to help answer our research question about young women’s perspectives related to weight gain during pregnancy.

Supplementary data

After developing the guiding questions, it is important to pilot test the interview. Having a good sense of the guide helps you to pace the interview (and not run out of time), use a conversational tone and make necessary adjustments to the questions.

Like all qualitative research, interviewing is iterative in nature—data collection and analysis occur simultaneously, which may result in changes to the guiding questions as the study progresses. Questions that are not effective may be replaced with other questions and additional probes can be added to explore new topics that are introduced by participants in previous interviews. 10

Step 6: establishing trust and rapport

Interviews are a special form of relationship, where the interviewer and interviewee converse about important and often personal topics. The interviewer must build rapport quickly by listening attentively and respectfully to the information shared by the interviewee. 19 As the interview progresses, the interviewer must continue to demonstrate respect, encourage the interviewee to share their perspectives and acknowledge the sensitive nature of the conversation. 20

To establish rapport, it is important to be authentic and open to the interviewee’s point of view. It is possible that the participants you recruit for your study will have preconceived notions about research, which may include mistrust. As a result, it is important to describe why you are conducting the research and how their participation is meaningful. In an interview relationship, the interviewee is the expert and should be treated as such—you are relying on the interviewee to enhance your understanding and add to your research. Small behaviours that can enhance rapport include: dressing professionally but not overly formal; avoiding jargon or slang; and using a normal conversational tone. Because interviewees will be discussing their experience, having some awareness of contextual or cultural factors that may influence their perspectives may be helpful as background knowledge.

Step 7: conducting the interview

Location and set-up.

The interview should have already been scheduled at a convenient time and location for the interviewee. The location should be private, ideally with a closed door, rather than a public place. It is helpful if there is a room where you can speak privately without interruption, and where it is quiet enough to hear and audio record the interview. Within the interview space, Josselson 15 suggests an arrangement with a comfortable distance between the interviewer and interviewee with a low table in between for the recorder and any materials (consent forms, questionnaires, water, and so on).

Beginning the interview

Many interviewers start with chatting to break the ice and attempt to establish commonalities, rapport and trust. Most interviews will need to begin with a brief explanation of the research study, consent/assent procedures, rationale for talking to that particular interviewee and description of the interview format and agenda. 11 It can also be helpful if the interviewer shares a little about who they are and why they are interested in the topic. The recording equipment should have already been tested thoroughly but interviewers may want to double-check that the audio equipment is working and remind participants about the reason for recording.

Interviewer stance

During the interview, the interviewer should adopt a friendly and non-judgemental attitude. You will want to maintain a warm and conversational tone, rather than a rote, question-answer approach. It is important to recognise the potential power differential as a researcher. Conveying a sense of being in the interview together and that you as the interviewer are a person just like the interviewee can help ease any discomfort. 15

Active listening

During a face-to-face interview, there is an opportunity to observe social and non-verbal cues of the interviewee. These cues may come in the form of voice, body language, gestures and intonation, and can supplement the interviewee’s verbal response and can give clues to the interviewer about the process of the interview. 21 Listening is the key to successful interviewing. 22 Listening should be ‘attentive, empathic, nonjudgmental, listening in order to invite, and engender talk’ 15 15 (p 66). Silence, nods, smiles and utterances can also encourage further elaboration from the interviewee.

Continuing the interview

As the interview progresses, the interviewer can repeat the words used by the interviewee, use planned and unplanned follow-up questions that invite further clarification, exploration or elaboration. As DiCicco-Bloom and Crabtree 10 explain: ‘Throughout the interview, the goal of the interviewer is to encourage the interviewee to share as much information as possible, unselfconsciously and in his or her own words’ (p 317). Some interviewees are more forthcoming and will offer many details of their experiences without much probing required. Others will require prompting and follow-up to elicit sufficient detail.

As a result, follow-up questions are equally important to the core questions in a semistructured interview. Prompts encourage people to continue talking and they can elicit more details needed to understand the topic. Examples of verbal probes are repeating the participant’s words, summarising the main idea or expressing interest with verbal agreement. 8 11 See table 6 for probing techniques and example probes we have used in our own interviewing.

Probing techniques for semistructured interviews (modified from Bernard 30 )

Probing technique	Description	Example
Wait time	Interviewer remains silent after asking a question. This allows the interviewee to think about their response and often encourages the interviewee to speak.	(Wait, do not respond with additional questioning until participant speaks)
Echo	Interviewer repeats or summarises the participant’s words, encouraging them to go into more detail.	.
Verbal agreement	Interviewer uses affirming words to encourage the interviewee to continue speaking.
Expansion	Interviewer asks participant to elaborate on a particular response.	. .
Explanation	Interviewer asks participant to clarify a specific comment.
Leading	Interviewer asks interviewee to explain their reasoning.	.

Step 8: memoing and reflection

After an interview, it is essential for the interviewer to begin to reflect on both the process and the content of the interview. During the actual interview, it can be difficult to take notes or begin reflecting. Even if you think you will remember a particular moment, you likely will not be able to recall each moment with sufficient detail. Therefore, interviewers should always record memos —notes about what you are learning from the data. 23 24 There are different approaches to recording memos: you can reflect on several specific ideas, or create a running list of thoughts. Memos are also useful for improving the quality of subsequent interviews.

Step 9: analysing the data

The data analysis strategy should also be developed during planning stages because analysis occurs concurrently with data collection. 25 The researcher will take notes, modify the data collection procedures and write reflective memos throughout the data collection process. This begins the process of data analysis.

The data analysis strategy used in your study will depend on your research question and qualitative design—see the study of Creswell for an overview of major qualitative approaches. 26 The general process for analysing and interpreting most interviews involves reviewing the data (in the form of transcripts, audio recordings or detailed notes), applying descriptive codes to the data and condensing and categorising codes to look for patterns. 24 27 These patterns can exist within a single interview or across multiple interviews depending on the research question and design. Qualitative computer software programs can be used to help organise and manage interview data.

Step 10: demonstrating the trustworthiness of the research

Similar to validity and reliability, qualitative research can be assessed on trustworthiness. 9 28 There are several criteria used to establish trustworthiness: credibility (whether the findings accurately and fairly represent the data), transferability (whether the findings can be applied to other settings and contexts), confirmability (whether the findings are biased by the researcher) and dependability (whether the findings are consistent and sustainable over time).

Step 11: presenting findings in a paper or report

When presenting the results of interview analysis, researchers will often report themes or narratives that describe the broad range of experiences evidenced in the data. This involves providing an in-depth description of participant perspectives and being sure to include multiple perspectives. 12 In interview research, the participant words are your data. Presenting findings in a report requires the integration of quotes into a more traditional written format.

Conclusions

Though semistructured interviews are often an effective way to collect open-ended data, there are some disadvantages as well. One common problem with interviewing is that not all interviewees make great participants. 12 29 Some individuals are hard to engage in conversation or may be reluctant to share about sensitive or personal topics. Difficulty interviewing some participants can affect experienced and novice interviewers. Some common problems include not doing a good job of probing or asking for follow-up questions, failure to actively listen, not having a well-developed interview guide with open-ended questions and asking questions in an insensitive way. Outside of pitfalls during the actual interview, other problems with semistructured interviewing may be underestimating the resources required to recruit participants, interview, transcribe and analyse the data.

Despite their limitations, semistructured interviews can be a productive way to collect open-ended data from participants. In our research, we have interviewed children and adolescents about their stress experiences and coping behaviours, young women about their thoughts and behaviours during pregnancy, practitioners about the care they provide to patients and countless other key informants about health-related topics. Because the intent is to understand participant experiences, the possible research topics are endless.

Due to the close relationships family physicians have with their patients, the unique settings in which they work, and in their advocacy, semistructured interviews are an attractive approach for family medicine researchers, even if working in a setting with limited research resources. When seeking to balance both the relational focus of interviewing and the necessary rigour of research, we recommend: prioritising listening over talking; using clear language and avoiding jargon; and deeply engaging in the interview process by actively listening, expressing empathy, demonstrating openness to the participant’s worldview and thanking the participant for helping you to understand their experience.

Further Reading

• Edwards R, & Holland J. (2013). What is qualitative interviewing?: A&C Black.
• Josselson R. Interviewing for qualitative inquiry: A relational approach. Guilford Press, 2013.
• Kvale S. InterViews: An Introduction to Qualitative Research Interviewing. SAGE, London, 1996.
• Pope C, & Mays N. (Eds). (2006). Qualitative research in health care.

Correction notice: This article has been corrected. Reference details have been updated.

Contributors: Both authors contributed equally to this work.

Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests: None declared.

Patient consent for publication: Not required.

Provenance and peer review: Not commissioned; internally peer reviewed.

Mastering Semi-Structured Interviews

Introduction

What is the difference between structured and semi-structured interviews, when to use a semi-structured interview, best practices for conducting semi-structured interviews, semi-structured interviews in qualitative research.

Interviews are an integral part of qualitative and social science research . While observational research explores what people do, interviews look at what people say and believe. The interview is an important research method to capture people's perspectives and experiences concerning relevant topics.

Three different types of interviews can be utilized in research. In this article, we will look at the semi-structured interview. This form of interview offers a balance between a rigid interaction that produces neatly organized data and a fluid conversation that can explore unexpected but relevant aspects of the phenomenon under study.

Among research methods , interviewing focuses on the experiences and perspectives that people have about a particular topic. In contrast, other research methods such as experiments and observations focus on what people do or how things work. However, people may look at the same cultural or social practice and think different things about it, making interviews important to capture potential nuances in experiences and interpretations.

Conducting an interview is a more complex task than simply talking with people. Qualitative researchers can adopt three different approaches to talking with interview respondents. The most straightforward form of interview is the structured interview , which is a rigid form of interview that asks a specific set of questions. It is fully structured in that all questions are specified beforehand and the interviewer poses the same questions to all participants, without any variations or asking any follow-up questions on the spot. A strength of structured interviews is that asking only predetermined questions produces uniform data that makes comparisons across participants easier, as answers from structured interviews can be quickly sorted into a matrix or spreadsheet for simple comparison.

Another type of interview is the semi-structured interview, which also has predetermined questions but allows for follow-up questions for deeper exploration. In this case, the interview can be seen as a formal conversation, with the researcher having a predetermined set of topics and questions they want to ask, while at the same time remaining open to asking other questions as the conversation unfolds. As a result, a semi-structured interview offers the necessary flexibility for researchers to explore any relevant ideas that may emerge as the participant answers questions and shares new information.

Advantages of semi-structured interviews

Semi-structured interviews allow the researcher to probe deeply into the perspectives of interview respondents, while structured interviews have a rigid format that does not allow for the interviewer to elicit more detail if given the opportunity.

The semi-structured format also provides the necessary guidance for researchers to stay focused on the key topics at hand. While the interview may go through the questions in a different order or explore additional topics, the predetermined questions in a semi-structured interview ensures that the important topics are sufficiently explored.

Disadvantages of semi-structured interviews

Unlike in a formal interview, the open-ended nature of semi-structured interviews can allow for the interview respondent to take the conversation in unanticipated directions. While this is a useful feature of semi-structured interviews, it is also important for the interviewer to guide the conversation toward the topic of study to ensure that the collected data will be relevant to the research question.

A semi-structured interview also requires the interviewer to engage in active listening to be able to take advantage of opportunities to ask probing questions. In this respect, interviewers may require training to ensure that they can effectively conduct a semi-structured interview that explores respondents' perspectives deeply enough while collecting data relevant to the research inquiry.

Unstructured interviews

One more distinction to keep in mind is that of the unstructured interview . While structured and semi-structured interviews have predetermined questions tailored to address the research question , unstructured interviews have no framework set before conducting the interview.

These kinds of interviews are meant to be more informal or exploratory in nature; they allow respondents to answer as freely as possible and permits the interviewer to follow the dialogue wherever it goes. While both semi-structured and unstructured interviews can employ spontaneous follow-up questions, semi-structured interviews are designed to ensure that a set of key questions are asked to all respondents to ensure relevant data is collected.

While interviews can follow predetermined structures to different degrees, interviewing as a data collection method is a social act that involves developing rapport with the interview respondent so that they feel comfortable to answer freely. This is also important to collect rich data that shed light on the phenomenon under study.

Keep in mind that any qualitative interview, regardless of type, focuses on open-ended questions. Any study that is more suited to closed-ended questions may find survey research more conducive to addressing their research inquiry.

Organize your interview data with ATLAS.ti

Powerful tools to manage and analyze research projects of all sizes. Start with a free trial today.

A semi-structured interview is ideal when you want to explore individuals' experiences and perspectives around a particular topic. It is also important to have a clearly defined research agenda with specific objectives that your interview respondents can address. Your research objectives can inform the core questions you can pose to respondents.

In addition, if you are still looking to inductively generate theory in areas that have little theoretical coherence or conceptualization, a semi-structured interview is ideal because it allows you to probe further into the ideas that emerge from your respondents. Semi-structured interviews are thus powerful data collection tools when you are looking to build a theory or explore individuals' experiences or perspectives.

Interviewing in qualitative research is not merely an act of conversing with research participants. It is a research method aimed at exploring the perspectives and ideas of research participants as deeply as possible.

When you conduct semi-structured interviews, it is important to intentionally consider every major element of the study, from the selection of participants to the questions asked, even the setting in which the interviews take place.

Preparing for a semi-structured interview

Think about which participants can adequately address the objectives in your study. For example, if your research inquiry deals with a specific cultural practice from a particular perspective, then you will benefit from choosing respondents who can best speak to that perspective.

Also reflect on how you will interact with your respondents. What is the best way to reach them and elicit their ideas? To engage in a meaningful conversation with your participants, it is important to pose questions in a way that is easy for others to understand, avoiding any jargon and preparing alternative ways to ask each question. Moreover, interview questions should be adjusted to each participant. Interviewing children is a different matter from interviewing adults. If the respondents' first language is different from yours, you may also want to consider adjusting your language to make yourself understood. The respondent's individual circumstances will play an important role in how you conduct your interview.

In addition, consider what equipment you will use to collect qualitative data in the form of audio or video recordings , and aim to record in as high a quality as possible. While the audio recorder on most smartphones is adequate enough to capture most conversations, you may want to think about using professional equipment if you are conducting interviews in public environments like cafés or parks. A camera may also be appropriate if you want to record facial expressions, gestures, and other body language for later analysis.

Semi-structured interview questions

A researcher should prepare an interview guide that lists all the necessary questions to be asked and topics to be explored. The guide can be flexible and researchers can ask the questions in whichever order naturally unfolds during the conversation. Nonetheless, having a guide helps ensure that the researcher is collecting data relevant to the research question.

When designing interview guides, consider how your questions are framed and how they might be received by the interview respondent. Avoid leading questions that may elicit socially desirable responses, and prepare alternative ways to word your questions in case participants don't understand a question.

Preparing follow-up questions

Probing questions make for effective follow-ups that encourage respondents to provide in-depth information about the topic at hand. A common challenge of interviews is that participants may provide very brief responses or not deeply engage with the conversation. Preparing prompts and probes can help researchers encourage participants to open up or provide more details if needed.

In general, an interviewer should invite the respondent to elaborate on answers when additional details can benefit the research. Taking advantage of such opportunities in a semi-structured interview can greatly contribute to the theoretical development arising from the interview study. These prompts and probes can be as simple as asking for more details, nodding along, or practicing silence. Another helpful tactic is to ask participants to provide an example or walk you through the story they are sharing.

The interview itself is just one of the components of the interview study. During and after the semi-structured interview, take the following into consideration to ensure rigorous data collection .

Collecting qualitative data in the form of interviews

In most cases, interview data takes the form of transcriptions of raw audio or video recordings of the interview conversations. It's important to ensure that you have the necessary equipment to record and transcribe the interview. Being able to count on high quality recordings is crucial to make transcription easier and more accurate.

While you can certainly analyze the actual recordings, textual data can make the analysis process easier and more manageable. You can use qualitative data analysis software such as ATLAS.ti to analyze multimedia or text data; another benefit of text data is that many additional analysis tools can be used to analyze the structure or contents of the data.

An interview researcher should also consider how the interview is conducted. After all, the two-way communication in a face-to-face interview has different effects on the interview respondent from an interview that is conducted online or by email. Be sure to familiarize yourself with the environment in which you will conduct the interview so that you can anticipate any issues that arise regarding clarity between you and your respondents.

During the course of any interview, it may benefit your analysis to capture detailed notes about the interactions you have with your respondents. A good practice is to note down any observations or impressions immediately after concluding each interview while the interaction is fresh in your mind. Many interviewers use these notes to remind them of potentially significant theoretical developments that can be used when coding the data.

Interviews with vulnerable populations

For interview projects that involve sensitive issues, the researcher should be mindful of how questions are posed and what is asked to avoid interview respondents becoming uncomfortable or anxious.

This is especially true in studies that involve children, people in conflict zones, and other vulnerable populations. The interviewer should take great care to balance data collection with the responsibility of protecting the well-being of their research participants.

Informed consent with interview respondents

In terms of addressing ethical considerations , the researcher should also ensure that they receive participants' consent before collecting any data . Informed consent is a crucial standard in research involving human participants, and it involves both the interviewer and interview respondent being cognizant of the purpose of the study, the procedures taken during the interview, and the measures in place to preserve the respondent's privacy and personal data .

Especially with respect to interviews that collect open-ended data from participants, researchers should ensure that respondents have an in-depth understanding of the interview study in which they are participating.

Preparing semi-structured interviews for analysis

Unlike interviews for news outlets or entertainment programs, the interview research process doesn't end at the conclusion of the conversation with the participant. A research paper is not simply a reporting of what was said in an interview or set of interviews. Instead, the respondents' utterances should be carefully and rigorously analyzed to determine what themes and patterns arise from the data and how these relate to the research question guiding the study.

Transcription of interview recordings is a standard practice for analyzing interviews. You can either manually transcribe interviews, outsource transcription to a professional service, or use software that automates the process. Whatever you choose, make sure that the transcription is accurate and has the level of detail (e.g., thinking sounds, pauses) that you are looking for in your analysis.

Coding and analyzing semi-structured interviews

Qualitative data typically undergoes a coding process in which data segments are labeled with descriptive codes. These codes help to identify patterns in the data. Ultimately, the goal of coding is to help the researcher condense and organize the data to address their research objectives.

For semi-structured interviews, consider first coding every answer based on the questions in the interview guide. This will allow you to compare respondents' answers to the same interview questions when viewing and analyzing each question code later on. You can supplement these codes with interpretive codes based on emerging themes to further explore patterns across participants' experiences or perspectives.

Make the most of your interview research with ATLAS.ti

Easily organize and analyze interview research projects with our intuitive interface. Start with a free trial today.

• Search Menu

Sign in through your institution

• Browse content in Arts and Humanities
• Browse content in Archaeology
• Anglo-Saxon and Medieval Archaeology
• Archaeological Methodology and Techniques
• Archaeology by Region
• Archaeology of Religion
• Archaeology of Trade and Exchange
• Biblical Archaeology
• Contemporary and Public Archaeology
• Environmental Archaeology
• Historical Archaeology
• History and Theory of Archaeology
• Industrial Archaeology
• Landscape Archaeology
• Mortuary Archaeology
• Prehistoric Archaeology
• Underwater Archaeology
• Zooarchaeology
• Browse content in Architecture
• Architectural Structure and Design
• History of Architecture
• Residential and Domestic Buildings
• Theory of Architecture
• Browse content in Art
• Art Subjects and Themes
• History of Art
• Industrial and Commercial Art
• Theory of Art
• Biographical Studies
• Byzantine Studies
• Browse content in Classical Studies
• Classical History
• Classical Philosophy
• Classical Mythology
• Classical Numismatics
• Classical Literature
• Classical Reception
• Classical Art and Architecture
• Classical Oratory and Rhetoric
• Greek and Roman Papyrology
• Greek and Roman Epigraphy
• Greek and Roman Law
• Greek and Roman Archaeology
• Late Antiquity
• Religion in the Ancient World
• Social History
• Digital Humanities
• Browse content in History
• Colonialism and Imperialism
• Diplomatic History
• Environmental History
• Genealogy, Heraldry, Names, and Honours
• Genocide and Ethnic Cleansing
• Historical Geography
• History by Period
• History of Emotions
• History of Agriculture
• History of Education
• History of Gender and Sexuality
• Industrial History
• Intellectual History
• International History
• Labour History
• Legal and Constitutional History
• Local and Family History
• Maritime History
• Military History
• National Liberation and Post-Colonialism
• Oral History
• Political History
• Public History
• Regional and National History
• Revolutions and Rebellions
• Slavery and Abolition of Slavery
• Social and Cultural History
• Theory, Methods, and Historiography
• Urban History
• World History
• Browse content in Language Teaching and Learning
• Language Learning (Specific Skills)
• Language Teaching Theory and Methods
• Browse content in Linguistics
• Applied Linguistics
• Cognitive Linguistics
• Computational Linguistics
• Forensic Linguistics
• Grammar, Syntax and Morphology
• Historical and Diachronic Linguistics
• History of English
• Language Evolution
• Language Reference
• Language Acquisition
• Language Variation
• Language Families
• Lexicography
• Linguistic Anthropology
• Linguistic Theories
• Linguistic Typology
• Phonetics and Phonology
• Psycholinguistics
• Sociolinguistics
• Translation and Interpretation
• Writing Systems
• Browse content in Literature
• Bibliography
• Children's Literature Studies
• Literary Studies (Romanticism)
• Literary Studies (American)
• Literary Studies (Asian)
• Literary Studies (European)
• Literary Studies (Eco-criticism)
• Literary Studies (Modernism)
• Literary Studies - World
• Literary Studies (1500 to 1800)
• Literary Studies (19th Century)
• Literary Studies (20th Century onwards)
• Literary Studies (African American Literature)
• Literary Studies (British and Irish)
• Literary Studies (Early and Medieval)
• Literary Studies (Fiction, Novelists, and Prose Writers)
• Literary Studies (Gender Studies)
• Literary Studies (Graphic Novels)
• Literary Studies (History of the Book)
• Literary Studies (Plays and Playwrights)
• Literary Studies (Poetry and Poets)
• Literary Studies (Postcolonial Literature)
• Literary Studies (Queer Studies)
• Literary Studies (Science Fiction)
• Literary Studies (Travel Literature)
• Literary Studies (War Literature)
• Literary Studies (Women's Writing)
• Literary Theory and Cultural Studies
• Mythology and Folklore
• Shakespeare Studies and Criticism
• Browse content in Media Studies
• Browse content in Music
• Applied Music
• Dance and Music
• Ethics in Music
• Ethnomusicology
• Gender and Sexuality in Music
• Medicine and Music
• Music Cultures
• Music and Media
• Music and Religion
• Music and Culture
• Music Education and Pedagogy
• Music Theory and Analysis
• Musical Scores, Lyrics, and Libretti
• Musical Structures, Styles, and Techniques
• Musicology and Music History
• Performance Practice and Studies
• Race and Ethnicity in Music
• Sound Studies
• Browse content in Performing Arts
• Browse content in Philosophy
• Aesthetics and Philosophy of Art
• Epistemology
• Feminist Philosophy
• History of Western Philosophy
• Metaphysics
• Moral Philosophy
• Non-Western Philosophy
• Philosophy of Language
• Philosophy of Mind
• Philosophy of Perception
• Philosophy of Science
• Philosophy of Action
• Philosophy of Law
• Philosophy of Religion
• Philosophy of Mathematics and Logic
• Practical Ethics
• Social and Political Philosophy
• Browse content in Religion
• Biblical Studies
• Christianity
• East Asian Religions
• History of Religion
• Judaism and Jewish Studies
• Qumran Studies
• Religion and Education
• Religion and Health
• Religion and Politics
• Religion and Science
• Religion and Law
• Religion and Art, Literature, and Music
• Religious Studies
• Browse content in Society and Culture
• Cookery, Food, and Drink
• Cultural Studies
• Customs and Traditions
• Ethical Issues and Debates
• Hobbies, Games, Arts and Crafts
• Natural world, Country Life, and Pets
• Popular Beliefs and Controversial Knowledge
• Sports and Outdoor Recreation
• Technology and Society
• Travel and Holiday
• Visual Culture
• Browse content in Law
• Arbitration
• Browse content in Company and Commercial Law
• Commercial Law
• Company Law
• Browse content in Comparative Law
• Systems of Law
• Competition Law
• Browse content in Constitutional and Administrative Law
• Government Powers
• Judicial Review
• Local Government Law
• Military and Defence Law
• Parliamentary and Legislative Practice
• Construction Law
• Contract Law
• Browse content in Criminal Law
• Criminal Procedure
• Criminal Evidence Law
• Sentencing and Punishment
• Employment and Labour Law
• Environment and Energy Law
• Browse content in Financial Law
• Banking Law
• Insolvency Law
• History of Law
• Human Rights and Immigration
• Intellectual Property Law
• Browse content in International Law
• Private International Law and Conflict of Laws
• Public International Law
• IT and Communications Law
• Jurisprudence and Philosophy of Law
• Law and Politics
• Law and Society
• Browse content in Legal System and Practice
• Courts and Procedure
• Legal Skills and Practice
• Legal System - Costs and Funding
• Primary Sources of Law
• Regulation of Legal Profession
• Medical and Healthcare Law
• Browse content in Policing
• Criminal Investigation and Detection
• Police and Security Services
• Police Procedure and Law
• Police Regional Planning
• Browse content in Property Law
• Personal Property Law
• Restitution
• Study and Revision
• Terrorism and National Security Law
• Browse content in Trusts Law
• Wills and Probate or Succession
• Browse content in Medicine and Health
• Browse content in Allied Health Professions
• Arts Therapies
• Clinical Science
• Dietetics and Nutrition
• Occupational Therapy
• Operating Department Practice
• Physiotherapy
• Radiography
• Speech and Language Therapy
• Browse content in Anaesthetics
• General Anaesthesia
• Clinical Neuroscience
• Browse content in Clinical Medicine
• Acute Medicine
• Cardiovascular Medicine
• Clinical Genetics
• Clinical Pharmacology and Therapeutics
• Dermatology
• Endocrinology and Diabetes
• Gastroenterology
• Genito-urinary Medicine
• Geriatric Medicine
• Infectious Diseases
• Medical Toxicology
• Medical Oncology
• Pain Medicine
• Palliative Medicine
• Rehabilitation Medicine
• Respiratory Medicine and Pulmonology
• Rheumatology
• Sleep Medicine
• Sports and Exercise Medicine
• Community Medical Services
• Critical Care
• Emergency Medicine
• Forensic Medicine
• Haematology
• History of Medicine
• Browse content in Medical Skills
• Clinical Skills
• Communication Skills
• Nursing Skills
• Surgical Skills
• Browse content in Medical Dentistry
• Oral and Maxillofacial Surgery
• Paediatric Dentistry
• Restorative Dentistry and Orthodontics
• Surgical Dentistry
• Medical Ethics
• Medical Statistics and Methodology
• Browse content in Neurology
• Clinical Neurophysiology
• Neuropathology
• Nursing Studies
• Browse content in Obstetrics and Gynaecology
• Gynaecology
• Occupational Medicine
• Ophthalmology
• Otolaryngology (ENT)
• Browse content in Paediatrics
• Neonatology
• Browse content in Pathology
• Chemical Pathology
• Clinical Cytogenetics and Molecular Genetics
• Histopathology
• Medical Microbiology and Virology
• Patient Education and Information
• Browse content in Pharmacology
• Psychopharmacology
• Browse content in Popular Health
• Caring for Others
• Complementary and Alternative Medicine
• Self-help and Personal Development
• Browse content in Preclinical Medicine
• Cell Biology
• Molecular Biology and Genetics
• Reproduction, Growth and Development
• Primary Care
• Professional Development in Medicine
• Browse content in Psychiatry
• Addiction Medicine
• Child and Adolescent Psychiatry
• Forensic Psychiatry
• Learning Disabilities
• Old Age Psychiatry
• Psychotherapy
• Browse content in Public Health and Epidemiology
• Epidemiology
• Public Health
• Browse content in Radiology
• Clinical Radiology
• Interventional Radiology
• Nuclear Medicine
• Radiation Oncology
• Reproductive Medicine
• Browse content in Surgery
• Cardiothoracic Surgery
• Gastro-intestinal and Colorectal Surgery
• General Surgery
• Neurosurgery
• Paediatric Surgery
• Peri-operative Care
• Plastic and Reconstructive Surgery
• Surgical Oncology
• Transplant Surgery
• Trauma and Orthopaedic Surgery
• Vascular Surgery
• Browse content in Science and Mathematics
• Browse content in Biological Sciences
• Aquatic Biology
• Biochemistry
• Bioinformatics and Computational Biology
• Developmental Biology
• Ecology and Conservation
• Evolutionary Biology
• Genetics and Genomics
• Microbiology
• Molecular and Cell Biology
• Natural History
• Plant Sciences and Forestry
• Research Methods in Life Sciences
• Structural Biology
• Systems Biology
• Zoology and Animal Sciences
• Browse content in Chemistry
• Analytical Chemistry
• Computational Chemistry
• Crystallography
• Environmental Chemistry
• Industrial Chemistry
• Inorganic Chemistry
• Materials Chemistry
• Medicinal Chemistry
• Mineralogy and Gems
• Organic Chemistry
• Physical Chemistry
• Polymer Chemistry
• Study and Communication Skills in Chemistry
• Theoretical Chemistry
• Browse content in Computer Science
• Artificial Intelligence
• Computer Architecture and Logic Design
• Game Studies
• Human-Computer Interaction
• Mathematical Theory of Computation
• Programming Languages
• Software Engineering
• Systems Analysis and Design
• Virtual Reality
• Browse content in Computing
• Business Applications
• Computer Security
• Computer Games
• Computer Networking and Communications
• Digital Lifestyle
• Graphical and Digital Media Applications
• Operating Systems
• Browse content in Earth Sciences and Geography
• Atmospheric Sciences
• Environmental Geography
• Geology and the Lithosphere
• Maps and Map-making
• Meteorology and Climatology
• Oceanography and Hydrology
• Palaeontology
• Physical Geography and Topography
• Regional Geography
• Soil Science
• Urban Geography
• Browse content in Engineering and Technology
• Agriculture and Farming
• Biological Engineering
• Civil Engineering, Surveying, and Building
• Electronics and Communications Engineering
• Energy Technology
• Engineering (General)
• Environmental Science, Engineering, and Technology
• History of Engineering and Technology
• Mechanical Engineering and Materials
• Technology of Industrial Chemistry
• Transport Technology and Trades
• Browse content in Environmental Science
• Applied Ecology (Environmental Science)
• Conservation of the Environment (Environmental Science)
• Environmental Sustainability
• Environmentalist Thought and Ideology (Environmental Science)
• Management of Land and Natural Resources (Environmental Science)
• Natural Disasters (Environmental Science)
• Nuclear Issues (Environmental Science)
• Pollution and Threats to the Environment (Environmental Science)
• Social Impact of Environmental Issues (Environmental Science)
• History of Science and Technology
• Browse content in Materials Science
• Ceramics and Glasses
• Composite Materials
• Metals, Alloying, and Corrosion
• Nanotechnology
• Browse content in Mathematics
• Applied Mathematics
• Biomathematics and Statistics
• History of Mathematics
• Mathematical Education
• Mathematical Finance
• Mathematical Analysis
• Numerical and Computational Mathematics
• Probability and Statistics
• Pure Mathematics
• Browse content in Neuroscience
• Cognition and Behavioural Neuroscience
• Development of the Nervous System
• Disorders of the Nervous System
• History of Neuroscience
• Invertebrate Neurobiology
• Molecular and Cellular Systems
• Neuroendocrinology and Autonomic Nervous System
• Neuroscientific Techniques
• Sensory and Motor Systems
• Browse content in Physics
• Astronomy and Astrophysics
• Atomic, Molecular, and Optical Physics
• Biological and Medical Physics
• Classical Mechanics
• Computational Physics
• Condensed Matter Physics
• Electromagnetism, Optics, and Acoustics
• History of Physics
• Mathematical and Statistical Physics
• Measurement Science
• Nuclear Physics
• Particles and Fields
• Plasma Physics
• Quantum Physics
• Relativity and Gravitation
• Semiconductor and Mesoscopic Physics
• Browse content in Psychology
• Affective Sciences
• Clinical Psychology
• Cognitive Psychology
• Cognitive Neuroscience
• Criminal and Forensic Psychology
• Developmental Psychology
• Educational Psychology
• Evolutionary Psychology
• Health Psychology
• History and Systems in Psychology
• Music Psychology
• Neuropsychology
• Organizational Psychology
• Psychological Assessment and Testing
• Psychology of Human-Technology Interaction
• Psychology Professional Development and Training
• Research Methods in Psychology
• Social Psychology
• Browse content in Social Sciences
• Browse content in Anthropology
• Anthropology of Religion
• Human Evolution
• Medical Anthropology
• Physical Anthropology
• Regional Anthropology
• Social and Cultural Anthropology
• Theory and Practice of Anthropology
• Browse content in Business and Management
• Business Ethics
• Business Strategy
• Business History
• Business and Technology
• Business and Government
• Business and the Environment
• Comparative Management
• Corporate Governance
• Corporate Social Responsibility
• Entrepreneurship
• Health Management
• Human Resource Management
• Industrial and Employment Relations
• Industry Studies
• Information and Communication Technologies
• International Business
• Knowledge Management
• Management and Management Techniques
• Operations Management
• Organizational Theory and Behaviour
• Pensions and Pension Management
• Public and Nonprofit Management
• Social Issues in Business and Management
• Strategic Management
• Supply Chain Management
• Browse content in Criminology and Criminal Justice
• Criminal Justice
• Criminology
• Forms of Crime
• International and Comparative Criminology
• Youth Violence and Juvenile Justice
• Development Studies
• Browse content in Economics
• Agricultural, Environmental, and Natural Resource Economics
• Asian Economics
• Behavioural Finance
• Behavioural Economics and Neuroeconomics
• Econometrics and Mathematical Economics
• Economic History
• Economic Systems
• Economic Methodology
• Economic Development and Growth
• Financial Markets
• Financial Institutions and Services
• General Economics and Teaching
• Health, Education, and Welfare
• History of Economic Thought
• International Economics
• Labour and Demographic Economics
• Law and Economics
• Macroeconomics and Monetary Economics
• Microeconomics
• Public Economics
• Urban, Rural, and Regional Economics
• Welfare Economics
• Browse content in Education
• Adult Education and Continuous Learning
• Care and Counselling of Students
• Early Childhood and Elementary Education
• Educational Equipment and Technology
• Educational Strategies and Policy
• Higher and Further Education
• Organization and Management of Education
• Philosophy and Theory of Education
• Schools Studies
• Secondary Education
• Teaching of a Specific Subject
• Teaching of Specific Groups and Special Educational Needs
• Teaching Skills and Techniques
• Browse content in Environment
• Applied Ecology (Social Science)
• Climate Change
• Conservation of the Environment (Social Science)
• Environmentalist Thought and Ideology (Social Science)
• Management of Land and Natural Resources (Social Science)
• Natural Disasters (Environment)
• Pollution and Threats to the Environment (Social Science)
• Social Impact of Environmental Issues (Social Science)
• Sustainability
• Browse content in Human Geography
• Cultural Geography
• Economic Geography
• Political Geography
• Browse content in Interdisciplinary Studies
• Communication Studies
• Museums, Libraries, and Information Sciences
• Browse content in Politics
• African Politics
• Asian Politics
• Chinese Politics
• Comparative Politics
• Conflict Politics
• Elections and Electoral Studies
• Environmental Politics
• Ethnic Politics
• European Union
• Foreign Policy
• Gender and Politics
• Human Rights and Politics
• Indian Politics
• International Relations
• International Organization (Politics)
• Irish Politics
• Latin American Politics
• Middle Eastern Politics
• Political Behaviour
• Political Economy
• Political Institutions
• Political Methodology
• Political Communication
• Political Philosophy
• Political Sociology
• Political Theory
• Politics and Law
• Politics of Development
• Public Policy
• Public Administration
• Qualitative Political Methodology
• Quantitative Political Methodology
• Regional Political Studies
• Russian Politics
• Security Studies
• State and Local Government
• UK Politics
• US Politics
• Browse content in Regional and Area Studies
• African Studies
• Asian Studies
• East Asian Studies
• Japanese Studies
• Latin American Studies
• Middle Eastern Studies
• Native American Studies
• Scottish Studies
• Browse content in Research and Information
• Research Methods
• Browse content in Social Work
• Addictions and Substance Misuse
• Adoption and Fostering
• Care of the Elderly
• Child and Adolescent Social Work
• Couple and Family Social Work
• Direct Practice and Clinical Social Work
• Emergency Services
• Human Behaviour and the Social Environment
• International and Global Issues in Social Work
• Mental and Behavioural Health
• Social Justice and Human Rights
• Social Policy and Advocacy
• Social Work and Crime and Justice
• Social Work Macro Practice
• Social Work Practice Settings
• Social Work Research and Evidence-based Practice
• Welfare and Benefit Systems
• Browse content in Sociology
• Childhood Studies
• Community Development
• Comparative and Historical Sociology
• Disability Studies
• Economic Sociology
• Gender and Sexuality
• Gerontology and Ageing
• Health, Illness, and Medicine
• Marriage and the Family
• Migration Studies
• Occupations, Professions, and Work
• Organizations
• Population and Demography
• Race and Ethnicity
• Social Theory
• Social Movements and Social Change
• Social Research and Statistics
• Social Stratification, Inequality, and Mobility
• Sociology of Religion
• Sociology of Education
• Sport and Leisure
• Urban and Rural Studies
• Browse content in Warfare and Defence
• Defence Strategy, Planning, and Research
• Land Forces and Warfare
• Military Administration
• Military Life and Institutions
• Naval Forces and Warfare
• Other Warfare and Defence Issues
• Peace Studies and Conflict Resolution
• Weapons and Equipment

• < Previous chapter
• Next chapter >

6 Semi-structured interviewing

• Published: June 2005
• Cite Icon Cite
• Permissions Icon Permissions

This chapter presents a guide to conducting effective semi-structured interviews. It discusses the nature of semi-structured interviews and why they should be used, as well as preparation, the logistics of conducting the interview, and reflexivity.

Personal account

• Sign in with email/username & password
• Get email alerts
• Save searches
• Purchase content
• Activate your purchase/trial code
• Add your ORCID iD

Institutional access

Sign in with a library card.

• Sign in with username/password
• Recommend to your librarian
• Institutional account management
• Get help with access

Access to content on Oxford Academic is often provided through institutional subscriptions and purchases. If you are a member of an institution with an active account, you may be able to access content in one of the following ways:

IP based access

Typically, access is provided across an institutional network to a range of IP addresses. This authentication occurs automatically, and it is not possible to sign out of an IP authenticated account.

Choose this option to get remote access when outside your institution. Shibboleth/Open Athens technology is used to provide single sign-on between your institution’s website and Oxford Academic.

• Click Sign in through your institution.
• Select your institution from the list provided, which will take you to your institution's website to sign in.
• When on the institution site, please use the credentials provided by your institution. Do not use an Oxford Academic personal account.
• Following successful sign in, you will be returned to Oxford Academic.

If your institution is not listed or you cannot sign in to your institution’s website, please contact your librarian or administrator.

Enter your library card number to sign in. If you cannot sign in, please contact your librarian.

Society Members

Society member access to a journal is achieved in one of the following ways:

Sign in through society site

Many societies offer single sign-on between the society website and Oxford Academic. If you see ‘Sign in through society site’ in the sign in pane within a journal:

• Click Sign in through society site.
• When on the society site, please use the credentials provided by that society. Do not use an Oxford Academic personal account.

If you do not have a society account or have forgotten your username or password, please contact your society.

Sign in using a personal account

Some societies use Oxford Academic personal accounts to provide access to their members. See below.

A personal account can be used to get email alerts, save searches, purchase content, and activate subscriptions.

Some societies use Oxford Academic personal accounts to provide access to their members.

Viewing your signed in accounts

Click the account icon in the top right to:

• View your signed in personal account and access account management features.
• View the institutional accounts that are providing access.

Signed in but can't access content

Oxford Academic is home to a wide variety of products. The institutional subscription may not cover the content that you are trying to access. If you believe you should have access to that content, please contact your librarian.

For librarians and administrators, your personal account also provides access to institutional account management. Here you will find options to view and activate subscriptions, manage institutional settings and access options, access usage statistics, and more.

Our books are available by subscription or purchase to libraries and institutions.

Month:	Total Views:
October 2022	6
November 2022	23
December 2022	5
January 2023	7
February 2023	13
March 2023	17
April 2023	12
May 2023	9
June 2023	4
July 2023	11
August 2023	13
September 2023	15
October 2023	10
November 2023	5
December 2023	5
January 2024	4
February 2024	13
March 2024	5
April 2024	6
May 2024	11
June 2024	10
July 2024	8
August 2024	5

• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Rights and permissions
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

• Reviews / Why join our community?
• For companies
• Frequently asked questions

Semi-Structured Interviews

What are semi-structured interviews.

Semi-structured interviews are a research method that uses both predetermined questions and open-ended exploration to gain more in-depth insights into participants' perspectives, attitudes, and experiences.

• Transcript loading…

Semi-structured interviews are commonly used in social science research, market research, and other fields where an understanding of people's attitudes, behaviors, and beliefs is important.

Key Characteristics of Semi-Structured Interviews

Semi-structured interviews have several key characteristics that differentiate them from other types of interviews: 

The flexible nature allows researchers to dive deeper into a topic and adapt the interview based on new insights or issues. Unlike structured interviews, which rely on a fixed set of questions and responses, semi-structured interviews allow for more open-ended discussion, which can lead to unexpected insights and perspectives.

Their emphasis is on participant perspectives and experiences. Rather than simply gathering participant data or information, the purpose of semi-structured interviews is to understand how participants think and feel about particular topics or issues. This approach allows researchers to understand better the social and cultural contexts in which participants live and work.

They are often used in research projects that aim to generate new ideas or theories rather than test existing ones. Because they allow for open-ended discussion and exploration, they can effectively generate new insights into complex social phenomena.

Types of Questions for Semi-Structured Interviews

Semi-structured interviews use a combination of predetermined questions and open-ended exploration to learn more about participants' perspectives. There are three main categories of questions you can use:

Open-ended Questions: These are broad, general questions that allow participants to express their thoughts and feelings on a topic without restriction. Open-ended questions typically begin with phrases like "Tell me about..." or "How do you feel about...". These questions help encourage participants to share their experiences and perspectives in their own words.

Closed-ended Questions: Closed-ended questions are more specific and provide the participant with predetermined responses. These questions typically begin with phrases like "Do you agree or disagree with..." or "Which option best describes...". Closed-ended questions can help gather data on specific attitudes or behaviors.

Probing Questions: Probing questions are follow-up questions that aim to clarify or expand upon a participant's response. These questions typically begin with phrases like "Can you tell me more about..." or "Why do you think that is...". Probing questions can help a researcher to understand a participant's thought process or experience.

Steps to Conduct a Successful Semi-Structured Interview

Proper preparation is key to conducting successful semi-structured interviews. Below are some tips for preparing for your interviews:

Define Your Research Questions: Before conducting interviews, it's important to understand your research questions and objectives clearly. This will help you develop a set of initial questions to guide your interview process.

Develop an Interview Guide: An interview guide is a list of questions and prompts designed to elicit information from participants. It should include open-ended and closed-ended questions and probing questions to encourage participants to elaborate on their responses.

Pilot Test Your Interview Guide: It's important to pilot test your interview guide with a small group of participants before conducting full-scale interviews. This will allow you to identify potential issues or areas where the questions must be revised.

Identify and Recruit Participants: Ensure that your sample is representative of the population you are studying. Consider using targeted sampling methods, such as snowball sampling or maximum variation sampling, to recruit participants who can provide diverse perspectives.

Schedule Interviews: Once you've identified and recruited participants, it's time to schedule interviews. Be sure to allow adequate time between interviews for transcription and analysis.

Conduct Interviews: During the interview process, it's important to establish rapport with participants and create a comfortable environment where they feel safe sharing their experiences and opinions. Be sure to follow your interview guide while allowing flexibility in response to unexpected information during the discussion.

Provide Compensation or Incentives to Participants: Consider offering compensation or incentives to participants to encourage their participation. Compensation can come in many forms, such as gift cards, cash, or vouchers. It can also be non-monetary, such as offering participants the opportunity to receive a summary of the study's findings or the chance to participate in future research projects. Compensation or incentives can help to show participants that their time and contributions are valued and appreciated.

Plan your research with this helpful checklist . Then, get ready to conduct semi-structured interviews! Download this template for help in creating different types of interview questions. 

Learn More about Semi Structured Interviews

Take our course on User Research – Methods and Best Practices. 

Read more about the process of conducting semi-structured interviews .

Learn how to analyze the data from your semi-structured interviews .  

Read this reflection on semi-structured interviews as a research instrument .  

Learn how to use the snowball sampling method to recruit participants.

Do you need more diversity in your study? Try maximum variation sampling .

Answer a Short Quiz to Earn a Gift

What are semi-structured interviews?

• Interviews that consist only of casual conversations without any structure.
• Interviews that follow a strict set of questions without deviation.
• Interviews that use both predetermined questions and allow open-ended exploration.

Why is building rapport important in semi-structured interviews?

• It allows the interviewer to dominate the conversation.
• It encourages participants to provide honest and detailed responses.
• It guarantees the interviewer sticks to the script.

Which type of question do researchers commonly use in semi-structured interviews to gain in-depth responses?

• Multiple-choice questions
• Open-ended questions
• Yes/no questions

Why are semi-structured interviews beneficial in research?

• They allow for the exploration of new topics based on responses.
• They are easy to analyze quantitatively.
• They require minimal preparation.

What is the purpose of probing questions in semi-structured interviews?

• To clarify and expand on participants' answers
• To gather specific predetermined data
• To limit the participants' responses

Better luck next time!

Do you want to improve your UX / UI Design skills? Join us now

Congratulations! You did amazing

You earned your gift with a perfect score! Let us send it to you.

Check Your Inbox

We’ve emailed your gift to [email protected] .

Literature on Semi-Structured Interviews

Here’s the entire UX literature on Semi-Structured Interviews by the Interaction Design Foundation, collated in one place:

Learn more about Semi-Structured Interviews

Take a deep dive into Semi-Structured Interviews with our course User Research – Methods and Best Practices .

How do you plan to design a product or service that your users will love , if you don't know what they want in the first place? As a user experience designer, you shouldn't leave it to chance to design something outstanding; you should make the effort to understand your users and build on that knowledge from the outset. User research is the way to do this, and it can therefore be thought of as the largest part of user experience design .

In fact, user research is often the first step of a UX design process—after all, you cannot begin to design a product or service without first understanding what your users want! As you gain the skills required, and learn about the best practices in user research, you’ll get first-hand knowledge of your users and be able to design the optimal product—one that’s truly relevant for your users and, subsequently, outperforms your competitors’ .

This course will give you insights into the most essential qualitative research methods around and will teach you how to put them into practice in your design work. You’ll also have the opportunity to embark on three practical projects where you can apply what you’ve learned to carry out user research in the real world . You’ll learn details about how to plan user research projects and fit them into your own work processes in a way that maximizes the impact your research can have on your designs. On top of that, you’ll gain practice with different methods that will help you analyze the results of your research and communicate your findings to your clients and stakeholders—workshops, user journeys and personas, just to name a few!

By the end of the course, you’ll have not only a Course Certificate but also three case studies to add to your portfolio. And remember, a portfolio with engaging case studies is invaluable if you are looking to break into a career in UX design or user research!

We believe you should learn from the best, so we’ve gathered a team of experts to help teach this course alongside our own course instructors. That means you’ll meet a new instructor in each of the lessons on research methods who is an expert in their field—we hope you enjoy what they have in store for you!

All open-source articles on Semi-Structured Interviews

How to do a thematic analysis of user interviews.

• 1.3k shares
• 4 years ago

Pros and Cons of Conducting User Interviews

How to Moderate User Interviews

Open Access—Link to us!

We believe in Open Access and the  democratization of knowledge . Unfortunately, world-class educational materials such as this page are normally hidden behind paywalls or in expensive textbooks.

If you want this to change , cite this page , link to us, or join us to help us democratize design knowledge !

Privacy Settings

Our digital services use necessary tracking technologies, including third-party cookies, for security, functionality, and to uphold user rights. Optional cookies offer enhanced features, and analytics.

Experience the full potential of our site that remembers your preferences and supports secure sign-in.

Governs the storage of data necessary for maintaining website security, user authentication, and fraud prevention mechanisms.

Enhanced Functionality

Saves your settings and preferences, like your location, for a more personalized experience.

Referral Program

We use cookies to enable our referral program, giving you and your friends discounts.

Error Reporting

We share user ID with Bugsnag and NewRelic to help us track errors and fix issues.

Optimize your experience by allowing us to monitor site usage. You’ll enjoy a smoother, more personalized journey without compromising your privacy.

Analytics Storage

Collects anonymous data on how you navigate and interact, helping us make informed improvements.

Differentiates real visitors from automated bots, ensuring accurate usage data and improving your website experience.

Lets us tailor your digital ads to match your interests, making them more relevant and useful to you.

Advertising Storage

Stores information for better-targeted advertising, enhancing your online ad experience.

Personalization Storage

Permits storing data to personalize content and ads across Google services based on user behavior, enhancing overall user experience.

Advertising Personalization

Allows for content and ad personalization across Google services based on user behavior. This consent enhances user experiences.

Enables personalizing ads based on user data and interactions, allowing for more relevant advertising experiences across Google services.

Receive more relevant advertisements by sharing your interests and behavior with our trusted advertising partners.

Enables better ad targeting and measurement on Meta platforms, making ads you see more relevant.

Allows for improved ad effectiveness and measurement through Meta’s Conversions API, ensuring privacy-compliant data sharing.

LinkedIn Insights

Tracks conversions, retargeting, and web analytics for LinkedIn ad campaigns, enhancing ad relevance and performance.

LinkedIn CAPI

Enhances LinkedIn advertising through server-side event tracking, offering more accurate measurement and personalization.

Google Ads Tag

Tracks ad performance and user engagement, helping deliver ads that are most useful to you.

Share Knowledge, Get Respect!

or copy link

Cite according to academic standards

Simply copy and paste the text below into your bibliographic reference list, onto your blog, or anywhere else. You can also just hyperlink to this page.

New to UX Design? We’re Giving You a Free ebook!

Download our free ebook The Basics of User Experience Design to learn about core concepts of UX design.

In 9 chapters, we’ll cover: conducting user interviews, design thinking, interaction design, mobile UX design, usability, UX research, and many more!

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Published: 15 September 2022

Interviews in the social sciences

• Eleanor Knott   ORCID: orcid.org/0000-0002-9131-3939 1 ,
• Aliya Hamid Rao   ORCID: orcid.org/0000-0003-0674-4206 1 ,
• Kate Summers   ORCID: orcid.org/0000-0001-9964-0259 1 &
• Chana Teeger   ORCID: orcid.org/0000-0002-5046-8280 1  

Nature Reviews Methods Primers volume  2 , Article number:  73 ( 2022 ) Cite this article

734k Accesses

77 Citations

190 Altmetric

Metrics details

• Interdisciplinary studies

In-depth interviews are a versatile form of qualitative data collection used by researchers across the social sciences. They allow individuals to explain, in their own words, how they understand and interpret the world around them. Interviews represent a deceptively familiar social encounter in which people interact by asking and answering questions. They are, however, a very particular type of conversation, guided by the researcher and used for specific ends. This dynamic introduces a range of methodological, analytical and ethical challenges, for novice researchers in particular. In this Primer, we focus on the stages and challenges of designing and conducting an interview project and analysing data from it, as well as strategies to overcome such challenges.

Similar content being viewed by others

The fundamental importance of method to theory

How ‘going online’ mediates the challenges of policy elite interviews

Participatory action research

Introduction.

In-depth interviews are a qualitative research method that follow a deceptively familiar logic of human interaction: they are conversations where people talk with each other, interact and pose and answer questions 1 . An interview is a specific type of interaction in which — usually and predominantly — a researcher asks questions about someone’s life experience, opinions, dreams, fears and hopes and the interview participant answers the questions 1 .

Interviews will often be used as a standalone method or combined with other qualitative methods, such as focus groups or ethnography, or quantitative methods, such as surveys or experiments. Although interviewing is a frequently used method, it should not be viewed as an easy default for qualitative researchers 2 . Interviews are also not suited to answering all qualitative research questions, but instead have specific strengths that should guide whether or not they are deployed in a research project. Whereas ethnography might be better suited to trying to observe what people do, interviews provide a space for extended conversations that allow the researcher insights into how people think and what they believe. Quantitative surveys also give these kinds of insights, but they use pre-determined questions and scales, privileging breadth over depth and often overlooking harder-to-reach participants.

In-depth interviews can take many different shapes and forms, often with more than one participant or researcher. For example, interviews might be highly structured (using an almost survey-like interview guide), entirely unstructured (taking a narrative and free-flowing approach) or semi-structured (using a topic guide ). Researchers might combine these approaches within a single project depending on the purpose of the interview and the characteristics of the participant. Whatever form the interview takes, researchers should be mindful of the dynamics between interviewer and participant and factor these in at all stages of the project.

In this Primer, we focus on the most common type of interview: one researcher taking a semi-structured approach to interviewing one participant using a topic guide. Focusing on how to plan research using interviews, we discuss the necessary stages of data collection. We also discuss the stages and thought-process behind analysing interview material to ensure that the richness and interpretability of interview material is maintained and communicated to readers. The Primer also tracks innovations in interview methods and discusses the developments we expect over the next 5–10 years.

We wrote this Primer as researchers from sociology, social policy and political science. We note our disciplinary background because we acknowledge that there are disciplinary differences in how interviews are approached and understood as a method.

Experimentation

Here we address research design considerations and data collection issues focusing on topic guide construction and other pragmatics of the interview. We also explore issues of ethics and reflexivity that are crucial throughout the research project.

Research design

Participant selection.

Participants can be selected and recruited in various ways for in-depth interview studies. The researcher must first decide what defines the people or social groups being studied. Often, this means moving from an abstract theoretical research question to a more precise empirical one. For example, the researcher might be interested in how people talk about race in contexts of diversity. Empirical settings in which this issue could be studied could include schools, workplaces or adoption agencies. The best research designs should clearly explain why the particular setting was chosen. Often there are both intrinsic and extrinsic reasons for choosing to study a particular group of people at a specific time and place 3 . Intrinsic motivations relate to the fact that the research is focused on an important specific social phenomenon that has been understudied. Extrinsic motivations speak to the broader theoretical research questions and explain why the case at hand is a good one through which to address them empirically.

Next, the researcher needs to decide which types of people they would like to interview. This decision amounts to delineating the inclusion and exclusion criteria for the study. The criteria might be based on demographic variables, like race or gender, but they may also be context-specific, for example, years of experience in an organization. These should be decided based on the research goals. Researchers should be clear about what characteristics would make an individual a candidate for inclusion in the study (and what would exclude them).

The next step is to identify and recruit the study’s sample . Usually, many more people fit the inclusion criteria than can be interviewed. In cases where lists of potential participants are available, the researcher might want to employ stratified sampling , dividing the list by characteristics of interest before sampling.

When there are no lists, researchers will often employ purposive sampling . Many researchers consider purposive sampling the most useful mode for interview-based research since the number of interviews to be conducted is too small to aim to be statistically representative 4 . Instead, the aim is not breadth, via representativeness, but depth via rich insights about a set of participants. In addition to purposive sampling, researchers often use snowball sampling . Both purposive and snowball sampling can be combined with quota sampling . All three types of sampling aim to ensure a variety of perspectives within the confines of a research project. A goal for in-depth interview studies can be to sample for range, being mindful of recruiting a diversity of participants fitting the inclusion criteria.

Study design

The total number of interviews depends on many factors, including the population studied, whether comparisons are to be made and the duration of interviews. Studies that rely on quota sampling where explicit comparisons are made between groups will require a larger number of interviews than studies focused on one group only. Studies where participants are interviewed over several hours, days or even repeatedly across years will tend to have fewer participants than those that entail a one-off engagement.

Researchers often stop interviewing when new interviews confirm findings from earlier interviews with no new or surprising insights (saturation) 4 , 5 , 6 . As a criterion for research design, saturation assumes that data collection and analysis are happening in tandem and that researchers will stop collecting new data once there is no new information emerging from the interviews. This is not always possible. Researchers rarely have time for systematic data analysis during data collection and they often need to specify their sample in funding proposals prior to data collection. As a result, researchers often draw on existing reports of saturation to estimate a sample size prior to data collection. These suggest between 12 and 20 interviews per category of participant (although researchers have reported saturation with samples that are both smaller and larger than this) 7 , 8 , 9 . The idea of saturation has been critiqued by many qualitative researchers because it assumes that meaning inheres in the data, waiting to be discovered — and confirmed — once saturation has been reached 7 . In-depth interview data are often multivalent and can give rise to different interpretations. The important consideration is, therefore, not merely how many participants are interviewed, but whether one’s research design allows for collecting rich and textured data that provide insight into participants’ understandings, accounts, perceptions and interpretations.

Sometimes, researchers will conduct interviews with more than one participant at a time. Researchers should consider the benefits and shortcomings of such an approach. Joint interviews may, for example, give researchers insight into how caregivers agree or debate childrearing decisions. At the same time, they may be less adaptive to exploring aspects of caregiving that participants may not wish to disclose to each other. In other cases, there may be more than one person interviewing each participant, such as when an interpreter is used, and so it is important to consider during the research design phase how this might shape the dynamics of the interview.

Data collection

Semi-structured interviews are typically organized around a topic guide comprised of an ordered set of broad topics (usually 3–5). Each topic includes a set of questions that form the basis of the discussion between the researcher and participant (Fig.  1 ). These topics are organized around key concepts that the researcher has identified (for example, through a close study of prior research, or perhaps through piloting a small, exploratory study) 5 .

a | Elaborated topics the researcher wants to cover in the interview and example questions. b | An example topic arc. Using such an arc, one can think flexibly about the order of topics. Considering the main question for each topic will help to determine the best order for the topics. After conducting some interviews, the researcher can move topics around if a different order seems to make sense.

Topic guide

One common way to structure a topic guide is to start with relatively easy, open-ended questions (Table  1 ). Opening questions should be related to the research topic but broad and easy to answer, so that they help to ease the participant into conversation.

After these broad, opening questions, the topic guide may move into topics that speak more directly to the overarching research question. The interview questions will be accompanied by probes designed to elicit concrete details and examples from the participant (see Table  1 ).

Abstract questions are often easier for participants to answer once they have been asked more concrete questions. In our experience, for example, questions about feelings can be difficult for some participants to answer, but when following probes concerning factual experiences these questions can become less challenging. After the main themes of the topic guide have been covered, the topic guide can move onto closing questions. At this stage, participants often repeat something they have said before, although they may sometimes introduce a new topic.

Interviews are especially well suited to gaining a deeper insight into people’s experiences. Getting these insights largely depends on the participants’ willingness to talk to the researcher. We recommend designing open-ended questions that are more likely to elicit an elaborated response and extended reflection from participants rather than questions that can be answered with yes or no.

Questions should avoid foreclosing the possibility that the participant might disagree with the premise of the question. Take for example the question: “Do you support the new family-friendly policies?” This question minimizes the possibility of the participant disagreeing with the premise of this question, which assumes that the policies are ‘family-friendly’ and asks for a yes or no answer. Instead, asking more broadly how a participant feels about the specific policy being described as ‘family-friendly’ (for example, a work-from-home policy) allows them to express agreement, disagreement or impartiality and, crucially, to explain their reasoning 10 .

For an uninterrupted interview that will last between 90 and 120 minutes, the topic guide should be one to two single-spaced pages with questions and probes. Ideally, the researcher will memorize the topic guide before embarking on the first interview. It is fine to carry a printed-out copy of the topic guide but memorizing the topic guide ahead of the interviews can often make the interviewer feel well prepared in guiding the participant through the interview process.

Although the topic guide helps the researcher stay on track with the broad areas they want to cover, there is no need for the researcher to feel tied down by the topic guide. For instance, if a participant brings up a theme that the researcher intended to discuss later or a point the researcher had not anticipated, the researcher may well decide to follow the lead of the participant. The researcher’s role extends beyond simply stating the questions; it entails listening and responding, making split-second decisions about what line of inquiry to pursue and allowing the interview to proceed in unexpected directions.

Optimizing the interview

The ideal place for an interview will depend on the study and what is feasible for participants. Generally, a place where the participant and researcher can both feel relaxed, where the interview can be uninterrupted and where noise or other distractions are limited is ideal. But this may not always be possible and so the researcher needs to be prepared to adapt their plans within what is feasible (and desirable for participants).

Another key tool for the interview is a recording device (assuming that permission for recording has been given). Recording can be important to capture what the participant says verbatim. Additionally, it can allow the researcher to focus on determining what probes and follow-up questions they want to pursue rather than focusing on taking notes. Sometimes, however, a participant may not allow the researcher to record, or the recording may fail. If the interview is not recorded we suggest that the researcher takes brief notes during the interview, if feasible, and then thoroughly make notes immediately after the interview and try to remember the participant’s facial expressions, gestures and tone of voice. Not having a recording of an interview need not limit the researcher from getting analytical value from it.

As soon as possible after each interview, we recommend that the researcher write a one-page interview memo comprising three key sections. The first section should identify two to three important moments from the interview. What constitutes important is up to the researcher’s discretion 9 . The researcher should note down what happened in these moments, including the participant’s facial expressions, gestures, tone of voice and maybe even the sensory details of their surroundings. This exercise is about capturing ethnographic detail from the interview. The second part of the interview memo is the analytical section with notes on how the interview fits in with previous interviews, for example, where the participant’s responses concur or diverge from other responses. The third part consists of a methodological section where the researcher notes their perception of their relationship with the participant. The interview memo allows the researcher to think critically about their positionality and practice reflexivity — key concepts for an ethical and transparent research practice in qualitative methodology 11 , 12 .

Ethics and reflexivity

All elements of an in-depth interview can raise ethical challenges and concerns. Good ethical practice in interview studies often means going beyond the ethical procedures mandated by institutions 13 . While discussions and requirements of ethics can differ across disciplines, here we focus on the most pertinent considerations for interviews across the research process for an interdisciplinary audience.

Ethical considerations prior to interview

Before conducting interviews, researchers should consider harm minimization, informed consent, anonymity and confidentiality, and reflexivity and positionality. It is important for the researcher to develop their own ethical sensitivities and sensibilities by gaining training in interview and qualitative methods, reading methodological and field-specific texts on interviews and ethics and discussing their research plans with colleagues.

Researchers should map the potential harm to consider how this can be minimized. Primarily, researchers should consider harm from the participants’ perspective (Box  1 ). But, it is also important to consider and plan for potential harm to the researcher, research assistants, gatekeepers, future researchers and members of the wider community 14 . Even the most banal of research topics can potentially pose some form of harm to the participant, researcher and others — and the level of harm is often highly context-dependent. For example, a research project on religion in society might have very different ethical considerations in a democratic versus authoritarian research context because of how openly or not such topics can be discussed and debated 15 .

The researcher should consider how they will obtain and record informed consent (for example, written or oral), based on what makes the most sense for their research project and context 16 . Some institutions might specify how informed consent should be gained. Regardless of how consent is obtained, the participant must be made aware of the form of consent, the intentions and procedures of the interview and potential forms of harm and benefit to the participant or community before the interview commences. Moreover, the participant must agree to be interviewed before the interview commences. If, in addition to interviews, the study contains an ethnographic component, it is worth reading around this topic (see, for example, Murphy and Dingwall 17 ). Informed consent must also be gained for how the interview will be recorded before the interview commences. These practices are important to ensure the participant is contributing on a voluntary basis. It is also important to remind participants that they can withdraw their consent at any time during the interview and for a specified period after the interview (to be decided with the participant). The researcher should indicate that participants can ask for anything shared to be off the record and/or not disseminated.

In terms of anonymity and confidentiality, it is standard practice when conducting interviews to agree not to use (or even collect) participants’ names and personal details that are not pertinent to the study. Anonymizing can often be the safer option for minimizing harm to participants as it is hard to foresee all the consequences of de-anonymizing, even if participants agree. Regardless of what a researcher decides, decisions around anonymity must be agreed with participants during the process of gaining informed consent and respected following the interview.

Although not all ethical challenges can be foreseen or planned for 18 , researchers should think carefully — before the interview — about power dynamics, participant vulnerability, emotional state and interactional dynamics between interviewer and participant, even when discussing low-risk topics. Researchers may then wish to plan for potential ethical issues, for example by preparing a list of relevant organizations to which participants can be signposted. A researcher interviewing a participant about debt, for instance, might prepare in advance a list of debt advice charities, organizations and helplines that could provide further support and advice. It is important to remember that the role of an interviewer is as a researcher rather than as a social worker or counsellor because researchers may not have relevant and requisite training in these other domains.

Box 1 Mapping potential forms of harm

Social: researchers should avoid causing any relational detriment to anyone in the course of interviews, for example, by sharing information with other participants or causing interview participants to be shunned or mistreated by their community as a result of participating.

Economic: researchers should avoid causing financial detriment to anyone, for example, by expecting them to pay for transport to be interviewed or to potentially lose their job as a result of participating.

Physical: researchers should minimize the risk of anyone being exposed to violence as a result of the research both from other individuals or from authorities, including police.

Psychological: researchers should minimize the risk of causing anyone trauma (or re-traumatization) or psychological anguish as a result of the research; this includes not only the participant but importantly the researcher themselves and anyone that might read or analyse the transcripts, should they contain triggering information.

Political: researchers should minimize the risk of anyone being exposed to political detriment as a result of the research, such as retribution.

Professional/reputational: researchers should minimize the potential for reputational damage to anyone connected to the research (this includes ensuring good research practices so that any researchers involved are not harmed reputationally by being involved with the research project).

The task here is not to map exhaustively the potential forms of harm that might pertain to a particular research project (that is the researcher’s job and they should have the expertise most suited to mapping such potential harms relative to the specific project) but to demonstrate the breadth of potential forms of harm.

Ethical considerations post-interview

Researchers should consider how interview data are stored, analysed and disseminated. If participants have been offered anonymity and confidentiality, data should be stored in a way that does not compromise this. For example, researchers should consider removing names and any other unnecessary personal details from interview transcripts, password-protecting and encrypting files and using pseudonyms to label and store all interview data. It is also important to address where interview data are taken (for example, across borders in particular where interview data might be of interest to local authorities) and how this might affect the storage of interview data.

Examining how the researcher will represent participants is a paramount ethical consideration both in the planning stages of the interview study and after it has been conducted. Dissemination strategies also need to consider questions of anonymity and representation. In small communities, even if participants are given pseudonyms, it might be obvious who is being described. Anonymizing not only the names of those participating but also the research context is therefore a standard practice 19 . With particularly sensitive data or insights about the participant, it is worth considering describing participants in a more abstract way rather than as specific individuals. These practices are important both for protecting participants’ anonymity but can also affect the ability of the researcher and others to return ethically to the research context and similar contexts 20 .

Reflexivity and positionality

Reflexivity and positionality mean considering the researcher’s role and assumptions in knowledge production 13 . A key part of reflexivity is considering the power relations between the researcher and participant within the interview setting, as well as how researchers might be perceived by participants. Further, researchers need to consider how their own identities shape the kind of knowledge and assumptions they bring to the interview, including how they approach and ask questions and their analysis of interviews (Box  2 ). Reflexivity is a necessary part of developing ethical sensibility as a researcher by adapting and reflecting on how one engages with participants. Participants should not feel judged, for example, when they share information that researchers might disagree with or find objectionable. How researchers deal with uncomfortable moments or information shared by participants is at their discretion, but they should consider how they will react both ahead of time and in the moment.

Researchers can develop their reflexivity by considering how they themselves would feel being asked these interview questions or represented in this way, and then adapting their practice accordingly. There might be situations where these questions are not appropriate in that they unduly centre the researchers’ experiences and worldview. Nevertheless, these prompts can provide a useful starting point for those beginning their reflexive journey and developing an ethical sensibility.

Reflexivity and ethical sensitivities require active reflection throughout the research process. For example, researchers should take care in interview memos and their notes to consider their assumptions, potential preconceptions, worldviews and own identities prior to and after interviews (Box  2 ). Checking in with assumptions can be a way of making sure that researchers are paying close attention to their own theoretical and analytical biases and revising them in accordance with what they learn through the interviews. Researchers should return to these notes (especially when analysing interview material), to try to unpack their own effects on the research process as well as how participants positioned and engaged with them.

Box 2 Aspects to reflect on reflexively

For reflexive engagement, and understanding the power relations being co-constructed and (re)produced in interviews, it is necessary to reflect, at a minimum, on the following.

Ethnicity, race and nationality, such as how does privilege stemming from race or nationality operate between the researcher, the participant and research context (for example, a researcher from a majority community may be interviewing a member of a minority community)

Gender and sexuality, see above on ethnicity, race and nationality

Social class, and in particular the issue of middle-class bias among researchers when formulating research and interview questions

Economic security/precarity, see above on social class and thinking about the researcher’s relative privilege and the source of biases that stem from this

Educational experiences and privileges, see above

Disciplinary biases, such as how the researcher’s discipline/subfield usually approaches these questions, possibly normalizing certain assumptions that might be contested by participants and in the research context

Political and social values

Lived experiences and other dimensions of ourselves that affect and construct our identity as researchers

In this section, we discuss the next stage of an interview study, namely, analysing the interview data. Data analysis may begin while more data are being collected. Doing so allows early findings to inform the focus of further data collection, as part of an iterative process across the research project. Here, the researcher is ultimately working towards achieving coherence between the data collected and the findings produced to answer successfully the research question(s) they have set.

The two most common methods used to analyse interview material across the social sciences are thematic analysis 21 and discourse analysis 22 . Thematic analysis is a particularly useful and accessible method for those starting out in analysis of qualitative data and interview material as a method of coding data to develop and interpret themes in the data 21 . Discourse analysis is more specialized and focuses on the role of discourse in society by paying close attention to the explicit, implicit and taken-for-granted dimensions of language and power 22 , 23 . Although thematic and discourse analysis are often discussed as separate techniques, in practice researchers might flexibly combine these approaches depending on the object of analysis. For example, those intending to use discourse analysis might first conduct thematic analysis as a way to organize and systematize the data. The object and intention of analysis might differ (for example, developing themes or interrogating language), but the questions facing the researcher (such as whether to take an inductive or deductive approach to analysis) are similar.

Preparing data

Data preparation is an important step in the data analysis process. The researcher should first determine what comprises the corpus of material and in what form it will it be analysed. The former refers to whether, for example, alongside the interviews themselves, analytic memos or observational notes that may have been taken during data collection will also be directly analysed. The latter refers to decisions about how the verbal/audio interview data will be transformed into a written form, making it suitable for processes of data analysis. Typically, interview audio recordings are transcribed to produce a written transcript. It is important to note that the process of transcription is one of transformation. The verbal interview data are transformed into a written transcript through a series of decisions that the researcher must make. The researcher should consider the effect of mishearing what has been said or how choosing to punctuate a sentence in a particular way will affect the final analysis.

Box  3 shows an example transcript excerpt from an interview with a teacher conducted by Teeger as part of her study of history education in post-apartheid South Africa 24 (Box  3 ). Seeing both the questions and the responses means that the reader can contextualize what the participant (Ms Mokoena) has said. Throughout the transcript the researcher has used square brackets, for example to indicate a pause in speech, when Ms Mokoena says “it’s [pause] it’s a difficult topic”. The transcription choice made here means that we see that Ms Mokoena has taken time to pause, perhaps to search for the right words, or perhaps because she has a slight apprehension. Square brackets are also included as an overt act of communication to the reader. When Ms Mokoena says “ja”, the English translation (“yes”) of the word in Afrikaans is placed in square brackets to ensure that the reader can follow the meaning of the speech.

Decisions about what to include when transcribing will be hugely important for the direction and possibilities of analysis. Researchers should decide what they want to capture in the transcript, based on their analytic focus. From a (post)positivist perspective 25 , the researcher may be interested in the manifest content of the interview (such as what is said, not how it is said). In that case, they may choose to transcribe intelligent verbatim . From a constructivist perspective 25 , researchers may choose to record more aspects of speech (including, for example, pauses, repetitions, false starts, talking over one another) so that these features can be analysed. Those working from this perspective argue that to recognize the interactional nature of the interview setting adequately and to avoid misinterpretations, features of interaction (pauses, overlaps between speakers and so on) should be preserved in transcription and therefore in the analysis 10 . Readers interested in learning more should consult Potter and Hepburn’s summary of how to present interaction through transcription of interview data 26 .

The process of analysing semi-structured interviews might be thought of as a generative rather than an extractive enterprise. Findings do not already exist within the interview data to be discovered. Rather, researchers create something new when analysing the data by applying their analytic lens or approach to the transcripts. At a high level, there are options as to what researchers might want to glean from their interview data. They might be interested in themes, whereby they identify patterns of meaning across the dataset 21 . Alternatively, they may focus on discourse(s), looking to identify how language is used to construct meanings and therefore how language reinforces or produces aspects of the social world 27 . Alternatively, they might look at the data to understand narrative or biographical elements 28 .

A further overarching decision to make is the extent to which researchers bring predetermined framings or understandings to bear on their data, or instead begin from the data themselves to generate an analysis. One way of articulating this is the extent to which researchers take a deductive approach or an inductive approach to analysis. One example of a truly inductive approach is grounded theory, whereby the aim of the analysis is to build new theory, beginning with one’s data 6 , 29 . In practice, researchers using thematic and discourse analysis often combine deductive and inductive logics and describe their process instead as iterative (referred to also as an abductive approach ) 30 , 31 . For example, researchers may decide that they will apply a given theoretical framing, or begin with an initial analytic framework, but then refine or develop these once they begin the process of analysis.

Box 3 Excerpt of interview transcript (from Teeger 24 )

Interviewer : Maybe you could just start by talking about what it’s like to teach apartheid history.

Ms Mokoena : It’s a bit challenging. You’ve got to accommodate all the kids in the class. You’ve got to be sensitive to all the racial differences. You want to emphasize the wrongs that were done in the past but you also want to, you know, not to make kids feel like it’s their fault. So you want to use the wrongs of the past to try and unite the kids …

Interviewer : So what kind of things do you do?

Ms Mokoena : Well I normally highlight the fact that people that were struggling were not just the blacks, it was all the races. And I give examples of the people … from all walks of life, all races, and highlight how they suffered as well as a result of apartheid, particularly the whites… . What I noticed, particularly my first year of teaching apartheid, I noticed that the black kids made the others feel responsible for what happened… . I had a lot of fights…. A lot of kids started hating each other because, you know, the others are white and the others were black. And they started saying, “My mother is a domestic worker because she was never allowed an opportunity to get good education.” …

Interviewer : I didn’t see any of that now when I was observing.

Ms Mokoena : … Like I was saying I think that because of the re-emphasis of the fact that, look, everybody did suffer one way or the other, they sort of got to see that it was everybody’s struggle … . They should now get to understand that that’s why we’re called a Rainbow Nation. Not everybody agreed with apartheid and not everybody suffered. Even all the blacks, not all blacks got to feel what the others felt . So ja [yes], it’s [pause] it’s a difficult topic, ja . But I think if you get the kids to understand why we’re teaching apartheid in the first place and you show the involvement of all races in all the different sides , then I think you have managed to teach it properly. So I think because of my inexperience then — that was my first year of teaching history — so I think I — maybe I over-emphasized the suffering of the blacks versus the whites [emphasis added].

Reprinted with permission from ref. 24 , Sage Publications.

From data to codes

Coding data is a key building block shared across many approaches to data analysis. Coding is a way of organizing and describing data, but is also ultimately a way of transforming data to produce analytic insights. The basic practice of coding involves highlighting a segment of text (this may be a sentence, a clause or a longer excerpt) and assigning a label to it. The aim of the label is to communicate some sort of summary of what is in the highlighted piece of text. Coding is an iterative process, whereby researchers read and reread their transcripts, applying and refining their codes, until they have a coding frame (a set of codes) that is applied coherently across the dataset and that captures and communicates the key features of what is contained in the data as it relates to the researchers’ analytic focus.

What one codes for is entirely contingent on the focus of the research project and the choices the researcher makes about the approach to analysis. At first, one might apply descriptive codes, summarizing what is contained in the interviews. It is rarely desirable to stop at this point, however, because coding is a tool to move from describing the data to interpreting the data. Suppose the researcher is pursuing some version of thematic analysis. In that case, it might be that the objects of coding are aspects of reported action, emotions, opinions, norms, relationships, routines, agreement/disagreement and change over time. A discourse analysis might instead code for different types of speech acts, tropes, linguistic or rhetorical devices. Multiple types of code might be generated within the same research project. What is important is that researchers are aware of the choices they are making in terms of what they are coding for. Moreover, through the process of refinement, the aim is to produce a set of discrete codes — in which codes are conceptually distinct, as opposed to overlapping. By using the same codes across the dataset, the researcher can capture commonalities across the interviews. This process of refinement involves relabelling codes and reorganizing how and where they are applied in the dataset.

From coding to analysis and writing

Data analysis is also an iterative process in which researchers move closer to and further away from the data. As they move away from the data, they synthesize their findings, thus honing and articulating their analytic insights. As they move closer to the data, they ground these insights in what is contained in the interviews. The link should not be broken between the data themselves and higher-order conceptual insights or claims being made. Researchers must be able to show evidence for their claims in the data. Figure  2 summarizes this iterative process and suggests the sorts of activities involved at each stage more concretely.

As well as going through steps 1 to 6 in order, the researcher will also go backwards and forwards between stages. Some stages will themselves be a forwards and backwards processing of coding and refining when working across different interview transcripts.

At the stage of synthesizing, there are some common quandaries. When dealing with a dataset consisting of multiple interviews, there will be salient and minority statements across different participants, or consensus or dissent on topics of interest to the researcher. A strength of qualitative interviews is that we can build in these nuances and variations across our data as opposed to aggregating them away. When exploring and reporting data, researchers should be asking how different findings are patterned and which interviews contain which codes, themes or tropes. Researchers should think about how these variations fit within the longer flow of individual interviews and what these variations tell them about the nature of their substantive research interests.

A further consideration is how to approach analysis within and across interview data. Researchers may look at one individual code, to examine the forms it takes across different participants and what they might be able to summarize about this code in the round. Alternatively, they might look at how a code or set of codes pattern across the account of one participant, to understand the code(s) in a more contextualized way. Further analysis might be done according to different sampling characteristics, where researchers group together interviews based on certain demographic characteristics and explore these together.

When it comes to writing up and presenting interview data, key considerations tend to rest on what is often termed transparency. When presenting the findings of an interview-based study, the reader should be able to understand and trace what the stated findings are based upon. This process typically involves describing the analytic process, how key decisions were made and presenting direct excerpts from the data. It is important to account for how the interview was set up and to consider the active part that the researcher has played in generating the data 32 . Quotes from interviews should not be thought of as merely embellishing or adding interest to a final research output. Rather, quotes serve the important function of connecting the reader directly to the underlying data. Quotes, therefore, should be chosen because they provide the reader with the most apt insight into what is being discussed. It is good practice to report not just on what participants said, but also on the questions that were asked to elicit the responses.

Researchers have increasingly used specialist qualitative data analysis software to organize and analyse their interview data, such as NVivo or ATLAS.ti. It is important to remember that such software is a tool for, rather than an approach or technique of, analysis. That said, software also creates a wide range of possibilities in terms of what can be done with the data. As researchers, we should reflect on how the range of possibilities of a given software package might be shaping our analytical choices and whether these are choices that we do indeed want to make.

Applications

This section reviews how and why in-depth interviews have been used by researchers studying gender, education and inequality, nationalism and ethnicity and the welfare state. Although interviews can be employed as a method of data collection in just about any social science topic, the applications below speak directly to the authors’ expertise and cutting-edge areas of research.

When it comes to the broad study of gender, in-depth interviews have been invaluable in shaping our understanding of how gender functions in everyday life. In a study of the US hedge fund industry (an industry dominated by white men), Tobias Neely was interested in understanding the factors that enable white men to prosper in the industry 33 . The study comprised interviews with 45 hedge fund workers and oversampled women of all races and men of colour to capture a range of experiences and beliefs. Tobias Neely found that practices of hiring, grooming and seeding are key to maintaining white men’s dominance in the industry. In terms of hiring, the interviews clarified that white men in charge typically preferred to hire people like themselves, usually from their extended networks. When women were hired, they were usually hired to less lucrative positions. In terms of grooming, Tobias Neely identifies how older and more senior men in the industry who have power and status will select one or several younger men as their protégés, to include in their own elite networks. Finally, in terms of her concept of seeding, Tobias Neely describes how older men who are hedge fund managers provide the seed money (often in the hundreds of millions of dollars) for a hedge fund to men, often their own sons (but not their daughters). These interviews provided an in-depth look into gendered and racialized mechanisms that allow white men to flourish in this industry.

Research by Rao draws on dozens of interviews with men and women who had lost their jobs, some of the participants’ spouses and follow-up interviews with about half the sample approximately 6 months after the initial interview 34 . Rao used interviews to understand the gendered experience and understanding of unemployment. Through these interviews, she found that the very process of losing their jobs meant different things for men and women. Women often saw job loss as being a personal indictment of their professional capabilities. The women interviewed often referenced how years of devaluation in the workplace coloured their interpretation of their job loss. Men, by contrast, were also saddened by their job loss, but they saw it as part and parcel of a weak economy rather than a personal failing. How these varied interpretations occurred was tied to men’s and women’s very different experiences in the workplace. Further, through her analysis of these interviews, Rao also showed how these gendered interpretations had implications for the kinds of jobs men and women sought to pursue after job loss. Whereas men remained tied to participating in full-time paid work, job loss appeared to be a catalyst pushing some of the women to re-evaluate their ties to the labour force.

In a study of workers in the tech industry, Hart used interviews to explain how individuals respond to unwanted and ambiguously sexual interactions 35 . Here, the researcher used interviews to allow participants to describe how these interactions made them feel and act and the logics of how they interpreted, classified and made sense of them 35 . Through her analysis of these interviews, Hart showed that participants engaged in a process she termed “trajectory guarding”, whereby they sought to monitor unwanted and ambiguously sexual interactions to avoid them from escalating. Yet, as Hart’s analysis proficiently demonstrates, these very strategies — which protect these workers sexually — also undermined their workplace advancement.

Drawing on interviews, these studies have helped us to understand better how gendered mechanisms, gendered interpretations and gendered interactions foster gender inequality when it comes to paid work. Methodologically, these studies illuminate the power of interviews to reveal important aspects of social life.

Nationalism and ethnicity

Traditionally, nationalism has been studied from a top-down perspective, through the lens of the state or using historical methods; in other words, in-depth interviews have not been a common way of collecting data to study nationalism. The methodological turn towards everyday nationalism has encouraged more scholars to go to the field and use interviews (and ethnography) to understand nationalism from the bottom up: how people talk about, give meaning, understand, navigate and contest their relation to nation, national identification and nationalism 36 , 37 , 38 , 39 . This turn has also addressed the gap left by those studying national and ethnic identification via quantitative methods, such as surveys.

Surveys can enumerate how individuals ascribe to categorical forms of identification 40 . However, interviews can question the usefulness of such categories and ask whether these categories are reflected, or resisted, by participants in terms of the meanings they give to identification 41 , 42 . Categories often pitch identification as a mutually exclusive choice; but identification might be more complex than such categories allow. For example, some might hybridize these categories or see themselves as moving between and across categories 43 . Hearing how people talk about themselves and their relation to nations, states and ethnicities, therefore, contributes substantially to the study of nationalism and national and ethnic forms of identification.

One particular approach to studying these topics, whether via everyday nationalism or alternatives, is that of using interviews to capture both articulations and narratives of identification, relations to nationalism and the boundaries people construct. For example, interviews can be used to gather self–other narratives by studying how individuals construct I–we–them boundaries 44 , including how participants talk about themselves, who participants include in their various ‘we’ groupings and which and how participants create ‘them’ groupings of others, inserting boundaries between ‘I/we’ and ‘them’. Overall, interviews hold great potential for listening to participants and understanding the nuances of identification and the construction of boundaries from their point of view.

Education and inequality

Scholars of social stratification have long noted that the school system often reproduces existing social inequalities. Carter explains that all schools have both material and sociocultural resources 45 . When children from different backgrounds attend schools with different material resources, their educational and occupational outcomes are likely to vary. Such material resources are relatively easy to measure. They are operationalized as teacher-to-student ratios, access to computers and textbooks and the physical infrastructure of classrooms and playgrounds.

Drawing on Bourdieusian theory 46 , Carter conceptualizes the sociocultural context as the norms, values and dispositions privileged within a social space 45 . Scholars have drawn on interviews with students and teachers (as well as ethnographic observations) to show how schools confer advantages on students from middle-class families, for example, by rewarding their help-seeking behaviours 47 . Focusing on race, researchers have revealed how schools can remain socioculturally white even as they enrol a racially diverse student population. In such contexts, for example, teachers often misrecognize the aesthetic choices made by students of colour, wrongly inferring that these students’ tastes in clothing and music reflect negative orientations to schooling 48 , 49 , 50 . These assessments can result in disparate forms of discipline and may ultimately shape educators’ assessments of students’ academic potential 51 .

Further, teachers and administrators tend to view the appropriate relationship between home and school in ways that resonate with white middle-class parents 52 . These parents are then able to advocate effectively for their children in ways that non-white parents are not 53 . In-depth interviews are particularly good at tapping into these understandings, revealing the mechanisms that confer privilege on certain groups of students and thereby reproduce inequality.

In addition, interviews can shed light on the unequal experiences that young people have within educational institutions, as the views of dominant groups are affirmed while those from disadvantaged backgrounds are delegitimized. For example, Teeger’s interviews with South African high schoolers showed how — because racially charged incidents are often framed as jokes in the broader school culture — Black students often feel compelled to ignore and keep silent about the racism they experience 54 . Interviews revealed that Black students who objected to these supposed jokes were coded by other students as serious or angry. In trying to avoid such labels, these students found themselves unable to challenge the racism they experienced. Interviews give us insight into these dynamics and help us see how young people understand and interpret the messages transmitted in schools — including those that speak to issues of inequality in their local school contexts as well as in society more broadly 24 , 55 .

The welfare state

In-depth interviews have also proved to be an important method for studying various aspects of the welfare state. By welfare state, we mean the social institutions relating to the economic and social wellbeing of a state’s citizens. Notably, using interviews has been useful to look at how policy design features are experienced and play out on the ground. Interviews have often been paired with large-scale surveys to produce mixed-methods study designs, therefore achieving both breadth and depth of insights.

In-depth interviews provide the opportunity to look behind policy assumptions or how policies are designed from the top down, to examine how these play out in the lives of those affected by the policies and whose experiences might otherwise be obscured or ignored. For example, the Welfare Conditionality project used interviews to critique the assumptions that conditionality (such as, the withdrawal of social security benefits if recipients did not perform or meet certain criteria) improved employment outcomes and instead showed that conditionality was harmful to mental health, living standards and had many other negative consequences 56 . Meanwhile, combining datasets from two small-scale interview studies with recipients allowed Summers and Young to critique assumptions around the simplicity that underpinned the design of Universal Credit in 2020, for example, showing that the apparently simple monthly payment design instead burdened recipients with additional money management decisions and responsibilities 57 .

Similarly, the Welfare at a (Social) Distance project used a mixed-methods approach in a large-scale study that combined national surveys with case studies and in-depth interviews to investigate the experience of claiming social security benefits during the COVID-19 pandemic. The interviews allowed researchers to understand in detail any issues experienced by recipients of benefits, such as delays in the process of claiming, managing on a very tight budget and navigating stigma and claiming 58 .

These applications demonstrate the multi-faceted topics and questions for which interviews can be a relevant method for data collection. These applications highlight not only the relevance of interviews, but also emphasize the key added value of interviews, which might be missed by other methods (surveys, in particular). Interviews can expose and question what is taken for granted and directly engage with communities and participants that might otherwise be ignored, obscured or marginalized.

Reproducibility and data deposition

There is a robust, ongoing debate about reproducibility in qualitative research, including interview studies. In some research paradigms, reproducibility can be a way of interrogating the rigour and robustness of research claims, by seeing whether these hold up when the research process is repeated. Some scholars have suggested that although reproducibility may be challenging, researchers can facilitate it by naming the place where the research was conducted, naming participants, sharing interview and fieldwork transcripts (anonymized and de-identified in cases where researchers are not naming people or places) and employing fact-checkers for accuracy 11 , 59 , 60 .

In addition to the ethical concerns of whether de-anonymization is ever feasible or desirable, it is also important to address whether the replicability of interview studies is meaningful. For example, the flexibility of interviews allows for the unexpected and the unforeseen to be incorporated into the scope of the research 61 . However, this flexibility means that we cannot expect reproducibility in the conventional sense, given that different researchers will elicit different types of data from participants. Sharing interview transcripts with other researchers, for instance, downplays the contextual nature of an interview.

Drawing on Bauer and Gaskell, we propose several measures to enhance rigour in qualitative research: transparency, grounding interpretations and aiming for theoretical transferability and significance 62 .

Researchers should be transparent when describing their methodological choices. Transparency means documenting who was interviewed, where and when (without requiring de-anonymization, for example, by documenting their characteristics), as well as the questions they were asked. It means carefully considering who was left out of the interviews and what that could mean for the researcher’s findings. It also means carefully considering who the researcher is and how their identity shaped the research process (integrating and articulating reflexivity into whatever is written up).

Second, researchers should ground their interpretations in the data. Grounding means presenting the evidence upon which the interpretation relies. Quotes and extracts should be extensive enough to allow the reader to evaluate whether the researcher’s interpretations are grounded in the data. At each step, researchers should carefully compare their own explanations and interpretations with alternative explanations. Doing so systematically and frequently allows researchers to become more confident in their claims. Here, researchers should justify the link between data and analysis by using quotes to justify and demonstrate the analytical point, while making sure the analytical point offers an interpretation of quotes (Box  4 ).

An important step in considering alternative explanations is to seek out disconfirming evidence 4 , 63 . This involves looking for instances where participants deviate from what the majority are saying and thus bring into question the theory (or explanation) that the researcher is developing. Careful analysis of such examples can often demonstrate the salience and meaning of what appears to be the norm (see Table  2 for examples) 54 . Considering alternative explanations and paying attention to disconfirming evidence allows the researcher to refine their own theories in respect of the data.

Finally, researchers should aim for theoretical transferability and significance in their discussions of findings. One way to think about this is to imagine someone who is not interested in the empirical study. Articulating theoretical transferability and significance usually takes the form of broadening out from the specific findings to consider explicitly how the research has refined or altered prior theoretical approaches. This process also means considering under what other conditions, aside from those of the study, the researcher thinks their theoretical revision would be supported by and why. Importantly, it also includes thinking about the limitations of one’s own approach and where the theoretical implications of the study might not hold.

Box 4 An example of grounding interpretations in data (from Rao 34 )

In an article explaining how unemployed men frame their job loss as a pervasive experience, Rao writes the following: “Unemployed men in this study understood unemployment to be an expected aspect of paid work in the contemporary United States. Robert, a white unemployed communications professional, compared the economic landscape after the Great Recession with the tragic events of September 11, 2001:

Part of your post-9/11 world was knowing people that died as a result of terrorism. The same thing is true with the [Great] Recession, right? … After the Recession you know somebody who was unemployed … People that really should be working.

The pervasiveness of unemployment rendered it normal, as Robert indicates.”

Here, the link between the quote presented and the analytical point Rao is making is clear: the analytical point is grounded in a quote and an interpretation of the quote is offered 34 .

Limitations and optimizations

When deciding which research method to use, the key question is whether the method provides a good fit for the research questions posed. In other words, researchers should consider whether interviews will allow them to successfully access the social phenomena necessary to answer their question(s) and whether the interviews will do so more effectively than other methods. Table  3 summarizes the major strengths and limitations of interviews. However, the accompanying text below is organized around some key issues, where relative strengths and weaknesses are presented alongside each other, the aim being that readers should think about how these can be balanced and optimized in relation to their own research.

Breadth versus depth of insight

Achieving an overall breadth of insight, in a statistically representative sense, is not something that is possible or indeed desirable when conducting in-depth interviews. Instead, the strength of conducting interviews lies in their ability to generate various sorts of depth of insight. The experiences or views of participants that can be accessed by conducting interviews help us to understand participants’ subjective realities. The challenge, therefore, is for researchers to be clear about why depth of insight is the focus and what we should aim to glean from these types of insight.

Naturalistic or artificial interviews

Interviews make use of a form of interaction with which people are familiar 64 . By replicating a naturalistic form of interaction as a tool to gather social science data, researchers can capitalize on people’s familiarity and expectations of what happens in a conversation. This familiarity can also be a challenge, as people come to the interview with preconceived ideas about what this conversation might be for or about. People may draw on experiences of other similar conversations when taking part in a research interview (for example, job interviews, therapy sessions, confessional conversations, chats with friends). Researchers should be aware of such potential overlaps and think through their implications both in how the aims and purposes of the research interview are communicated to participants and in how interview data are interpreted.

Further, some argue that a limitation of interviews is that they are an artificial form of data collection. By taking people out of their daily lives and asking them to stand back and pass comment, we are creating a distance that makes it difficult to use such data to say something meaningful about people’s actions, experiences and views. Other approaches, such as ethnography, might be more suitable for tapping into what people actually do, as opposed to what they say they do 65 .

Dynamism and replicability

Interviews following a semi-structured format offer flexibility both to the researcher and the participant. As the conversation develops, the interlocutors can explore the topics raised in much more detail, if desired, or pass over ones that are not relevant. This flexibility allows for the unexpected and the unforeseen to be incorporated into the scope of the research.

However, this flexibility has a related challenge of replicability. Interviews cannot be reproduced because they are contingent upon the interaction between the researcher and the participant in that given moment of interaction. In some research paradigms, replicability can be a way of interrogating the robustness of research claims, by seeing whether they hold when they are repeated. This is not a useful framework to bring to in-depth interviews and instead quality criteria (such as transparency) tend to be employed as criteria of rigour.

Accessing the private and personal

Interviews have been recognized for their strength in accessing private, personal issues, which participants may feel more comfortable talking about in a one-to-one conversation. Furthermore, interviews are likely to take a more personable form with their extended questions and answers, perhaps making a participant feel more at ease when discussing sensitive topics in such a context. There is a similar, but separate, argument made about accessing what are sometimes referred to as vulnerable groups, who may be difficult to make contact with using other research methods.

There is an associated challenge of anonymity. There can be types of in-depth interview that make it particularly challenging to protect the identities of participants, such as interviewing within a small community, or multiple members of the same household. The challenge to ensure anonymity in such contexts is even more important and difficult when the topic of research is of a sensitive nature or participants are vulnerable.

Increasingly, researchers are collaborating in large-scale interview-based studies and integrating interviews into broader mixed-methods designs. At the same time, interviews can be seen as an old-fashioned (and perhaps outdated) mode of data collection. We review these debates and discussions and point to innovations in interview-based studies. These include the shift from face-to-face interviews to the use of online platforms, as well as integrating and adapting interviews towards more inclusive methodologies.

Collaborating and mixing

Qualitative researchers have long worked alone 66 . Increasingly, however, researchers are collaborating with others for reasons such as efficiency, institutional incentives (for example, funding for collaborative research) and a desire to pool expertise (for example, studying similar phenomena in different contexts 67 or via different methods). Collaboration can occur across disciplines and methods, cases and contexts and between industry/business, practitioners and researchers. In many settings and contexts, collaboration has become an imperative 68 .

Cheek notes how collaboration provides both advantages and disadvantages 68 . For example, collaboration can be advantageous, saving time and building on the divergent knowledge, skills and resources of different researchers. Scholars with different theoretical or case-based knowledge (or contacts) can work together to build research that is comparative and/or more than the sum of its parts. But such endeavours also carry with them practical and political challenges in terms of how resources might actually be pooled, shared or accounted for. When undertaking such projects, as Morse notes, it is worth thinking about the nature of the collaboration and being explicit about such a choice, its advantages and its disadvantages 66 .

A further tension, but also a motivation for collaboration, stems from integrating interviews as a method in a mixed-methods project, whether with other qualitative researchers (to combine with, for example, focus groups, document analysis or ethnography) or with quantitative researchers (to combine with, for example, surveys, social media analysis or big data analysis). Cheek and Morse both note the pitfalls of collaboration with quantitative researchers: that quality of research may be sacrificed, qualitative interpretations watered down or not taken seriously, or tensions experienced over the pace and different assumptions that come with different methods and approaches of research 66 , 68 .

At the same time, there can be real benefits of such mixed-methods collaboration, such as reaching different and more diverse audiences or testing assumptions and theories between research components in the same project (for example, testing insights from prior quantitative research via interviews, or vice versa), as long as the skillsets of collaborators are seen as equally beneficial to the project. Cheek provides a set of questions that, as a starting point, can be useful for guiding collaboration, whether mixed methods or otherwise. First, Cheek advises asking all collaborators about their assumptions and understandings concerning collaboration. Second, Cheek recommends discussing what each perspective highlights and focuses on (and conversely ignores or sidelines) 68 .

A different way to engage with the idea of collaboration and mixed methods research is by fostering greater collaboration between researchers in the Global South and Global North, thus reversing trends of researchers from the Global North extracting knowledge from the Global South 69 . Such forms of collaboration also align with interview innovations, discussed below, that seek to transform traditional interview approaches into more participatory and inclusive (as part of participatory methodologies).

Digital innovations and challenges

The ongoing COVID-19 pandemic has centred the question of technology within interview-based fieldwork. Although conducting synchronous oral interviews online — for example, via Zoom, Skype or other such platforms — has been a method used by a small constituency of researchers for many years, it became (and remains) a necessity for many researchers wanting to continue or start interview-based projects while COVID-19 prevents face-to-face data collection.

In the past, online interviews were often framed as an inferior form of data collection for not providing the kinds of (often necessary) insights and forms of immersion face-to-face interviews allow 70 , 71 . Online interviews do tend to be more decontextualized than interviews conducted face-to-face 72 . For example, it is harder to recognize, engage with and respond to non-verbal cues 71 . At the same time, they broaden participation to those who might not have been able to access or travel to sites where interviews would have been conducted otherwise, for example people with disabilities. Online interviews also offer more flexibility in terms of scheduling and time requirements. For example, they provide more flexibility around precarious employment or caring responsibilities without having to travel and be away from home. In addition, online interviews might also reduce discomfort between researchers and participants, compared with face-to-face interviews, enabling more discussion of sensitive material 71 . They can also provide participants with more control, enabling them to turn on and off the microphone and video as they choose, for example, to provide more time to reflect and disconnect if they so wish 72 .

That said, online interviews can also introduce new biases based on access to technology 72 . For example, in the Global South, there are often urban/rural and gender gaps between who has access to mobile phones and who does not, meaning that some population groups might be overlooked unless researchers sample mindfully 71 . There are also important ethical considerations when deciding between online and face-to-face interviews. Online interviews might seem to imply lower ethical risks than face-to-face interviews (for example, they lower the chances of identification of participants or researchers), but they also offer more barriers to building trust between researchers and participants 72 . Interacting only online with participants might not provide the information needed to assess risk, for example, participants’ access to a private space to speak 71 . Just because online interviews might be more likely to be conducted in private spaces does not mean that private spaces are safe, for example, for victims of domestic violence. Finally, online interviews prompt further questions about decolonizing research and engaging with participants if research is conducted from afar 72 , such as how to include participants meaningfully and challenge dominant assumptions while doing so remotely.

A further digital innovation, modulating how researchers conduct interviews and the kinds of data collected and analysed, stems from the use and integration of (new) technology, such as WhatsApp text or voice notes to conduct synchronous or asynchronous oral or written interviews 73 . Such methods can provide more privacy, comfort and control to participants and make recruitment easier, allowing participants to share what they want when they want to, using technology that already forms a part of their daily lives, especially for young people 74 , 75 . Such technology is also emerging in other qualitative methods, such as focus groups, with similar arguments around greater inclusivity versus traditional offline modes. Here, the digital challenge might be higher for researchers than for participants if they are less used to such technology 75 . And while there might be concerns about the richness, depth and quality of written messages as a form of interview data, Gibson reports that the reams of transcripts that resulted from a study using written messaging were dense with meaning to be analysed 75 .

Like with online and face-to-face interviews, it is important also to consider the ethical questions and challenges of using such technology, from gaining consent to ensuring participant safety and attending to their distress, without cues, like crying, that might be more obvious in a face-to-face setting 75 , 76 . Attention to the platform used for such interviews is also important and researchers should be attuned to the local and national context. For example, in China, many platforms are neither legal nor available 76 . There, more popular platforms — like WeChat — can be highly monitored by the government, posing potential risks to participants depending on the topic of the interview. Ultimately, researchers should consider trade-offs between online and offline interview modalities, being attentive to the social context and power dynamics involved.

The next 5–10 years

Continuing to integrate (ethically) this technology will be among the major persisting developments in interview-based research, whether to offer more flexibility to researchers or participants, or to diversify who can participate and on what terms.

Pushing the idea of inclusion even further is the potential for integrating interview-based studies within participatory methods, which are also innovating via integrating technology. There is no hard and fast line between researchers using in-depth interviews and participatory methods; many who employ participatory methods will use interviews at the beginning, middle or end phases of a research project to capture insights, perspectives and reflections from participants 77 , 78 . Participatory methods emphasize the need to resist existing power and knowledge structures. They broaden who has the right and ability to contribute to academic knowledge by including and incorporating participants not only as subjects of data collection, but as crucial voices in research design and data analysis 77 . Participatory methods also seek to facilitate local change and to produce research materials, whether for academic or non-academic audiences, including films and documentaries, in collaboration with participants.

In responding to the challenges of COVID-19, capturing the fraught situation wrought by the pandemic and the momentum to integrate technology, participatory researchers have sought to continue data collection from afar. For example, Marzi has adapted an existing project to co-produce participatory videos, via participants’ smartphones in Medellin, Colombia, alongside regular check-in conversations/meetings/interviews with participants 79 . Integrating participatory methods into interview studies offers a route by which researchers can respond to the challenge of diversifying knowledge, challenging assumptions and power hierarchies and creating more inclusive and collaborative partnerships between participants and researchers in the Global North and South.

Brinkmann, S. & Kvale, S. Doing Interviews Vol. 2 (Sage, 2018). This book offers a good general introduction to the practice and design of interview-based studies.

Silverman, D. A Very Short, Fairly Interesting And Reasonably Cheap Book About Qualitative Research (Sage, 2017).

Yin, R. K. Case Study Research And Applications: Design And Methods (Sage, 2018).

Small, M. L. How many cases do I need?’ On science and the logic of case selection in field-based research. Ethnography 10 , 5–38 (2009). This article convincingly demonstrates how the logic of qualitative research differs from quantitative research and its goal of representativeness.

Google Scholar  

Gerson, K. & Damaske, S. The Science and Art of Interviewing (Oxford Univ. Press, 2020).

Glaser, B. G. & Strauss, A. L. The Discovery Of Grounded Theory: Strategies For Qualitative Research (Aldine, 1967).

Braun, V. & Clarke, V. To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales. Qual. Res. Sport Exerc. Health 13 , 201–216 (2021).

Guest, G., Bunce, A. & Johnson, L. How many interviews are enough? An experiment with data saturation and variability. Field Methods 18 , 59–82 (2006).

Vasileiou, K., Barnett, J., Thorpe, S. & Young, T. Characterising and justifying sample size sufficiency in interview-based studies: systematic analysis of qualitative health research over a 15-year period. BMC Med. Res. Methodol. 18 , 148 (2018).

Silverman, D. How was it for you? The Interview Society and the irresistible rise of the (poorly analyzed) interview. Qual. Res. 17 , 144–158 (2017).

Jerolmack, C. & Murphy, A. The ethical dilemmas and social scientific tradeoffs of masking in ethnography. Sociol. Methods Res. 48 , 801–827 (2019).

MathSciNet   Google Scholar  

Reyes, V. Ethnographic toolkit: strategic positionality and researchers’ visible and invisible tools in field research. Ethnography 21 , 220–240 (2020).

Guillemin, M. & Gillam, L. Ethics, reflexivity and “ethically important moments” in research. Qual. Inq. 10 , 261–280 (2004).

Summers, K. For the greater good? Ethical reflections on interviewing the ‘rich’ and ‘poor’ in qualitative research. Int. J. Soc. Res. Methodol. 23 , 593–602 (2020). This article argues that, in qualitative interview research, a clearer distinction needs to be drawn between ethical commitments to individual research participants and the group(s) to which they belong, a distinction that is often elided in existing ethics guidelines.

Yusupova, G. Exploring sensitive topics in an authoritarian context: an insider perspective. Soc. Sci. Q. 100 , 1459–1478 (2019).

Hemming, J. in Surviving Field Research: Working In Violent And Difficult Situations 21–37 (Routledge, 2009).

Murphy, E. & Dingwall, R. Informed consent, anticipatory regulation and ethnographic practice. Soc. Sci. Med. 65 , 2223–2234 (2007).

Kostovicova, D. & Knott, E. Harm, change and unpredictability: the ethics of interviews in conflict research. Qual. Res. 22 , 56–73 (2022). This article highlights how interviews need to be considered as ethically unpredictable moments where engaging with change among participants can itself be ethical.

Andersson, R. Illegality, Inc.: Clandestine Migration And The Business Of Bordering Europe (Univ. California Press, 2014).

Ellis, R. What do we mean by a “hard-to-reach” population? Legitimacy versus precarity as barriers to access. Sociol. Methods Res. https://doi.org/10.1177/0049124121995536 (2021).

Article   Google Scholar  

Braun, V. & Clarke, V. Thematic Analysis: A Practical Guide (Sage, 2022).

Alejandro, A. & Knott, E. How to pay attention to the words we use: the reflexive review as a method for linguistic reflexivity. Int. Stud. Rev. https://doi.org/10.1093/isr/viac025 (2022).

Alejandro, A., Laurence, M. & Maertens, L. in International Organisations and Research Methods: An Introduction (eds Badache, F., Kimber, L. R. & Maertens, L.) (Michigan Univ. Press, in the press).

Teeger, C. “Both sides of the story” history education in post-apartheid South Africa. Am. Sociol. Rev. 80 , 1175–1200 (2015).

Crotty, M. The Foundations Of Social Research: Meaning And Perspective In The Research Process (Routledge, 2020).

Potter, J. & Hepburn, A. Qualitative interviews in psychology: problems and possibilities. Qual. Res. Psychol. 2 , 281–307 (2005).

Taylor, S. What is Discourse Analysis? (Bloomsbury Publishing, 2013).

Riessman, C. K. Narrative Analysis (Sage, 1993).

Corbin, J. M. & Strauss, A. Grounded theory research: Procedures, canons and evaluative criteria. Qual. Sociol. 13 , 3–21 (1990).

Timmermans, S. & Tavory, I. Theory construction in qualitative research: from grounded theory to abductive analysis. Sociol. Theory 30 , 167–186 (2012).

Fereday, J. & Muir-Cochrane, E. Demonstrating rigor using thematic analysis: a hybrid approach of inductive and deductive coding and theme development. Int. J. Qual. Meth. 5 , 80–92 (2006).

Potter, J. & Hepburn, A. Eight challenges for interview researchers. Handb. Interview Res. 2 , 541–570 (2012).

Tobias Neely, M. Fit to be king: how patrimonialism on Wall Street leads to inequality. Socioecon. Rev. 16 , 365–385 (2018).

Rao, A. H. Gendered interpretations of job loss and subsequent professional pathways. Gend. Soc. 35 , 884–909 (2021). This article used interview data from unemployed men and women to illuminate how job loss becomes a pivotal moment shaping men’s and women’s orientation to paid work, especially in terms of curtailing women’s participation in paid work.

Hart, C. G. Trajectory guarding: managing unwanted, ambiguously sexual interactions at work. Am. Sociol. Rev. 86 , 256–278 (2021).

Goode, J. P. & Stroup, D. R. Everyday nationalism: constructivism for the masses. Soc. Sci. Q. 96 , 717–739 (2015).

Antonsich, M. The ‘everyday’ of banal nationalism — ordinary people’s views on Italy and Italian. Polit. Geogr. 54 , 32–42 (2016).

Fox, J. E. & Miller-Idriss, C. Everyday nationhood. Ethnicities 8 , 536–563 (2008).

Yusupova, G. Cultural nationalism and everyday resistance in an illiberal nationalising state: ethnic minority nationalism in Russia. Nations National. 24 , 624–647 (2018).

Kiely, R., Bechhofer, F. & McCrone, D. Birth, blood and belonging: identity claims in post-devolution Scotland. Sociol. Rev. 53 , 150–171 (2005).

Brubaker, R. & Cooper, F. Beyond ‘identity’. Theory Soc. 29 , 1–47 (2000).

Brubaker, R. Ethnicity Without Groups (Harvard Univ. Press, 2004).

Knott, E. Kin Majorities: Identity And Citizenship In Crimea And Moldova From The Bottom-Up (McGill Univ. Press, 2022).

Bucher, B. & Jasper, U. Revisiting ‘identity’ in international relations: from identity as substance to identifications in action. Eur. J. Int. Relat. 23 , 391–415 (2016).

Carter, P. L. Stubborn Roots: Race, Culture And Inequality In US And South African Schools (Oxford Univ. Press, 2012).

Bourdieu, P. in Cultural Theory: An Anthology Vol. 1, 81–93 (eds Szeman, I. & Kaposy, T.) (Wiley-Blackwell, 2011).

Calarco, J. M. Negotiating Opportunities: How The Middle Class Secures Advantages In School (Oxford Univ. Press, 2018).

Carter, P. L. Keepin’ It Real: School Success Beyond Black And White (Oxford Univ. Press, 2005).

Carter, P. L. ‘Black’ cultural capital, status positioning and schooling conflicts for low-income African American youth. Soc. Probl. 50 , 136–155 (2003).

Warikoo, N. K. The Diversity Bargain Balancing Acts: Youth Culture in the Global City (Univ. California Press, 2011).

Morris, E. W. “Tuck in that shirt!” Race, class, gender and discipline in an urban school. Sociol. Perspect. 48 , 25–48 (2005).

Lareau, A. Social class differences in family–school relationships: the importance of cultural capital. Sociol. Educ. 60 , 73–85 (1987).

Warikoo, N. Addressing emotional health while protecting status: Asian American and white parents in suburban America. Am. J. Sociol. 126 , 545–576 (2020).

Teeger, C. Ruptures in the rainbow nation: how desegregated South African schools deal with interpersonal and structural racism. Sociol. Educ. 88 , 226–243 (2015). This article leverages ‘ deviant ’ cases in an interview study with South African high schoolers to understand why the majority of participants were reluctant to code racially charged incidents at school as racist.

Ispa-Landa, S. & Conwell, J. “Once you go to a white school, you kind of adapt” black adolescents and the racial classification of schools. Sociol. Educ. 88 , 1–19 (2015).

Dwyer, P. J. Punitive and ineffective: benefit sanctions within social security. J. Soc. Secur. Law 25 , 142–157 (2018).

Summers, K. & Young, D. Universal simplicity? The alleged simplicity of Universal Credit from administrative and claimant perspectives. J. Poverty Soc. Justice 28 , 169–186 (2020).

Summers, K. et al. Claimants’ Experiences Of The Social Security System During The First Wave Of COVID-19 . https://www.distantwelfare.co.uk/winter-report (2021).

Desmond, M. Evicted: Poverty And Profit In The American City (Crown Books, 2016).

Reyes, V. Three models of transparency in ethnographic research: naming places, naming people and sharing data. Ethnography 19 , 204–226 (2018).

Robson, C. & McCartan, K. Real World Research (Wiley, 2016).

Bauer, M. W. & Gaskell, G. Qualitative Researching With Text, Image And Sound: A Practical Handbook (SAGE, 2000).

Lareau, A. Listening To People: A Practical Guide To Interviewing, Participant Observation, Data Analysis And Writing It All Up (Univ. Chicago Press, 2021).

Lincoln, Y. S. & Guba, E. G. Naturalistic Inquiry (Sage, 1985).

Jerolmack, C. & Khan, S. Talk is cheap. Sociol. Methods Res. 43 , 178–209 (2014).

Morse, J. M. Styles of collaboration in qualitative inquiry. Qual. Health Res. 18 , 3–4 (2008).

ADS   Google Scholar  

Lamont, M. et al. Getting Respect: Responding To Stigma And Discrimination In The United States, Brazil And Israel (Princeton Univ. Press, 2016).

Cheek, J. Researching collaboratively: implications for qualitative research and researchers. Qual. Health Res. 18 , 1599–1603 (2008).

Botha, L. Mixing methods as a process towards indigenous methodologies. Int. J. Soc. Res. Methodol. 14 , 313–325 (2011).

Howlett, M. Looking at the ‘field’ through a zoom lens: methodological reflections on conducting online research during a global pandemic. Qual. Res. https://doi.org/10.1177/1468794120985691 (2021).

Reñosa, M. D. C. et al. Selfie consents, remote rapport and Zoom debriefings: collecting qualitative data amid a pandemic in four resource-constrained settings. BMJ Glob. Health 6 , e004193 (2021).

Mwambari, D., Purdeková, A. & Bisoka, A. N. Covid-19 and research in conflict-affected contexts: distanced methods and the digitalisation of suffering. Qual. Res. https://doi.org/10.1177/1468794121999014 (2021).

Colom, A. Using WhatsApp for focus group discussions: ecological validity, inclusion and deliberation. Qual. Res. https://doi.org/10.1177/1468794120986074 (2021).

Kaufmann, K. & Peil, C. The mobile instant messaging interview (MIMI): using WhatsApp to enhance self-reporting and explore media usage in situ. Mob. Media Commun. 8 , 229–246 (2020).

Gibson, K. Bridging the digital divide: reflections on using WhatsApp instant messenger interviews in youth research. Qual. Res. Psychol. 19 , 611–631 (2020).

Lawrence, L. Conducting cross-cultural qualitative interviews with mainland Chinese participants during COVID: lessons from the field. Qual. Res. https://doi.org/10.1177/1468794120974157 (2020).

Ponzoni, E. Windows of understanding: broadening access to knowledge production through participatory action research. Qual. Res. 16 , 557–574 (2016).

Kong, T. S. Gay and grey: participatory action research in Hong Kong. Qual. Res. 18 , 257–272 (2018).

Marzi, S. Participatory video from a distance: co-producing knowledge during the COVID-19 pandemic using smartphones. Qual. Res. https://doi.org/10.1177/14687941211038171 (2021).

Kvale, S. & Brinkmann, S. InterViews: Learning The Craft Of Qualitative Research Interviewing (Sage, 2008).

Rao, A. H. The ideal job-seeker norm: unemployment and marital privileges in the professional middle-class. J. Marriage Fam. 83 , 1038–1057 (2021).

Rivera, L. A. Ivies, extracurriculars and exclusion: elite employers’ use of educational credentials. Res. Soc. Stratif. Mobil. 29 , 71–90 (2011).

Download references

Acknowledgements

The authors are grateful to the MY421 team and students for prompting how best to frame and communicate issues pertinent to in-depth interview studies.

Author information

Authors and affiliations.

Department of Methodology, London School of Economics, London, UK

Eleanor Knott, Aliya Hamid Rao, Kate Summers & Chana Teeger

You can also search for this author in PubMed   Google Scholar

Contributions

The authors contributed equally to all aspects of the article.

Corresponding author

Correspondence to Eleanor Knott .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature Reviews Methods Primers thanks Jonathan Potter and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

A pre-written interview outline for a semi-structured interview that provides both a topic structure and the ability to adapt flexibly to the content and context of the interview and the interaction between the interviewer and participant. Others may refer to the topic guide as an interview protocol.

Here we refer to the participants that take part in the study as the sample. Other researchers may refer to the participants as a participant group or dataset.

This involves dividing a population into smaller groups based on particular characteristics, for example, age or gender, and then sampling randomly within each group.

A sampling method where the guiding logic when deciding who to recruit is to achieve the most relevant participants for the research topic, in terms of being rich in information or insights.

Researchers ask participants to introduce the researcher to others who meet the study’s inclusion criteria.

Similar to stratified sampling, but participants are not necessarily randomly selected. Instead, the researcher determines how many people from each category of participants should be recruited. Recruitment can happen via snowball or purposive sampling.

A method for developing, analysing and interpreting patterns across data by coding in order to develop themes.

An approach that interrogates the explicit, implicit and taken-for-granted dimensions of language as well as the contexts in which it is articulated to unpack its purposes and effects.

A form of transcription that simplifies what has been said by removing certain verbal and non-verbal details that add no further meaning, such as ‘ums and ahs’ and false starts.

The analytic framework, theoretical approach and often hypotheses, are developed prior to examining the data and then applied to the dataset.

The analytic framework and theoretical approach is developed from analysing the data.

An approach that combines deductive and inductive components to work recursively by going back and forth between data and existing theoretical frameworks (also described as an iterative approach). This approach is increasingly recognized not only as a more realistic but also more desirable third alternative to the more traditional inductive versus deductive binary choice.

A theoretical apparatus that emphasizes the role of cultural processes and capital in (intergenerational) social reproduction.

Rights and permissions

Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Cite this article.

Knott, E., Rao, A.H., Summers, K. et al. Interviews in the social sciences. Nat Rev Methods Primers 2 , 73 (2022). https://doi.org/10.1038/s43586-022-00150-6

Download citation

Accepted : 14 July 2022

Published : 15 September 2022

DOI : https://doi.org/10.1038/s43586-022-00150-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

‘life minus illness = recovery’: a phenomenological study about experiences and meanings of recovery among individuals with serious mental illness from southern india.

• Srishti Hegde
• Shalini Quadros
• Vinita A. Acharya

Community Mental Health Journal (2024)

Is e-business breaking down barriers for Bangladesh’s young female entrepreneurs during the COVID-19 pandemic? A qualitative study

• Md. Fouad Hossain Sarker
• Sayed Farrukh Ahmed
• Md. Salman Sohel

SN Social Sciences (2024)

Between the dog and the wolf: an interpretative phenomenological analysis of bicultural, sexual minority people’s lived experiences

• Emelie Louise Miller
• Ingrid Zakrisson

Discover Psychology (2024)

Acknowledging that Men are Moral and Harmed by Gender Stereotypes Increases Men’s Willingness to Engage in Collective Action on Behalf of Women

• Alexandra Vázquez
• Lucía López-Rodríguez
• Marco Brambilla

Sex Roles (2024)

Supporting mid-career students’ psychological needs to improve motivation and retention in post-graduate courses

• Gillian Kirk
• Carly Sanbrook

The Australian Educational Researcher (2024)

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Semistructured interviewing in primary care research: a balance of relationship and rigour

Lisa M Vaughn .

https://doi.org/ 10.1136/fmch-2018-000057

Semistructured in-depth interviews are commonly used in qualitative research and are the most frequent qualitative data source in health services research. This method typically consists of a dialogue between researcher and participant, guided by a flexible interview protocol and supplemented by follow-up questions, probes and comments. The method allows the researcher to collect open-ended data, to explore participant thoughts, feelings and beliefs about a particular topic and to delve deeply into personal and sometimes sensitive issues. The purpose of this article was to identify and describe the essential skills to designing and conducting semistructured interviews in family medicine and primary care research settings. We reviewed the literature on semistructured interviewing to identify key skills and components for using this method in family medicine and primary care research settings. Overall, semistructured interviewing requires both a relational focus and practice in the skills of facilitation. Skills include: (1) determining the purpose and scope of the study; (2) identifying participants; (3) considering ethical issues; (4) planning logistical aspects; (5) developing the interview guide; (6) establishing trust and rapport; (7) conducting the interview; (8) memoing and reflection; (9) analysing the data; (10) demonstrating the trustworthiness of the research; and (11) presenting findings in a paper or report. Semistructured interviews provide an effective and feasible research method for family physicians to conduct in primary care research settings. Researchers using semistructured interviews for data collection should take on a relational focus and consider the skills of interviewing to ensure quality. Semistructured interviewing can be a powerful tool for family physicians, primary care providers and other health services researchers to use to understand the thoughts, beliefs and experiences of individuals. Despite the utility, semistructured interviews can be intimidating and challenging for researchers not familiar with qualitative approaches. In order to elucidate this method, we provide practical guidance for researchers, including novice researchers and those with few resources, to use semistructured interviewing as a data collection strategy. We provide recommendations for the essential steps to follow in order to best implement semistructured interviews in family medicine and primary care research settings.

• Introduction

Semistructured interviews can be used by family medicine researchers in clinical settings or academic settings even with few resources. In contrast to large-scale epidemiological studies, or even surveys, a family medicine researcher can conduct a highly meaningful project with interviews with as few as 8–12 participants. For example, Chang and her colleagues, all family physicians, conducted semistructured interviews with 10 providers to understand their perspectives on weight gain in pregnant patients. 1 The interviewers asked questions about providers’ overall perceptions on weight gain, their clinical approach to weight gain during pregnancy and challenges when managing weight gain among pregnant patients. Additional examples conducted by or with family physicians or in primary care settings are summarised in table 1 . 1–6

From our perspective as seasoned qualitative researchers, conducting effective semistructured interviews requires: (1) a relational focus, including active engagement and curiosity, and (2) practice in the skills of interviewing. First, a relational focus emphasises the unique relationship between interviewer and interviewee. To obtain quality data, interviews should not be conducted with a transactional question-answer approach but rather should be unfolding, iterative interactions between the interviewer and interviewee. Second, interview skills can be learnt. Some of us will naturally be more comfortable and skilful at conducting interviews but all aspects of interviews are learnable and through practice and feedback will improve. Throughout this article, we highlight strategies to balance relationship and rigour when conducting semistructured interviews in primary care and the healthcare setting.

Qualitative research interviews are ‘attempts to understand the world from the subjects’ point of view, to unfold the meaning of peoples’ experiences, to uncover their lived world prior to scientific explanations’ (p 1). 7 Qualitative research interviews unfold as an interviewer asks questions of the interviewee in order to gather subjective information about a particular topic or experience. Though the definitions and purposes of qualitative research interviews vary slightly in the literature, there is common emphasis on the experiences of interviewees and the ways in which the interviewee perceives the world (see table 2 for summary of definitions from seminal texts).

The most common type of interview used in qualitative research and the healthcare context is semistructured interview. 8 Figure 1 highlights the key features of this data collection method, which is guided by a list of topics or questions with follow-up questions, probes and comments. Typically, the sequencing and wording of the questions are modified by the interviewer to best fit the interviewee and interview context. Semistructured interviews can be conducted in multiple ways (ie, face to face, telephone, text/email, individual, group, brief, in-depth), each of which have advantages and disadvantages. We will focus on the most common form of semistructured interviews within qualitative research—individual, face-to-face, in-depth interviews.

Key characteristics of semistructured interviews.

Purpose of semistructured interviews

The overall purpose of using semistructured interviews for data collection is to gather information from key informants who have personal experiences, attitudes, perceptions and beliefs related to the topic of interest. Researchers can use semistructured interviews to collect new, exploratory data related to a research topic, triangulate other data sources or validate findings through member checking (respondent feedback about research results). 9 If using a mixed methods approach, semistructured interviews can also be used in a qualitative phase to explore new concepts to generate hypotheses or explain results from a quantitative phase that tests hypotheses. Semistructured interviews are an effective method for data collection when the researcher wants: (1) to collect qualitative, open-ended data; (2) to explore participant thoughts, feelings and beliefs about a particular topic; and (3) to delve deeply into personal and sometimes sensitive issues.

Designing and conducting semistructured interviews

In the following section, we provide recommendations for the steps required to carefully design and conduct semistructured interviews with emphasis on applications in family medicine and primary care research (see table 3 ).

Steps for designing and conducting semistructured interviews

Step 1: determining the purpose and scope of the study.

The purpose of the study is the primary objective of your project and may be based on an anecdotal experience, a review of the literature or previous research finding. The purpose is developed in response to an identified gap or problem that needs to be addressed.

Research questions are the driving force of a study because they are associated with every other aspect of the design. They should be succinct and clearly indicate that you are using a qualitative approach. Qualitative research questions typically start with ‘What’, ‘How’ or ‘Why’ and focus on the exploration of a single concept based on participant perspectives. 10

Step 2: identifying participants

After deciding on the purpose of the study and research question(s), the next step is to determine who will provide the best information to answer the research question. Good interviewees are those who are available, willing to be interviewed and have lived experiences and knowledge about the topic of interest. 11 12 Working with gatekeepers or informants to get access to potential participants can be extremely helpful as they are trusted sources that control access to the target sample.

Sampling strategies are influenced by the research question and the purpose of the study. Unlike quantitative studies, statistical representativeness is not the goal of qualitative research. There is no calculation of statistical power and the goal is not a large sample size. Instead, qualitative approaches seek an in-depth and detailed understanding and typically use purposeful sampling. See the study of Hatch for a summary of various types of purposeful sampling that can be used for interview studies. 12

‘How many participants are needed?’ The most common answer is, ‘it depends’—it depends on the purpose of the study, what kind of study is planned and what questions the study is trying to answer. 12–14 One common standard in qualitative sample sizes is reaching thematic saturation, which refers to the point at which no new thematic information is gathered from participants. Malterud and colleagues discuss the concept of information power , or a qualitative equivalent to statistical power, to determine how many interviews should be collected in a study. They suggest that the size of a sample should depend on the aim, homogeneity of the sample, theory, interview quality and analytic strategy. 14

Step 3: considering ethical issues

An ethical attitude should be present from the very beginning of the research project even before you decide who to interview. 15 This ethical attitude should incorporate respect, sensitivity and tact towards participants throughout the research process. Because semistructured interviewing often requires the participant to reveal sensitive and personal information directly to the interviewer, it is important to consider the power imbalance between the researcher and the participant. In healthcare settings, the interviewer or researcher may be a part of the patient’s healthcare team or have contact with the healthcare team. The researchers should ensure the interviewee that their participation and answers will not influence the care they receive or their relationship with their providers. Other issues to consider include: reducing the risk of harm; protecting the interviewee’s information; adequately informing interviewees about the study purpose and format; and reducing the risk of exploitation. 10

Step 4: planning logistical aspects

Careful planning particularly around the technical aspects of interviews can be the difference between a great interview and a not so great interview. During the preparation phase, the researcher will need to plan and make decisions about the best ways to contact potential interviewees, obtain informed consent, arrange interview times and locations convenient for both participant and researcher, and test recording equipment. Although many experienced researchers have found themselves conducting interviews in less than ideal locations, the interview location should avoid (or at least minimise) interruptions and be appropriate for the interview (quiet, private and able to get a clear recording). 16 For some research projects, the participants’ homes may make sense as the best interview location. 16

Initial contacts can be made through telephone or email and followed up with more details so the individual can make an informed decision about whether they wish to be interviewed. Potential participants should know what to expect in terms of length of time, purpose of the study, why they have been selected and who will be there. In addition, participants should be informed that they can refuse to answer questions or can withdraw from the study at any time, including during the interview itself.

Audio recording the interview is recommended so that the interviewer can concentrate on the interview and build rapport rather than being distracted with extensive note taking 16 (see table 4 for audio-recording tips). Participants should be informed that audio recording is used for data collection and that they can refuse to be audio recorded should they prefer.

Most researchers will want to have interviews transcribed verbatim from the audio recording. This allows you to refer to the exact words of participants during the analysis. Although it is possible to conduct analyses from the audio recordings themselves or from notes, it is not ideal. However, transcription can be extremely time consuming and, if not done yourself, can be costly.

In the planning phase of research, you will want to consider whether qualitative research software (eg, NVivo, ATLAS.ti, MAXQDA, Dedoose, and so on) will be used to assist with organising, managing and analysis. While these tools are helpful in the management of qualitative data, it is important to consider your research budget, the cost of the software and the learning curve associated with using a new system.

Step 5: developing the interview guide

Semistructured interviews include a short list of ‘guiding’ questions that are supplemented by follow-up and probing questions that are dependent on the interviewee’s responses. 8 17 All questions should be open ended, neutral, clear and avoid leading language. In addition, questions should use familiar language and avoid jargon.

Most interviews will start with an easy, context-setting question before moving to more difficult or in-depth questions. 17 Table 5 gives details of the types of guiding questions including ‘grand tour’ questions, 18 core questions and planned and unplanned follow-up questions.

To illustrate, online supplementary appendix A presents a sample interview guide from our study of weight gain during pregnancy among young women. We start with the prompt, ‘Tell me about how your pregnancy has been so far’ to initiate conversation about their thoughts and feelings during pregnancy. The subsequent questions will elicit responses to help answer our research question about young women’s perspectives related to weight gain during pregnancy.

After developing the guiding questions, it is important to pilot test the interview. Having a good sense of the guide helps you to pace the interview (and not run out of time), use a conversational tone and make necessary adjustments to the questions.

Like all qualitative research, interviewing is iterative in nature—data collection and analysis occur simultaneously, which may result in changes to the guiding questions as the study progresses. Questions that are not effective may be replaced with other questions and additional probes can be added to explore new topics that are introduced by participants in previous interviews. 10

Step 6: establishing trust and rapport

Interviews are a special form of relationship, where the interviewer and interviewee converse about important and often personal topics. The interviewer must build rapport quickly by listening attentively and respectfully to the information shared by the interviewee. 19 As the interview progresses, the interviewer must continue to demonstrate respect, encourage the interviewee to share their perspectives and acknowledge the sensitive nature of the conversation. 20

To establish rapport, it is important to be authentic and open to the interviewee’s point of view. It is possible that the participants you recruit for your study will have preconceived notions about research, which may include mistrust. As a result, it is important to describe why you are conducting the research and how their participation is meaningful. In an interview relationship, the interviewee is the expert and should be treated as such—you are relying on the interviewee to enhance your understanding and add to your research. Small behaviours that can enhance rapport include: dressing professionally but not overly formal; avoiding jargon or slang; and using a normal conversational tone. Because interviewees will be discussing their experience, having some awareness of contextual or cultural factors that may influence their perspectives may be helpful as background knowledge.

Step 7: conducting the interview

Location and set-up.

The interview should have already been scheduled at a convenient time and location for the interviewee. The location should be private, ideally with a closed door, rather than a public place. It is helpful if there is a room where you can speak privately without interruption, and where it is quiet enough to hear and audio record the interview. Within the interview space, Josselson 15 suggests an arrangement with a comfortable distance between the interviewer and interviewee with a low table in between for the recorder and any materials (consent forms, questionnaires, water, and so on).

Beginning the interview

Many interviewers start with chatting to break the ice and attempt to establish commonalities, rapport and trust. Most interviews will need to begin with a brief explanation of the research study, consent/assent procedures, rationale for talking to that particular interviewee and description of the interview format and agenda. 11 It can also be helpful if the interviewer shares a little about who they are and why they are interested in the topic. The recording equipment should have already been tested thoroughly but interviewers may want to double-check that the audio equipment is working and remind participants about the reason for recording.

Interviewer stance

During the interview, the interviewer should adopt a friendly and non-judgemental attitude. You will want to maintain a warm and conversational tone, rather than a rote, question-answer approach. It is important to recognise the potential power differential as a researcher. Conveying a sense of being in the interview together and that you as the interviewer are a person just like the interviewee can help ease any discomfort. 15

Active listening

During a face-to-face interview, there is an opportunity to observe social and non-verbal cues of the interviewee. These cues may come in the form of voice, body language, gestures and intonation, and can supplement the interviewee’s verbal response and can give clues to the interviewer about the process of the interview. 21 Listening is the key to successful interviewing. 22 Listening should be ‘attentive, empathic, nonjudgmental, listening in order to invite, and engender talk’ 15 15 (p 66). Silence, nods, smiles and utterances can also encourage further elaboration from the interviewee.

Continuing the interview

As the interview progresses, the interviewer can repeat the words used by the interviewee, use planned and unplanned follow-up questions that invite further clarification, exploration or elaboration. As DiCicco-Bloom and Crabtree 10 explain: ‘Throughout the interview, the goal of the interviewer is to encourage the interviewee to share as much information as possible, unselfconsciously and in his or her own words’ (p 317). Some interviewees are more forthcoming and will offer many details of their experiences without much probing required. Others will require prompting and follow-up to elicit sufficient detail.

As a result, follow-up questions are equally important to the core questions in a semistructured interview. Prompts encourage people to continue talking and they can elicit more details needed to understand the topic. Examples of verbal probes are repeating the participant’s words, summarising the main idea or expressing interest with verbal agreement. 8 11 See table 6 for probing techniques and example probes we have used in our own interviewing.

Step 8: memoing and reflection

After an interview, it is essential for the interviewer to begin to reflect on both the process and the content of the interview. During the actual interview, it can be difficult to take notes or begin reflecting. Even if you think you will remember a particular moment, you likely will not be able to recall each moment with sufficient detail. Therefore, interviewers should always record memos —notes about what you are learning from the data. 23 24 There are different approaches to recording memos: you can reflect on several specific ideas, or create a running list of thoughts. Memos are also useful for improving the quality of subsequent interviews.

Step 9: analysing the data

The data analysis strategy should also be developed during planning stages because analysis occurs concurrently with data collection. 25 The researcher will take notes, modify the data collection procedures and write reflective memos throughout the data collection process. This begins the process of data analysis.

The data analysis strategy used in your study will depend on your research question and qualitative design—see the study of Creswell for an overview of major qualitative approaches. 26 The general process for analysing and interpreting most interviews involves reviewing the data (in the form of transcripts, audio recordings or detailed notes), applying descriptive codes to the data and condensing and categorising codes to look for patterns. 24 27 These patterns can exist within a single interview or across multiple interviews depending on the research question and design. Qualitative computer software programs can be used to help organise and manage interview data.

Step 10: demonstrating the trustworthiness of the research

Similar to validity and reliability, qualitative research can be assessed on trustworthiness. 9 28 There are several criteria used to establish trustworthiness: credibility (whether the findings accurately and fairly represent the data), transferability (whether the findings can be applied to other settings and contexts), confirmability (whether the findings are biased by the researcher) and dependability (whether the findings are consistent and sustainable over time).

Step 11: presenting findings in a paper or report

When presenting the results of interview analysis, researchers will often report themes or narratives that describe the broad range of experiences evidenced in the data. This involves providing an in-depth description of participant perspectives and being sure to include multiple perspectives. 12 In interview research, the participant words are your data. Presenting findings in a report requires the integration of quotes into a more traditional written format.

• Conclusions

Though semistructured interviews are often an effective way to collect open-ended data, there are some disadvantages as well. One common problem with interviewing is that not all interviewees make great participants. 12 29 Some individuals are hard to engage in conversation or may be reluctant to share about sensitive or personal topics. Difficulty interviewing some participants can affect experienced and novice interviewers. Some common problems include not doing a good job of probing or asking for follow-up questions, failure to actively listen, not having a well-developed interview guide with open-ended questions and asking questions in an insensitive way. Outside of pitfalls during the actual interview, other problems with semistructured interviewing may be underestimating the resources required to recruit participants, interview, transcribe and analyse the data.

Despite their limitations, semistructured interviews can be a productive way to collect open-ended data from participants. In our research, we have interviewed children and adolescents about their stress experiences and coping behaviours, young women about their thoughts and behaviours during pregnancy, practitioners about the care they provide to patients and countless other key informants about health-related topics. Because the intent is to understand participant experiences, the possible research topics are endless.

Due to the close relationships family physicians have with their patients, the unique settings in which they work, and in their advocacy, semistructured interviews are an attractive approach for family medicine researchers, even if working in a setting with limited research resources. When seeking to balance both the relational focus of interviewing and the necessary rigour of research, we recommend: prioritising listening over talking; using clear language and avoiding jargon; and deeply engaging in the interview process by actively listening, expressing empathy, demonstrating openness to the participant’s worldview and thanking the participant for helping you to understand their experience.

• Further Reading

Edwards R, & Holland J. (2013). What is qualitative interviewing?: A&C Black.

Josselson R. Interviewing for qualitative inquiry: A relational approach. Guilford Press, 2013.

Kvale S. InterViews: An Introduction to Qualitative Research Interviewing. SAGE, London, 1996.

Pope C, & Mays N. (Eds). (2006). Qualitative research in health care.

• Supplementary files
• Publication history

-->

> > to return to Interviewing

	© RWJF 2008 P.O. Box 2316 College Road East and Route 1 Princeton, NJ 08543	-->Citation: Cohen D, Crabtree B. "Qualitative Research Guidelines Project." July 2006.

Reconciling Methodological Paradigms: Employing Large Language Models as Novice Qualitative Research Assistants in Talent Management Research

Qualitative data collection and analysis approaches, such as those employing interviews and focus groups, provide rich insights into customer attitudes, sentiment, and behavior. However, manually analyzing qualitative data requires extensive time and effort to identify relevant topics and thematic insights. This study proposes a novel approach to address this challenge by leveraging Retrieval Augmented Generation (RAG) based Large Language Models (LLMs) for analyzing interview transcripts. The novelty of this work lies in strategizing the research inquiry as one that is augmented by an LLM that serves as a novice research assistant. This research explores the mental model of LLMs to serve as novice qualitative research assistants for researchers in the talent management space. A RAG-based LLM approach is extended to enable topic modeling of semi-structured interview data, showcasing the versatility of these models beyond their traditional use in information retrieval and search. Our findings demonstrate that the LLM-augmented RAG approach can successfully extract topics of interest, with significant coverage compared to manually generated topics from the same dataset. This establishes the viability of employing LLMs as novice qualitative research assistants. Additionally, the study recommends that researchers leveraging such models lean heavily on quality criteria used in traditional qualitative research to ensure rigor and trustworthiness of their approach. Finally, the paper presents key recommendations for industry practitioners seeking to reconcile the use of LLMs with established qualitative research paradigms, providing a roadmap for the effective integration of these powerful, albeit novice, AI tools in the analysis of qualitative datasets within talent management research.

1. Introduction

Talent management researchers frequently work backwards from their customers, the employees at the organization. Understanding employee sentiment and behavior often involves conducting deep-dive interviews, explanatory in nature – e.g., demystifying the why behind customer choices, attitudes or behaviors (e.g., (Leino and Räihä, 2007 ) ). Talent management research, at its core, seeks to use science to equip every employee with resources to help them best navigate their careers (Zhao, 2023 ) .

Consequently, qualitative research methodology plays a critical role in talent management. Many of the key considerations around employee engagement, motivation, and workforce culture involve subjective, context-dependent factors that are best explored through in-depth interviews, focus groups, and other qualitative data collection approaches. Talent management professionals often rely on rich qualitative datasets to gain deep insights into employee experiences, organizational dynamics, and the nuances of human capital. However, these qualitative paradigms can clash with the more positivist, quantitative worldview that underlies many of the analytic tools used to evaluate talent management data. Talent management researchers may find that standard statistical techniques and data visualization approaches struggle to fully capture the complexities inherent in qualitative datasets, leading to potential misinterpretations or oversimplifications of the human elements involved in managing an organization’s workforce. Navigating this tension between qualitative and quantitative approaches is an ongoing challenge for talent management professionals.

Large language models (LLMs) like BERT, GPT-3 and PaLM have demonstrated strong aptitude for summarization (e.g., (Yang et al . , 2023 ) ), classification (e.g., (Pelaez et al . , 2024 ) ), and information extraction (e.g., (Dunn et al . , 2022 ) ) for text-based data. Consequently, LLMs are also increasingly being leveraged within talent management contexts for tasks such as interview analysis. However, language models are themselves designed primarily from a quantitative, data-driven paradigm. These models are trained on vast troves of text data using statistical machine learning techniques optimized for numerical patterns and correlations. While powerful at extracting insights from large-scale datasets, LLMs can often struggle to fully capture the nuanced, contextual nature of language (Bender et al . , 2021 ) , (Dwivedi et al . , 2023 ) that is critical for qualitative information sourced from interviews, focus groups, and other qualitative research methods common in talent management.

Talent management professionals must therefore continuously navigate a tension between the quantitative orientation of their analytical tools and the qualitative richness of the human dynamics they seek to understand. Bridging this gap requires innovative approaches that combine the opportunity for scale and speed offered by LLM-powered analysis augmented by borrowing evaluative nuances of traditional qualitative techniques. Talent leaders, thus, must carefully select and configure their AI-powered tools to ensure the voices and experiences of employees are authentically represented, rather than reduced to oversimplified metrics. Mastering this balance is an ongoing challenge, but one that is critical for talent management to yield truly holistic and impactful insights.

This paper presents results from leveraging LLMs as a novice qualitative researcher to augment qualitative research workstreams, specifically for data generated through semi-structured interviews.

The purpose of this paper is two-fold – 1) provide an overview of a successful implementation of a Retrieval Augmented Generation-based model for analyzing semi-structured interviews, and more importantly, 2) enumerate pragmatic take-aways and learnings drawing from traditional qualitative research to help fellow industry practitioners in reconciling the methodological paradigms. We posit the second purpose to be valuable to the larger discussion within talent management research communities on how and where to integrate AI capabilities across different talent management workstreams.

2. Quantitative and Qualitative Paradigms

Quantitative and qualitative research represent two fundamental paradigms or philosophical frameworks that guide research strategies, methods, analysis, and use of results (Yilmaz, 2013 ) . While both methodological approaches seek to rigorously study research problems, they are based on distinct assumptions and procedures adapted to investigating particular types of questions and drawing different conclusions. Quantitative research is based on the assumptions of positivism, the philosophical tradition premised on the application of natural science methods to the study of social reality and beyond (Bryman, 2016 ) . Quantitative researchers believe that objective facts and truths about human behavior and society can be measured and quantified numerically. Quantitative methods such as surveys, structured observations, and experiments aim to test hypotheses derived from theories by examining relationships between precisely measured variables statistically analyzed using large sample sizes (Creswell and Creswell, 2017 ) . These methods seek to minimize subjectivity and generalize findings to a population. In contrast, qualitative research aligns with interpretivist and constructivist philosophical traditions by embracing subjectivity and focused meaning-making by and with research participants (Denzin et al . , 2023 ) .

Qualitative researchers often use an inductive approach aimed at discovering and understanding processes, experiences, and worldviews by collecting non-numerical data through methods like in-depth interviews, ethnographic fieldwork, and document analysis. Findings derive from themes that emerge openly from the data rather than testing predetermined hypotheses. Samples tend to be small and purposely selected to illuminate a phenomenon in depth and detail. The aim is particularization rather than generalization, with a priority on ecological validity and multiple realities situated in time, place, culture, and context.

While debates once positioned these paradigms in opposition, contemporary mixed methods research leverages the complementary strengths of quantitative and qualitative approaches (Halcomb and Hickman, 2015 ) . Mixed methods investigations integrate quantitative and qualitative data collection and analysis within a single program of inquiry by combining these approaches in creative ways to deepen understanding (Creamer, 2017 ) (Creamer, 2018 ) (Greene, 2008 ) . This reconciliation of methodological perspectives offers opportunities to generate more robust, contextualized insights to address complex research problems. The use of large language models (LLMs) as novice qualitative research assistants, as explored in this paper, can be considered an exercise in mixed methods research design.

Prior to LLMs, in previous work, Natural Language Processing based modeling of qualitative data from social science contexts, have also been used as "novice insight" augmented by the more expert contextualization provided by human researchers (e.g., (Bhaduri, 2018 ) , (Bhaduri et al . , 2021 ) ). Popular traditional topic modeling techniques (e.g. Latent Dirichlet Allocation), however, suffer from several limitations (e.g. specifying number of clusters) when compared to existing deep learning-based methods. They also often fail to capture the contextual nuances and ambiguities inherent in natural language, as they rely heavily on predefined rules and patterns (Devlin, 2018 ) (Radford et al . , 2019 ) . This can make it challenging to handle the complexities and variations present in real-world text data, and may require domain-specific knowledge or fine-tuning to achieve acceptable performance (Lee and Hsiang, 2019 ) . Recent advancements in LLMs, such as BERT and GPT, have largely overcome these limitations by leveraging deep neural networks to learn rich, contextual representations from large amounts of text data (Vaswani et al . , 2017 ) (Devlin, 2018 ) . These powerful models can capture subtle semantic and pragmatic features of language, and demonstrate strong generalization capabilities through transfer learning (Brown, 2020 ) (Radford et al . , 2019 ) .

Further, in traditional qualitative research, thematic analysis is the process of gathering themes across topics from qualitative data, such as interview data, through iteratively analyzing the dataset for topics of interest (Creamer, 2017 ) . Inductive coding and deductive coding are two approaches to analyzing data from semi-structured interviews. Inductive coding involves starting with raw data and gradually developing codes and categories based on patterns and topics that emerge from the data as the researcher manually interacts with it (Patton, 2014 ) (Strauss and Corbin, 1998 ) . This approach is bottom-up, where the data drives the development of codes and theories (Glaser, 1965 ) . Deductive coding, on the other hand, involves starting with preconceived codes or theories and applying them to the data (Pearse, 2019 ) . This approach is top-down, where existing theories or frameworks guide the coding process (Maxwell, 2018 ) . Researchers in industry typically work backwards from research question of interest. Most of the research questions in industry driving qualitative data collection are also explanatory (i.e., tend to explain the quantitative findings such as low customer satisfaction, low product adoption numbers), rather than exploratory (i.e., ethnography of a community of interest or a phenomenon) and as a result deductive approaches are often more popular than inductive coding.

Ultimately, by augmenting traditional deep-dive qualitative analysis with the time and resource efficient pattern recognition and text processing capabilities of LLMs, researchers can integrate quantitative and qualitative techniques to enhance the speed, depth, and rigor of their investigations. This mental model of a novice-LLM approach holds promise for bridging the divide between positivist and interpretive paradigms, ultimately working towards a more comprehensive understanding of the phenomenon under study.

We used an open-source dataset (Paskevicius, 2018 ) to demonstrate how an LLM prompted as a novice researcher can enhance traditional qualitative deductive thematic coding. This dataset was originally collected to explore educators’ experiences implementing open educational practices (Paskevicius, 2018 ) . The dataset contains eight transcripts each from hour-long interviews conducted with educators to understand how they are using openly accessible sources of knowledge and open-source tools. The original research involved a deep-dive qualitative analysis through using a phenomenological approach to extract topics manually from the dataset. We chose this open-source dataset for two reasons – 1) structural match to proprietary dataset, and 2) rich description and manually identified topics by an expert to serve as a gold standard to measure the efficacy of our LLM based approach. Semi-structured interviews provide critical insights through participant perspectives, making them foundational in various industry settings.

The semi-structured approach used to create this dataset is a close match to proprietary talent management data from our organization, where employees are interviewed on a particular phenomenon to get deeper understanding of their related sentiment, attitudes, and behaviors. Manually extracted topics serve as gold standard for benchmarking findings from our LLM-based approach. The paper (Paskevicius, 2018 ) describing the dataset explains the manual process establishing how each transcript was read twice: first, for a comprehensive analysis, and subsequently, to initiate a thematic exploration. Additional reviewing continued as codes and topics emerged and intersected among the interviews. A manual qualitative coding approach was applied at each iteration to reveal themes, following constant comparison methodology (Glaser, 1965 ) .

We posit that our approach, as demonstrated on this sample semi-structured interview dataset, can easily extend to multiple industry settings in talent management research where researchers conduct interviews and focus groups.

4. Thematic Analysis Using LLMs

In traditional, manual qualitative research, deductive thematic analysis process begins with the researcher first formulating the research questions. Then, upon collection of the data, such as interview transcripts, the researcher iterates manually through the transcripts to identify and extract themes or topics of interest. This labor-intensive process involves carefully reading through the data, taking notes, and organizing the topics iteratively into broader coherent themes that address the research questions. The researcher may go through multiple rounds of coding and analysis to refine the themes and ensure they comprehensively capture the key insights from the data. Our approach finds that LLMs can quickly uncover topics of interest from the dataset which can then be iterated upon to garner broader themes of interest across topics. Thus, for our novice-LLM led approach, we leveraged the power of Large Language Models (LLMs) as a novice research assistant in the thematic analysis process. Specifically, we used the open-source framework called Langchain to create dynamic prompt templates, such as few-shot prompts and chain of thoughts, that guided the LLM in performing topic modeling and generating insights from the interview transcripts. We then opted to use Anthropic’s Claude2 model to execute these prompts and extract the relevant themes.

To initiate the analysis, we first selected a main research question and corresponding sub-questions from our dataset (Paskevicius, 2018 ) . We then fed these research questions, along with the interview transcripts, into the LLM-powered Langchain framework. The model was able to quickly identify and summarize the key topics, and iteratively, themes emerging from the data. This approach provided a quick yet relatively comprehensive analysis that would have taken a human researcher significant time and effort to reproduce manually.

4.1. Thematic analysis enhanced through Retrieval Augmented Generation (RAG)

In our LLM based approaches, we experiment with four methods - zero-shot prompting, few-shot prompting, chain-of-thought reasoning, and Retrieval Augmented Generation based Question Answering. In zero-shot prompting we provide a single prompt to the model. In few-shot prompting, we provide a set of topics and anecdotes to the model as examples. In the chain of thought (COT) approach, we provide a set of instructions for the model to follow. Finally, for Retrieval Augmented Generation (RAG) we provide context and questions to the model, from which it extracts information.

Zero-shot prompts are simple instructions or tasks given to an LLM that have not been specifically trained on that task. It serves as a baseline because it demonstrates the model’s fundamental ability to understand and respond to prompts based solely on its pre-training (Kong et al . , 2023 ) . In few-shot prompting, a small set of examples illustrating the desired outcome are manually selected and provided to the LLM. These examples allow the model to understand the tasks at hand and generate similar results (Brown, 2020 ) . Chain-of-thought prompting provides a set of intermediate steps to guide the LLM to mimic human-like reasoning. This significantly improves the capability of the LLM to understand complex reasoning and generate better topics (Wei et al . , [n. d.] ) . Retrieval-augmented generation (RAG) combines the capabilities of an LLM with a retrieval system to source and integrate additional information into its responses (Lewis et al . , 2020 ) . This effort provides contextually richer and ultimately more accurate outputs. We do this by providing all the interview transcripts to the LLM as a custom knowledge base. Two considerations helped the RAG approach outperform the other approaches:

4.1.1. Focused Analysis:

In our approach, LLM searches the knowledge base to find and retrieve parts of documents that are most relevant to the question in the query. This narrows the focus to the most relevant information and ensures attention to critical topics and nuances.

4.1.2. Context Dilution/Managing Information Overload:

Using all transcripts as input in a single instance creates information overload scenarios, ultimately leading to dilution of important topics or nuances. If the dataset is too large or complex, LLM might lose track of what’s most relevant to specific query, leading hallucinations. Hallucinations or inaccuracies within this context refers to instances where the model generates information which is not grounded in input data. In our approach, the use of RAG mitigates some of the hallucination by anchoring LLM responses relevant information, and providing a form of contextual validation for the output.

Distillbert-base-uncased	Precision	Recall	F1-Score
Chain of Thought	67%	62%	64%
Few Shot	72%	67%	70%
Zero Shot	68%	66%	67%
RAG	79%	80%	79%
Bert-base-uncased	Precision	Recall	F1-Score
Chain of Thought	56%	48%	52%
Few Shot	64%	56%	60%
Zero Shot	59%	55%	57%
RAG	70%	70%	70%
Roberta-large	Precision	Recall	F1-Score
Chain of Thought	89%	85%	87%
Few Shot	90%	87%	88%
Zero Shot	89%	86%	88%
RAG	92%	91%	91%

5. Findings

In the paper describing the dataset leveraged for this work, the authors collected and conducted a manual analysis (Paskevicius, 2018 ) . Their research led to identification of significant, recurring topics within the interviews. Our evaluation strategy uses these manually generated topics from the paper’s work as gold standard to compare against topics generated by the LLMs-based approach. We use Precision (Equation 1), Recall (Equation 2), and F1-score (Equation 3) to benchmark topics generated by our LLM-augmented qualitative research approach against the topics generated by the human researcher.

(1)

(2)

(3)

(4)

These metrics are the current evaluation standard for classification models, but they can be adapted for text generation tasks (Zhang et al . , 2019 ) . Precision and Recall measure the proportion of correctly identified positive cases. In the context of our experiment, every word from predicted text gets matched to a word in the referenced text to compute recall. This process is inverted to then compute precision. The precision and recall values are then combined to compute an F1 score. These metrics use cosine similarity (Equation 4) in which each predicted word is paired with its closest corresponding word from the reference text with the aim of maximizing the similarity score.

In Table 1, the performance of various LLM prompting techniques including Chain of Thought, Few Shot, Zero Shot and RAG, are compared across different embedding models (Distillbert-base-uncased, Bert-base-uncased, and Roberta-large). This comparison aims to evaluate the robustness and effectiveness of these prompting techniques. Our results indicate that while each prompting technique shows varying level of precision, recall and F1-score, RAG consistently outperform the others on all three metrics, achieving highest performance across all models.

Example: Keywords from LDA Topic One

Students

Course

Develop

People

Institution

Project

Science

Discipline

Material

Start

Example: Output from LLM approach

Collaboration: Co-creating resources and connecting with others

Corresponding Anecdote: You can also in your teaching have students connect with people outside the

course in various ways. Like, maybe some people outside the course are commenting on blogs

and student are getting in a conversation around that.

6. Learnings

Treating large language models (LLMs) as novice research assistants during thematic analysis offered valuable insights for our research. By framing the LLM as a novice collaborator with little knowledge or insight of the context, prompts can be crafted to better guide the model and leverage its capabilities. Used prudently, similar novice LLM-augmented approaches can significantly increase time and resource efficiency compared to traditional qualitative coding methods in talent management research. The following sections explore some of our key learnings that may benefit other researchers considering designing LLMs as novice researchers to optimize thematic analysis.

6.1. Approaching LLMs as Novice Research Assistants can help prepare better prompts

A novice is a person who, “has no experience with the situations in which they are expected to perform tasks” (Benner, 1982 ) . The novice is thus at a basic proficiency level for skill acquisition, with limited information and prior experience related to a task at hand (Montfort et al . , 2013 ) . For large qualitative datasets analyzed using LLMs we propose that a novice-led approach to analysis is a good fit. In our approach the human behaves as an expert prompting the novice LLM to provide insights related to topics of interest. We found this framework as a helpful mental model to ground the primary researcher prompting the LLM as they iteratively uncover insights from the dataset.

6.2. Used prudently, LLMs can help increase time effectiveness and resource efficiency

LLMs have advanced the field of natural language processing with their ability to understand and generate responses that closely mimic human language (Shanahan, 2024 ) . The strengths of LLMs extend beyond metrics, these models are adept at processing vast amounts of text rapidly, demonstrating a level of topic modeling that can mimic human analysis. Manual topic modeling is human labor intensive and time inefficient (Clarke and Braun, 2017 ) . LLMs also enhance efficiency by streamlining the processing of large datasets, allowing for the extraction of topics from qualitative data more quickly. Improvisations of these model using techniques like few-shot and zero-shot learning capabilities further reduce the need for expensive data labeling and annotations. In a nutshell, LLMs boost speed, reduce human effort, scale to massive datasets, and lower labeling costs. However, human expertise is still essential for judgment, validation and end-to-end framework design.

6.3. LLM augmented approaches offer significant increase in ease and enhanced context compared to traditional NLP approaches.

Using a RAG approach towards an LLM-augmented qualitative research analyzing semi-structure interviews shows great promise compared to natural language processing methods like Latent Dirichlet allocation (LDA). Currently, there are no widely accepted methods for comparing the two approaches as there is no bridge to compare keywords to themes, except from a human-evaluator ease of interpretability standpoint. We performed topic modeling analysis on the same dataset with the broader aim of finding themes. Manually comparing both approaches, each researcher of this workstream independently found that any of the approaches using an LLM yielded much greater context and consequently, better interpretability than the traditional LDA approach. This is likely because, with LDA, the model outputs a list of words and probability for each topic. With these words, the researcher would then have to manually define the topic. While this approach increases researcher flexibility, it remains time and resource consuming. In contrast, with the LLM approach, the output is richer in context of what particular topics mean. For example, our LDA model yielded 5 topics (see: Appendix A Figure 3). The first 10 words for topic 1 can also be seen in Table 2. Putting these words together into a comprehensive theme can be challenging without more context. However, an LLM is able to generate context grounded in the participant’s voice for researchers to work with. An example of an extracted theme and its corresponding anecdote using an LLM can also be seen in Table 2, above.

7. Recommendations

Traditional qualitative research is evaluated based on several criteria that ensure quality and rigor of the research, both in terms of methods as well as findings. Prior research has established four criteria for increased rigor and trustworthiness of qualitative research studies around credibility, dependability, confirmability, and transferability (Lincoln and Guba, 1988 ) . We recommend three ways in which quality criteria from traditional qualitative research can be used by practitioners employing LLM augmented analysis of qualitative data.

7.1. Establishing credibility of findings by incorporating mechanism for member checks.

Member checks, i.e., the strategy of soliciting insights from research participants on research findings, are often relied on as the gold standard for increasing trustworthiness of qualitative research approaches (e.g., (Patton, 2014 ) (Kornbluh, 2015 ) ). Qualitative researchers employing LLMs can work on deepening their understanding of the research context using appropriate data-collection methods and tools that work best for particular contexts, as well as conduct adequate member checking to ensure the accuracy of findings.

7.2. Practicing increased researcher reflexivity.

Qualitative researchers are recommended that they acknowledge and address their own biases, thus recognizing the influence of their own experiences and opinions on the research process (Finlay, 2002 ) . Similar exercises on reflectivity can also be helpful for researchers augmenting qualitative data analysis through employing LLMs. Researcher reflexivity in such instances can extend to querying the LLM to ask for rationale on why certain topics were extracted, grounding topics in anecdotes from the transcripts, and recognizing the influence the human researcher’s prior knowledge and biases will have on the prompts used. Future work in extending LLMs for qualitative research should continue to draw on evaluation criteria grounded in traditional qualitative research paradigm.

7.3. Increasing transparency of decisions made throughout the research study.

Qualitative researchers are recommended to thoroughly document all decisions that guide their analysis process by providing thick descriptions, allowing for increased transparency. This practice enhances reliability and reproducibility of the research (Lincoln and Guba, 1988 ) . Qualitative researchers employing LLMs should also similarly strategize maximizing transparency through mechanisms such as documenting changes in workflow, sharing prompts, and detailing model preferences.

8. Closing Thoughts

The approach outlined in this paper offers a promising avenue for industry-based talent management practitioners seeking to increase the time and resource efficiency of qualitative interview data analysis. By leveraging large language models (LLMs) as novice qualitative research assistants, organizations can potentially accelerate the coding, categorization, and thematic synthesis of rich interview data - a critical bottleneck in many talent management research initiatives.

However, as the field of LLM-assisted qualitative research matures, it will be essential to not only benchmark model performance against traditional quantitative evaluation metrics, but also consider quality criteria more prominent within the qualitative research paradigm. Factors such as credibility, transferability, dependability, and confirmability will need to be carefully evaluated as LLMs are integrated into qualitative workflows. Furthermore, the ethical use of AI assistants in sensitive domains like talent management will require close, multi-disciplinary attention to issues at the intersection of data privacy, algorithmic bias, and model transparency, for which researchers will have to be trained (Mackenzie et al . , 2024 ) .

Future research should seek to establish guidelines and best practices for LLM-augmented qualitative analysis that uphold the rigor and trustworthiness expected within the qualitative research community. Only by doing so can talent management scholars and practitioners unlock the full potential of these powerful language models, while respecting the epistemological foundations of qualitative inquiry. As the field evolves, we believe that a judicious, ethically-grounded approach to LLM integration can yield substantial gains in research efficiency and organizational impact.

• Bender et al . (2021) Emily M Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the dangers of stochastic parrots: Can language models be too big?. In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency . 610–623.
• Benner (1982) Patricia Benner. 1982. From novice to expert. AJN The American Journal of Nursing 82, 3 (1982), 402–407.
• Bhaduri (2018) Sreyoshi Bhaduri. 2018. NLP in Engineering Education-Demonstrating the use of Natural Language Processing Techniques for Use in Engineering Education Classrooms and Research. (2018).
• Bhaduri et al . (2021) Sreyoshi Bhaduri, Michelle Soledad, Tamoghna Roy, Homero Murzi, and Tamara Knott. 2021. A Semester Like No Other: Use of Natural Language Processing for Novice-Led Analysis on End-of-Semester Responses on Students’ Experience of Changing Learning Environments Due to COVID-19. In 2021 ASEE Virtual Annual Conference Content Access .
• Brown (2020) Tom B Brown. 2020. Language models are few-shot learners. arXiv preprint ArXiv:2005.14165 (2020).
• Bryman (2016) Alan Bryman. 2016. Social research methods . Oxford university press.
• Clarke and Braun (2017) Victoria Clarke and Virginia Braun. 2017. Thematic analysis. The journal of positive psychology 12, 3 (2017), 297–298.
• Creamer (2017) Elizabeth G Creamer. 2017. An introduction to fully integrated mixed methods research . sage publications.
• Creamer (2018) Elizabeth G Creamer. 2018. Striving for methodological integrity in mixed methods research: The difference between mixed methods and mixed-up methods. , 526–530 pages.
• Creswell and Creswell (2017) John W Creswell and J David Creswell. 2017. Research design: Qualitative, quantitative, and mixed methods approaches . Sage publications.
• Denzin et al . (2023) Norman K Denzin, Yvonna S Lincoln, Michael D Giardina, and Gaile S Cannella. 2023. The Sage handbook of qualitative research . Sage publications.
• Devlin (2018) Jacob Devlin. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018).
• Dunn et al . (2022) Alexander Dunn, John Dagdelen, Nicholas Walker, Sanghoon Lee, Andrew S Rosen, Gerbrand Ceder, Kristin Persson, and Anubhav Jain. 2022. Structured information extraction from complex scientific text with fine-tuned large language models. arXiv preprint arXiv:2212.05238 (2022).
• Dwivedi et al . (2023) Yogesh K Dwivedi, Nir Kshetri, Laurie Hughes, Emma Louise Slade, Anand Jeyaraj, Arpan Kumar Kar, Abdullah M Baabdullah, Alex Koohang, Vishnupriya Raghavan, Manju Ahuja, et al . 2023. Opinion Paper:“So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. International Journal of Information Management 71 (2023), 102642.
• Finlay (2002) Linda Finlay. 2002. Negotiating the swamp: the opportunity and challenge of reflexivity in research practice. Qualitative research 2, 2 (2002), 209–230.
• Glaser (1965) Barney G Glaser. 1965. The constant comparative method of qualitative analysis. Social problems 12, 4 (1965), 436–445.
• Greene (2008) Jennifer C Greene. 2008. Is mixed methods social inquiry a distinctive methodology? Journal of mixed methods research 2, 1 (2008), 7–22.
• Halcomb and Hickman (2015) Elizabeth J Halcomb and Louise Hickman. 2015. Mixed methods research. (2015).
• Kong et al . (2023) Aobo Kong, Shiwan Zhao, Hao Chen, Qicheng Li, Yong Qin, Ruiqi Sun, and Xin Zhou. 2023. Better zero-shot reasoning with role-play prompting. arXiv preprint arXiv:2308.07702 (2023).
• Kornbluh (2015) Mariah Kornbluh. 2015. Combatting challenges to establishing trustworthiness in qualitative research. Qualitative research in psychology 12, 4 (2015), 397–414.
• Lee and Hsiang (2019) Jieh-Sheng Lee and Jieh Hsiang. 2019. Patentbert: Patent classification with fine-tuning a pre-trained bert model. arXiv preprint arXiv:1906.02124 (2019).
• Leino and Räihä (2007) Juha Leino and Kari-Jouko Räihä. 2007. Case amazon: ratings and reviews as part of recommendations. In Proceedings of the 2007 ACM conference on Recommender systems . 137–140.
• Lewis et al . (2020) Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al . 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems 33 (2020), 9459–9474.
• Lincoln and Guba (1988) Yvonna S Lincoln and Egon G Guba. 1988. Criteria for Assessing Naturalistic Inquiries as Reports. (1988).
• Mackenzie et al . (2024) Tammy Mackenzie, Leslie Salgado, Sreyoshi Bhaduri, Victoria Kuketz, Solenne Savoia, and Lilianny Virguez. 2024. Beyond the Algorithm: Empowering AI Practitioners through Liberal Education. In 2024 ASEE Annual Conference & Exposition .
• Maxwell (2018) Joseph A Maxwell. 2018. Collecting qualitative data: A realist approach. The SAGE handbook of qualitative data collection (2018), 19–32.
• Montfort et al . (2013) Devlin B Montfort, Geoffrey L Herman, Shane A Brown, Holly M Matusovich, and Ruth A Streveler. 2013. Novice-led paired thematic analysis: A method for conceptual change in engineering. In 2013 ASEE Annual Conference & Exposition . 23–933.
• Paskevicius (2018) Michael Paskevicius. 2018. Exploring educators experiences implementing open educational practices . Ph. D. Dissertation.
• Patton (2014) Michael Quinn Patton. 2014. Qualitative research & evaluation methods: Integrating theory and practice . Sage publications.
• Pearse (2019) Noel Pearse. 2019. An illustration of deductive analysis in qualitative research. In 18th European conference on research methodology for business and management studies . 264.
• Pelaez et al . (2024) Sergio Pelaez, Gaurav Verma, Barbara Ribeiro, and Philip Shapira. 2024. Large-scale text analysis using generative language models: A case study in discovering public value expressions in AI patents. Quantitative Science Studies 5, 1 (2024), 153–169.
• Radford et al . (2019) Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al . 2019. Language models are unsupervised multitask learners. OpenAI blog 1, 8 (2019), 9.
• Shanahan (2024) Murray Shanahan. 2024. Talking about large language models. Commun. ACM 67, 2 (2024), 68–79.
• Strauss and Corbin (1998) Anselm Strauss and Juliet Corbin. 1998. Basics of qualitative research techniques. (1998).
• Vaswani et al . (2017) Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. 2017. Attention is All you Need. In Advances in Neural Information Processing Systems , I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (Eds.), Vol. 30. Curran Associates, Inc. https://proceedings.neurips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf
• Wei et al . ([n. d.]) J Wei, X Wang, D Schuurmans, M Bosma, F Xia, and E Chi. [n. d.]. & Zhou, D.(2022). Chain-of-thought prompting elicits reasoning in large language models ([n. d.]), 24824–24837.
• Yang et al . (2023) Binxia Yang, Xudong Luo, Kaili Sun, and Michael Y Luo. 2023. Recent progress on text summarisation based on bert and gpt. In International Conference on Knowledge Science, Engineering and Management . Springer, 225–241.
• Yilmaz (2013) Kaya Yilmaz. 2013. Comparison of quantitative and qualitative research traditions: Epistemological, theoretical, and methodological differences. European journal of education 48, 2 (2013), 311–325.
• Zhang et al . (2019) Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2019. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675 (2019).
• Zhao (2023) Wanqun Zhao. 2023. Using Science to Support and Develop Employees in the Tech Workforce—An Opportunity for Multidisciplinary Pursuits in Engineering Education. In 2023 ASEE Annual Conference & Exposition .

Appendix A Results from analyzing the same dataset using an LDA Approach.

Traditional topic modeling using approaches such as Latent Dirichlet Allocation (LDA) often present the most representative words for each generated topic. For instance, for Topic 1 words such as "students", "develop", "institution", "science", etc. were found important. Attempting to interpret the underlying thematic meaning of these word lists can be challenging without additional contextual information about how those words were used within the original corpus. In contrast, large language models (LLMs) have demonstrated the capability to synthesize the semantically related words and phrases into more coherent topical representations. This ability of LLMs to generate primitive yet formative contextual information threading together words and phrases of interest and thereby provide researchers with a more insightful starting point for further analysis and interpretation of the latent topics uncovered through the LDA process.

• Open access
• Published: 19 August 2024

The impact of study habits and personal factors on the academic achievement performances of medical students

• Mohammed A. Aljaffer 1 ,
• Ahmad H. Almadani 1 ,
• Abdullah S. AlDughaither 2 ,
• Ali A. Basfar 2 ,
• Saad M. AlGhadir 2 ,
• Yahya A. AlGhamdi 2 ,
• Bassam N. AlHubaysh 2 ,
• Osamah A. AlMayouf 2 ,
• Saleh A. AlGhamdi 3 ,
• Tauseef Ahmad 4 &
• Hamza M. Abdulghani 5  

BMC Medical Education volume  24 , Article number:  888 ( 2024 ) Cite this article

61 Accesses

2 Altmetric

Metrics details

Academic achievement is essential for all students seeking a successful career. Studying habits and routines is crucial in achieving such an ultimate goal.

This study investigates the association between study habits, personal factors, and academic achievement, aiming to identify factors that distinguish academically successful medical students.

A cross-sectional study was conducted at the College of Medicine, King Saud University, Riyadh, Saudi Arabia. The participants consisted of 1st through 5th-year medical students, with a sample size of 336. The research team collected study data using an electronic questionnaire containing three sections: socio-demographic data, personal characteristics, and study habits.

The study results indicated a statistically significant association between self-fulfillment as a motivation toward studying and academic achievement ( p  = 0.04). The results also showed a statistically significant correlation between recalling recently memorized information and academic achievement ( p  = 0.05). Furthermore, a statistically significant association between preferring the information to be presented in a graphical form rather than a written one and academic achievement was also found ( p  = 0.03). Students who were satisfied with their academic performance had 1.6 times greater chances of having a high-grade point average (OR = 1.6, p  = 0.08).

The results of this study support the available literature, indicating a correlation between study habits and high academic performance. Further multicenter studies are warranted to differentiate between high-achieving students and their peers using qualitative, semi-structured interviews. Educating the students about healthy study habits and enhancing their learning skills would also be of value.

Peer Review reports

Introduction

Academic performance is a common indicator used to measure student achievement [ 1 , 2 ]. It is a compound process influenced by many factors, among which is study habits [ 2 , 3 ]. Study habit is defined as different individual behavior in relation to studying, and is a combination of study methods and skills [ 2 , 3 , 4 ]. Put differently, study habits involve various techniques that would increase motivation and transform the study process into an effective one, thus enhancing learning [ 5 ]. Students’ perspectives and approaches toward studying were found to be the key factors in predicting their academic success [ 6 , 7 ]. However, these learning processes vary from one student to another due to variations in the students’ cognitive processing [ 8 ].

The study habits of students are the regular practices and habits they exhibit during the learning process [ 9 , 10 ]. Over time, several study habits have been developed, such as time management, setting appropriate goals, choosing a comfortable study environment, taking notes effectively, choosing main ideas, and being organized [ 11 ]. Global research shows that study habits impact academic performance and are the most important predictor of it [ 12 ]. It is difficult for medical students to organize and learn a lot of information, and they need to employ study skills to succeed [ 1 , 2 , 5 , 13 ].

Different lifestyle and social factors could affect students’ academic performance. For instance, Jafari et al. found that native students had better study habits compared to dormitory students [ 1 ]. This discrepancy between native and dormitory students was also indicated by Jouhari et al. who illustrated that dormitory students scored lower in attitude, test strategies, choosing main ideas, and concentration [ 10 ]. Regarding sleeping habits, Curcio G et al. found that students with a regular and adequate sleeping pattern had higher Grade Point Average (GPA) scores [ 14 ]. Lifestyle factors, such as watching television and listening to music, were shown to be unremarkable in affecting students’ grades [ 15 , 16 ]. Social media applications, including WhatsApp, Facebook, and Twitter, distract students during learning [ 16 , 17 ].

Motivation was found to be a major factor in students’ academic success. Bonsaksen et al. found that students who chose “to seek meaning” when studying were associated with high GPA scores [ 18 ]. In addition, low scores on “fear of failure” and high scores on “achieving” correlated with a higher GPA [ 8 , 18 ].

Resource-wise, Alzahrani et al. found that 82.7% of students relied on textbooks assigned by the department, while 46.6% mainly relied on the department’s lecture slides [ 19 ]. The study also indicated that 78.8% perceived that the scientific contents of the lectures were adequate [ 19 ]. Another study found that most students relied on the lecture slides (> 83%) along with their notes, followed by educational videos (76.1%), and reference textbooks (46.1%) [ 20 ]. Striking evidence in that study, as well as in another study, indicated that most students tended to avoid textbooks and opted for lecture slides, especially when preparing for exams [ 20 , 21 ].

Several researchers studied the association between different factors and academic performance; however, more is needed to know about this association in the process of education among medical students [ 15 , 20 , 22 ], with some limitations to the conducted studies. Such limitations include the study sample and using self-reported questionnaires, which may generate inaccurate results. Moreover, in Saudi Arabia in particular, the literature concerning the topic remains limited. Since many students are unsatisfied with their performance and seek improvement [ 10 ], the present study was designed and conducted.

Unlike other studies in the region, this study aims to investigate the relationship between study habits and personal factors and measure their influence on academic achievement. The results of this study could raise awareness regarding the effect of study habits and personal factors on students’ performance and would also guide them toward achieving academic success. The study also seeks to identify the factors that distinguish academically successful students from their peers.

Study design, setting, and participants

This observational cross-sectional study, which took place between June and December 2022, was conducted among students attending the College of Medicine at King Saud University (KSU), Riyadh, Saudi Arabia. Its targeted population included all male and female medical students (first to fifth years) attending KSU during the academic year 2021/2022. Whereas, students at other colleges and universities, those who failed to complete the questionnaire, interns (the students who already graduated), and those who were enrolled in the university’s preparatory year, were all excluded from the current study. The sample size was calculated based on a study conducted in 2015 by Lana Al Shawwa [ 15 ]. Using the sample size formula for a single proportion (0.79), the required sample size was 255 using a confidence interval of 95% and a margin of error of 5%. After adding a 20% margin to accommodate non-responses and incomplete responses, the calculated sample size required for this study was 306. However, our research team collected a total of 336 participants for this study to ensure complete representation.

Study instrument

The research team developed and used an electronic questionnaire. The rationale is that no standardized questionnaire measuring the study objectives was found in the literature. However, the questionnaire was tested on a pilot of 15 students to test its clarity and address any possible misconceptions and ambiguity. The study questionnaire was distributed randomly to this cohort, who were asked to fill out the questionnaire. The students reported a complete understanding of the questionnaire’s contents, so the same questionnaire was used without any modifications. The questionnaire, written in English, consisted of three parts. The first part included eleven questions about the socio-demographic status of the participants. The second part contained twenty-one questions examining personal factors such as sleep and caffeine consumption. The last part included twenty-one questions regarding students’ study habits. The questionnaire was constructed based on an ordinal Likert scale which had: strongly agree, agree, neutral, disagree, and strongly disagree as possible answers. The questionnaire was sent to participants through email and social media applications like Twitter and WhatsApp to increase the study response. An informed consent that clearly states the study’s purpose was taken from all participants at the beginning of the questionnaire. In addition, all participants were assured that the collected data would be anonymous and confidential. Each participant was represented by a code for the sole purpose of analyzing the data. Furthermore, no incentives or rewards were given to the participants for their participation.

Study variables

Socio-demographic information (such as age, gender, and academic year), and personal factors (such as motivation, sleeping status, caffeine consumption, and self-management) were the independent variables. Study habits such as attendance, individual versus group study, memorization techniques, revision, learning style, and strategies were also independent variables.

Academic achievement refers to a student’s success in gaining knowledge and understanding in various subjects, as well as the ability to apply that knowledge effectively [ 23 ]. It is a measure of the student’s progress throughout the educational journey, encompassing both academic achievements and personal growth [ 3 , 24 ]. Academic achievement is judged based on the student’s GPA or performance score. In this study, students’ GPA scores, awareness, and satisfaction regarding their academic performance were the dependent variables.

We divided the study sample into two groups based on the GPA. We considered students with high GPAs to be exposed (i.e. exposed to the study habits we are investigating), and students with low GPAs to be the control group. The purpose of this study was to determine why an exposed group of students gets high grades and what study factors they adopt. Based on this exposure (high achieving students), we concluded what methods they used to achieve higher grades. Those in the first group had a GPA greater or equal to 4.5 (out of 5), while those in the second group had a GPA less than 4.5. The students’ data were kept confidential and never used for any other purpose.

Data analysis

The data collected were analyzed by using IBM SPSS Statistical software for Windows version 24.0. Descriptive statistics such as frequency and percentage were used to describe the socio-demographic data in a tabular form. Furthermore, data for categorical variables, including different study habits, motivation factors, memorizing and revising factors, and lifestyle factors, were tabulated and analyzed using the odds ratio test. Finally, we calculated the odds ratio statistic and a p-value of 0.05 to report the statistical significance of our results.

Ethical approval and consent to Participate

Before conducting the study, the research team obtained the Ethics Committee Approval from the Institutional Review Board of the College of Medicine, KSU, Riyadh, Saudi Arabia (project No. E-22-7044). Participants’ agreement/consent to participate was guaranteed by choosing “agree” after reading the consent form at the beginning of the questionnaire. Participation was voluntary, and consent was obtained from all participants. The research team carried out all methods following relevant guidelines and regulations.

The total 336 medical students participated in the study. All participants completed the study questionnaire, and there were no missing or incomplete data, with all of them being able to participate. As shown in Table  1 9.3% of participants were between 18 and 20, 44.9% were between the ages of 21 and 22, and 35.8% were 23–28 years old. In the current study, 62.5% of the participants were males and 37.5% were females. The proportion of first-year students was 21.4%, 20.8% of second-year students, 20.8% of third-year students, 18.2% of fourth-year students, and 18.8% of fifth-year students, according to academic year levels. Regarding GPA scores, 36.9% scored 4.75-5 and 32.4% scored 4.5–4.74. 23.8% achieved 4-4.49, 6.5% achieved 3-3.99, and only 0.4% achieved 2.99 or less. Participants lived with their families in 94.6% of cases, with friends in 1.2% of cases, and alone in 4.2% of cases. For smoking habits, 86.3% did not smoke, 11% reported using vapes, 2.1% used cigarettes, and 0.6% used Shisha. 91.4% of the participants did not report any chronic illnesses; however, 8.6% did. In addition, 83% had no mental illness, 8.9% had anxiety, 6% had depression, and 2.1% reported other mental illnesses.

Table  2 shows motivational factors associated with academic performance. There was a clear difference in motivation factors between students with high and low achievement in the current study. Students with high GPAs were 1.67 times more motivated toward their careers (OR = 1.67, p  = 0.09) than those with low GPAs. Furthermore, significant differences were found between those students who had self-fulfillment or ambitions in life they had ~ 2 times higher (OR = 1.93, p  = 0.04) GPA scores than low GPA students. Exam results did not motivate exposed or high GPA students (46%) or control students with low GPA students (41%), but the current study showed test results had little impact on low achiever students (OR = 1.03, p  = 0.88). Furthermore, 72.6% of high achievers were satisfied with their academic performance, while only 41% of low achiever students were satisfied. Therefore, students who were satisfied with their academic performance had 1.6 times greater chances of a higher GPA (OR = 1.6, p  = 0.08). Students who get support and help from those around them are more likely to get high GPAs (OR = 1.1, p  = 0.73) than those who do not receive any support. When students reported feeling a sense of family responsibility, the odds (odds ratio) of their receiving higher grades were 1.15 times higher (OR = 1.15, p  = 0.6) compared to those who did not feel a sense of family responsibility. The p-value, which indicates the level of statistical significance, was 0.6.

Table  3 shows the study habits of higher achiever students and low achiever students. Most of the high-achieving students (79.0%) attended most of the lectures and had 1.6 times higher chances of getting higher grades (OR = 1.6, p  = 0.2) than those who did not attend regular lectures. The current study found that studying alone had no significant impact on academic achievement in either group. However, those students who had studied alone had lower GPAs (OR = 1.07, p  = 0.81). The current study findings reported 29.8% of students walk or stand while studying rather than sit, and they had 1.57 times higher GPA chances compared to students with lower GPAs (OR = 0.73, p  = 0.27). High achievers (54.0%) preferred studying early in the morning, and these students had higher chances of achieving good GPAs (OR = 1.3, p  = 0.28) than low achiever groups of students. The number of students with high achievement (39.5%) went through the lecture before the lesson was taught. These students had 1.08 times higher chances of achieving than low achiever groups of students. Furthermore, students who made a weekly study schedule had 1.3 times higher chances of being good academic achievers than those who did not (OR = 1.3, p  = 0.37). Additionally, high-achieving students paid closer attention to the lecturer (1.2 times higher). In addition, students with high GPAs spent more time studying when exam dates approached (OR = 1.3, p  = 0.58).

Table  4 demonstrates the relationship between memorizing and revising with high and low GPA students. It was found that high achiever students (58.9%) studied lectures daily and had 1.4 times higher chances of achieving high grades (OR = 1.4, p  = 0.16) than the other group. It was found that most of the high achievers (62.1%) skim the lecture beforehand before memorizing it, which led to 1.8 times higher chances of getting good grades in this exam (OR = 1.8, p  = 0.06). One regular activity reported by high GPA students (82.3%) was recalling what had just been memorized. For this recalling technique, we found a significant difference between low-achieving students (OR = 0.8, p  = 0.63) and high-achieving students (OR = 1.83, p  = 0.05). A high achiever student writes notes before speaking out for the memorizing method, which gives 1.2 times greater chances of getting high grades (OR = 1.2, p  = 0.55) than a student who does not write notes. A major difference in the current study was that high GPA achievers (70.2%) revise lectures more frequently than low GPA achievers (57.1%). They had 1.5 times more chances of getting high grades if they practiced and revised this method (OR = 1.5, p  = 0.13).

Table  5 illustrates the relationship between negative lifestyle factors and students’ academic performance. The current study found that students are less likely to get high exam grades when they smoke. Students who smoke cigarettes and those who vape are 1.14 and 1.07 times respectively more likely to have a decrease in GPA than those who do not smoke. Those students with chronic illnesses had 1.22 times higher chances of a downgrade in the exam (OR = 1.22, p  = 0.49). Additionally, students with high GPAs had higher mental pressures (Anxiety = 1.2, Depression = 1.18, and other mental pressures = 1.57) than those with low GPAs.

Learning is a multifaceted process that evolves throughout our lifetimes. The leading indicator that sets students apart is their academic achievement. Hence, it is crucial to investigate the factors that influence it. The present study examined the relationship between different study habits, personal characteristics, and academic achievement among medical students. In medical education, and more so in Saudi Arabia, there needs to be more understanding regarding such vital aspects.

Regarding motivational factors, the present study found some differences between high and low achievers. Students with high GPA scores were more motivated toward their future careers (OR = 1.67, p  = 0.09). The study also indicated that students who had ambitions and sought self-fulfillment were more likely to have high GPA scores, which were statistically significant (OR = 1.93, p  = 0.04). This was consistent with Bin Abdulrahman et al. [ 20 ], who indicated that the highest motivation was self-fulfillment and satisfying family dreams, followed by a high educational level, aspirations to join a high-quality residency program, and high income. Their study also found that few students were motivated by the desire to be regarded as unique students. We hypothesize that this probably goes back to human nature, where a highly rewarding incentive becomes the driving force of our work. Hence, schools should utilize this finding in exploring ways to enhance students’ motivation toward learning.

The present study did not find a significant effect of previous exam results on academic performance (OR = 1.03, p  = 0.88). However, some studies reported that more than half of the high-achieving students admitted that high scores acquired on previous assessments are an important motivational factor [ 15 , 25 , 26 ]. We hypothesize that as students score higher marks, they become pleased and feel confident with their study approach. This finding shows how positive measurable results influence the students’ mentality.

The present study also explored the social environment surrounding medical students. The results indicated that those who were supported by their friends or family were slightly more likely to score higher GPAs (OR = 1.1, p  = 0.73); however, the results did not reach a statistical significance. We hypothesize that a supportive and understanding environment would push the students to be patient and look for a brighter future. Our study results were consistent with previous published studies, which showed an association [ 3 , 27 , 28 , 29 , 30 ]. We hypothesize that students who spend most of their time with their families had less time to study, which made their study time more valuable. The findings of this study will hopefully raise awareness concerning the precious time that students have each day.

The association of different study habits among medical students with high and low GPAs was also studied in our study. It was noted that the high-achieving students try to attend their lectures compared to the lower achievers. This was in line with the previous published studies, which showed that significant differences were observed between the two groups regarding the attendance of lectures, tutorials, practical sessions, and clinical teachings [ 31 , 32 ]. The present study found that most students prefer to study alone, regardless of their level of academic achievement (82.1%). This finding is consistent with the study by Khalid A Bin Abdulrahman et al., which also showed that most students, regardless of their GPA, favored studying alone [ 20 ].

The present study findings suggest that a small number of students (29.8%) prefer to walk or stand while studying rather than sit, with most being high achievers (OR = 1.57, P  = 0.15). A study reported that 40.3% of students with high GPAs seemed to favor a certain posture or body position, such as sitting or lying on the floor [ 15 ]. These contradictory findings might indicate that which position to adopt while studying should come down to personal preference and what feels most comfortable to each student. The present study also found that high achievers are more likely to prefer studying early in the morning (OR = 1.3, P  = 0.28). The authors did not find similar studies investigating this same association in the literature. However, mornings might allow for more focused studying with fewer distractions, which has been shown to be associated with higher achievement in medical students [ 3 , 15 , 33 ].

Our study also found that 39.5% of the academically successful students reviewed pre-work or went through the material before they were taught it (OR = 1.08, p  = 0.75), and 25% were neutral. Similar findings were reported in other studies, showing that academically successful students prepared themselves by doing their pre-work, watching videos, and revising slides [ 3 , 9 , 34 ]. Our study showed that 75% of high-achieving students tend to listen attentively to the lecturer (OR = 1.2, p  = 0.48). Al Shawa et al. found no significant differences between the high achievers and low achievers when talking about attending lectures [ 15 ]. This could be due to the quality of teachers and the environment of the college or university.

Regarding the relationship between memorizing and revising with high and low GPA students, the present study found that students who study lectures daily are more likely to score higher than those who do not (OR = 1.4, p  = 0.16). This finding is consistent with other studies [ 3 , 19 , 35 ]. For skimming lectures beforehand, an appreciable agreement was noted by high GPA students (62.1%), while only (42%) of low GPA students agreed to it. Similarly, previous published studies also found that highlighting and reading the content before memorization were both common among high-achieving students [ 15 , 36 ]. Furthermore, the present study has found recalling what has just been memorized to be statistically significantly associated with high GPA students (OR = 1.83, p  = 0.05). Interestingly, we could not find any study that investigated this as an important factor, which could be justified by the high specificity of this question. Besides, when it comes to writing down/speaking out what has just been memorized, our study has found no recognizable differences between high-achieving students (75%) and low-achieving students (69%), as both categories had remarkably high percentages of reading and writing while studying.

The present study has found no statistical significance between regularly revising the lectures and high GPA ( p  > 0.05), unlike the study conducted by Deborah A. Sleight et al. [ 37 ]. The difference in findings between our study and Deborah A. Sleight et al. might be due to a limitation of our study, namely the similar backgrounds of our participants. Another explanation could be related to curricular differences between the institutions where the two studies were conducted. Moreover, a statistically significant correlation between not preferring the data being presented in a written form instead of a graphical form and high GPA scores have been found in their study ( p  < 0.05). However, a study conducted by Deborah A. Sleight et al. indicated that 66% of high achievers used notes prepared by other classmates compared to 84% of low achievers. Moreover, their study showed that only 59% of high achievers used tables and graphs prepared by others compared to 92% of low achievers. About 63% and 61% of the students in their study reported using self-made study aids for revision and memory aids, respectively [ 37 ].

The present study also examined the effects of smoking and chronic and mental illness, but found no statistical significance; the majority of both groups responded by denying these factors’ presence in their life. A similar finding by Al Shawwa et al. showed no statistical significance of smoking and caffeine consumption between low GPA and high GPA students [ 15 ]. We hypothesize that our findings occurred due to the study’s broad approach to examining such factors rather than delving deeper into them.

High-achieving students’ habits and factors contributing to their academic achievement were explored in the present study. High-achieving students were found to be more motivated and socially supported than their peers. Moreover, students who attended lectures, concentrated during lectures, studied early in the morning, prepared their weekly schedule, and studied more when exams approached were more likely to have high GPA scores. Studying techniques, including skimming before memorizing, writing what was memorized, active recall, and consistent revision, were adopted by high-achievers. To gain deeper insight into students’ strategies, it is recommended that qualitative semi-structured interviews be conducted to understand what distinguishes high-achieving students from their peers. Future studies should also explore differences between public and private university students. Additionally, further research is needed to confirm this study’s findings and provide guidance to all students. Future studies should collect a larger sample size from a variety of universities in order to increase generalizability.

Limitations and recommendations

The present study has some limitations. All the study’s findings indicated possible associations rather than causation; hence, the reader should approach the results of this study with caution. We recommend in-depth longitudinal studies to provide more insight into the different study habits and their impact on academic performance. Another limitation is that the research team created a self-reported questionnaire to address the study objectives, which carries a potential risk of bias. Hence, we recommend conducting interviews and having personal encounters with the study’s participants to reduce the risk of bias and better understand how different factors affect their academic achievement. A third limitation is that the research team only used the GPA scores as indicators of academic achievement. We recommend conducting other studies and investigating factors that cannot be solely reflected by the GPA, such as the student’s clinical performance and skills. Lastly, all participants included in the study share one background and live in the same environment. Therefore, the study’s findings do not necessarily apply to students who do not belong to such a geographic area and point in time. We recommend that future studies consider the sociodemographic and socioeconomic variations that exist among the universities in Saudi Arabia.

Availability of data materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations

Grade Point Average

King Saud University

Institutional review board

Statistical package for the social sciences

Jafari H, Aghaei A, Khatony A. Relationship between study habits and academic achievement in students of medical sciences in Kermanshah-Iran. Adv Med Educ Pract. 2019;10:637–43.

Article   Google Scholar  

Abid N, Aslam S, Alghamdi AA, Kumar T. Relationships among students’ reading habits, study skills, and academic achievement in English at the secondary level. Front Psychol. 2023;14:1020269.

Abdulghani HM, Al-Drees AA, Khalil MS, Ahmad F, Ponnamperuma GG, Amin Z. What factors determine academic achievement in high achieving undergraduate medical students? A qualitative study. Med Teach. 2014;36(Suppl 1):S43–48.

Muntean LM, Nireștean A, Sima-Comaniciu A, Mărușteri M, Zăgan CA, Lukacs E. The relationship between personality, motivation and academic performance at Medical students from Romania. Int J Environ Res Public Health 2022, 19(15).

Reza HM, Alireza HJIJME. Investigating study Habits of Library and Information Sciences Students of Isfahan University and Isfahan University of Medical Sciences. 2014, 14:751–757.

Kurtz SM, Silverman JD. The Calgary-Cambridge Referenced Observation guides: an aid to defining the curriculum and organizing the teaching in communication training programmes. Med Educ. 1996;30(2):83–9.

Pun J, Kong B. An exploratory study of communication training for Chinese medicine practitioners in Hong Kong to integrate patients’ conventional medical history. BMC Complement Med Ther. 2023;23(1):10.

İlçin N, Tomruk M, Yeşilyaprak SS, Karadibak D, Savcı S. The relationship between learning styles and academic performance in TURKISH physiotherapy students. BMC Med Educ. 2018;18(1):291.

McKeirnan KC, Colorafi K, Kim AP, Stewart AS, Remsberg CM, Vu M, Bray BS. Study behaviors Associated with Student pharmacists’ academic success in an active Classroom Pharmacy Curriculum. Am J Pharm Educ. 2020;84(7):ajpe7695.

Jouhari Z, Haghani F, Changiz T. Assessment of medical students’ learning and study strategies in self-regulated learning. J Adv Med Educ Professionalism. 2016;4(2):72–9.

Google Scholar  

Proctor BE, Prevatt FF, Adams KSS, Reaser A, Petscher Y. Study skills profiles of normal-achieving and academically-struggling College students. J Coll Student Dev. 2006;47(1):37–51.

Kyauta AMASY, Garba HS. The role of guidance and counseling service on academic performance among students of umar suleiman college of education, Gashua, Yobe State, Nigeria. KIU J Humanit. 2017;2(2):59–66.

Eva KW, Bordage G, Campbell C, Galbraith R, Ginsburg S, Holmboe E, Regehr G. Towards a program of assessment for health professionals: from training into practice. Adv Health Sci Education: Theory Pract. 2016;21(4):897–913.

Curcio G, Ferrara M, De Gennaro L. Sleep loss, learning capacity and academic performance. Sleep Med Rev. 2006;10(5):323–37.

Al Shawwa L, Abulaban AA, Abulaban AA, Merdad A, Baghlaf S, Algethami A, Abu-Shanab J, Balkhoyor A. Factors potentially influencing academic performance among medical students. Adv Med Educ Pract. 2015;6:65–75.

Ibrahim NK, Baharoon BS, Banjar WF, Jar AA, Ashor RM, Aman AA, Al-Ahmadi JR. Mobile Phone Addiction and its relationship to Sleep Quality and Academic Achievement of Medical students at King Abdulaziz University, Jeddah, Saudi Arabia. J Res Health Sci. 2018;18(3):e00420.

Alkhalaf AM, Tekian A, Park YS. The impact of WhatsApp use on academic achievement among Saudi medical students. Med Teach. 2018;40(sup1):S10–4.

Bonsaksen T, Brown T, Lim HB, Fong K. Approaches to studying predict academic performance in undergraduate occupational therapy students: a cross-cultural study. BMC Med Educ. 2017;17(1):76.

Alzahrani HA, Alzahrani OH. Learning strategies of medical students in the surgery department, Jeddah, Saudi Arabia. Adv Med Educ Pract. 2012;3:79–87.

Bin Abdulrahman KA, Khalaf AM, Bin Abbas FB, Alanazi OT. Study habits of highly effective medical students. Adv Med Educ Pract. 2021;12:627–33.

Jameel T, Gazzaz ZJ, Baig M, Tashkandi JM, Alharenth NS, Butt NS, Shafique A, Iftikhar R. Medical students’ preferences towards learning resources and their study habits at King Abdulaziz University, Jeddah, Saudi Arabia. BMC Res Notes. 2019;12(1):30.

Abdulghani HM, Alrowais NA, Bin-Saad NS, Al-Subaie NM, Haji AM, Alhaqwi AI. Sleep disorder among medical students: relationship to their academic performance. Med Teach. 2012;34(Suppl 1):S37–41.

Hwang G-J, Wang S-Y, Lai C-L. Effects of a social regulation-based online learning framework on students’ learning achievements and behaviors in mathematics. Comput Educ. 2021;160:104031.

Gamage KAA, Dehideniya D, Ekanayake SY. The role of personal values in learning approaches and student achievements. Behav Sci (Basel Switzerland) 2021, 11(7).

Linn Z, Tashiro Y, Morio K, Hori H. Peer evaluations of group work in different years of medical school and academic achievement: how are they related? BMC Med Educ. 2022;22(1):102.

Avonts M, Michels NR, Bombeke K, Hens N, Coenen S, Vanderveken OM, De Winter BY. Does peer teaching improve academic results and competencies during medical school? A mixed methods study. BMC Med Educ. 2022;22(1):431.

Topor DR, Keane SP, Shelton TL, Calkins SD. Parent involvement and student academic performance: a multiple mediational analysis. J Prev Interv Community. 2010;38(3):183–97.

Veas A, Castejón JL, Miñano P, Gilar-Corbí R. Relationship between parent involvement and academic achievement through metacognitive strategies: a multiple multilevel mediation analysis. Br J Educ Psychol. 2019;89(2):393–411.

Núñez JC, Regueiro B, Suárez N, Piñeiro I, Rodicio ML, Valle A. Student Perception of teacher and parent involvement in Homework and Student Engagement: the mediating role of motivation. Front Psychol. 2019;10:1384.

Abdulghani AH, Ahmad T, Abdulghani HM. The impact of COVID-19 pandemic on anxiety and depression among physical therapists in Saudi Arabia: a cross-sectional study. BMC Med Educ. 2022;22(1):751.

Park KH, Park JH, Kim S, Rhee JA, Kim JH, Ahn YJ, Han JJ, Suh DJ. Students’ perception of the educational environment of medical schools in Korea: findings from a nationwide survey. Korean J Med Educ. 2015;27(2):117–30.

Ahrberg K, Dresler M, Niedermaier S, Steiger A, Genzel L. The interaction between sleep quality and academic performance. J Psychiatr Res. 2012;46(12):1618–22.

Dikker S, Haegens S, Bevilacqua D, Davidesco I, Wan L, Kaggen L, McClintock J, Chaloner K, Ding M, West T, et al. Morning brain: real-world neural evidence that high school class times matter. Soc Cognit Affect Neurosci. 2020;15(11):1193–202.

Pittenger AL, Dimitropoulos E, Foag J, Bishop D, Panizza S, Bishop JR. Closing the Classroom Theory to practice gap by simulating a Psychiatric Pharmacy Practice Experience. Am J Pharm Educ. 2019;83(10):7276.

Walck-Shannon EM, Rowell SF, Frey RF. To what extent do Study habits relate to performance? CBE Life Sci Educ. 2021;20(1):ar6.

Abdulghani HM, Alanazi K, Alotaibi R, Alsubeeh NA, Ahmad T, Haque S. Prevalence of potential dropout thoughts and their influential factors among Saudi Medical Students. 2023, 13(1):21582440221146966.

Sleight DA, Mavis BE. Study skills and academic performance among second-Year Medical students in Problem-based learning. Med Educ Online. 2006;11(1):4599.

Download references

Acknowledgements

The authors are grateful to the Deanship of Scientific Research, King Saud University, for.

support through the Vice Deanship of Scientific Research Chairs.

Author information

Authors and affiliations.

Department of Psychiatry, College of Medicine, King Saud University, Riyadh, Saudi Arabia

Mohammed A. Aljaffer & Ahmad H. Almadani

College of Medicine, King Saud University, Riyadh, Saudi Arabia

Abdullah S. AlDughaither, Ali A. Basfar, Saad M. AlGhadir, Yahya A. AlGhamdi, Bassam N. AlHubaysh & Osamah A. AlMayouf

Department of Psychiatry, College of Medicine, Imam Mohammad Ibn Saud Islamic University, Riyadh, Saudi Arabia

Saleh A. AlGhamdi

Department of Medical Education, College of Medicine, King Saud University, P.O. Box: 230155, Riyadh, 11321, Saudi Arabia

Tauseef Ahmad

Department of Medical Education and Family Medicine, College of Medicine, King Saud University, Riyadh, Saudi Arabia

Hamza M. Abdulghani

You can also search for this author in PubMed   Google Scholar

Contributions

Conception or design: AHA, MAA, and HMA. Acquisition, analysis, or interpretation of data: AAB, SMA, ASA, YAA, BNA, OAA and SAA. Drafting the work or revising: TA, AHA, ASA AAB. Final approval of the manuscript: MAA, HMA., AHA, and TA. The author(s) read and approved the final manuscript.

Corresponding author

Correspondence to Tauseef Ahmad .

Ethics declarations

Conflict of interest.

The Authors declare that they have no conflict of interest.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Human ethics and consent to participate declarations

Additional information, publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Aljaffer, M.A., Almadani, A.H., AlDughaither, A.S. et al. The impact of study habits and personal factors on the academic achievement performances of medical students. BMC Med Educ 24 , 888 (2024). https://doi.org/10.1186/s12909-024-05889-y

Download citation

Received : 26 September 2023

Accepted : 12 August 2024

Published : 19 August 2024

DOI : https://doi.org/10.1186/s12909-024-05889-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Medical students
• Study habits
• Academic achievement
• Saudi Arabia

BMC Medical Education

ISSN: 1472-6920

• General enquiries: [email protected]

Supervisors’ emotion regulation in research supervision: navigating dilemmas in an accountability-based context

• Open access
• Published: 18 May 2024

Cite this article

You have full access to this open access article

• Jiying Han 1 ,
• Lei Jin 1 &
• Hongbiao Yin   ORCID: orcid.org/0000-0001-5424-587X 2  

969 Accesses

Explore all metrics

Given the complexity and high demands of research supervision and the intricate emotional experiences of supervisors, there is a need to explore how they regulate their emotions, particularly across various disciplinary backgrounds. The current study explored the emotion regulation strategies employed by research supervisors during the process of supervising graduate students. Based on data collected through semi-structured interviews, observations, and documentation from six research supervisors in different institutions in China, seven emotion regulation strategies employed by research supervisors were identified and further categorized into two groups, that is, antecedent-focused (prevention, intervention, reinterpretation, reconcentration, and detachment) and response-focused (suppression and expression) emotion regulation strategies. The findings shed light on the dilemmas faced by supervisors and the paradox aroused from the context-dependent and non-standardized nature of research supervision within an accountability-based managerial context. The implications for supervisors’ emotion regulation in authentic supervisory situations are discussed, and insights for universities’ policy-making are offered.

Similar content being viewed by others

Holding Abusive Managers in Contempt: Why and When Experienced Abusive Supervision Motivates Enacted Interpersonal Justice Toward Subordinates

Towards Safe and Equitable Relationships: Sociocultural Attunement in Supervision

“Take a step back”: teacher strategies for managing heightened emotions

Explore related subjects.

• Artificial Intelligence

Avoid common mistakes on your manuscript.

Introduction

Since the 1990s, educational research has undergone an “affective turn” as a result of the critique of the long-standing Cartesian dualism between emotionality and rationality (Zembylas, 2021 ). Over the following three decades, the dynamic and complex nature of teacher emotion has been explored from various perspectives and approaches (Agudo, 2018 ). Since emotion can significantly impact various stages of the teaching process, either facilitating or hindering it (Yin, 2016a , 2016b ), opportunities for emotion regulation can be identified in educational contexts at any time (Taxer & Gross, 2018 ). In higher education, although emotion regulation has been proven significant to teacher development and well-being (Xie, 2021 ), the majority of research has been conducted within the context of classroom instruction (Tao et al., 2022 ), leaving that of research supervision in graduate education unexplored.

In graduate education, emotion plays an important role in the supervisory process and relationship building which involves a series of emotional interactions essential for both supervisors and graduate students. The existing research has demonstrated an increasing need for supervisors to develop emotion regulation skills to cope with the challenges and provide emotional support in research supervision (Wollast et al., 2023 ). On the one hand, supervisors need to employ emotion regulation strategies in the challenging supervisory contexts, as accountability-based policies and the blurring of personal and academic relationships between supervisors and graduate students may trigger complex emotional experiences such as anxiety and worry for supervisors (Xu, 2021 ). On the other hand, the provision of support from supervisors is strongly linked to the emotional well-being and research success of graduate students (Janssen & Vuuren, 2021 ; Wollast et al., 2023 ). Specifically, supervisors’ emotion regulation plays a crucial role in providing emotional support to graduate students, which in turn has a positive impact on graduate students’ well-being and their belief about their further academic pursuits (Han & Xu, 2023 ; Wollast et al., 2023 ).

Of the limited research on emotion in graduate education, much has been conducted to investigate the influence of graduate students’ emotion regulation on their mental health and academic engagement (Saleem et al., 2022 ). However, there is a paucity of studies which have researched supervisors’ emotions and emotion regulation during the supervisory process. With the aim of unpacking how research supervisors employ emotion regulation strategies in real supervisory scenarios to effectively fulfill their roles, and to gain insights into the nature of research supervision, this qualitative study explores the emotion regulation strategies used by supervisors in the process of research supervision.

Literature review

Teacher emotion and emotion regulation.

Emotion, once considered inferior to cognition, has gained increasing attention in the social sciences, including in educational research (Han & Xu, 2023 ). The current recognition of the intricate interplay between emotion and cognition in teaching and learning highlights the importance of emphasizing teacher emotion in both teacher development and teacher well-being (Chen & Cheng, 2022 ). Emotion is complex and difficult to define (Chen & Cheng, 2022 ), and the connotation of emotion has shifted from an intrapersonal perspective to a relational one, emphasizing interactions between individuals and their environment during emotion generation (Campos et al., 2011 ).

Under the relational view of emotion, individuals can achieve social goals in most jobs involving interpersonal interactions through emotion regulation (Brotheridge & Grandey, 2002 ). Emotion regulation refers to “the processes by which individuals influence which emotions they have, when they have them, and how they experienced and expressed their emotions” (Gross, 1998 , p. 275). In the educational field, a growing interest of research in emotion regulation has emerged since the 1990s (Yin, 2016a , 2016b ; Zembylas, 2021 ), as teaching has been viewed as “an emotional practice” (Hargreaves, 1998 , p. 835). Due to the importance of emotion in teachers’ professional lives, it is crucial for teachers to regulate their emotions to achieve improved teaching and learning outcomes. Specifically, enhancing positive emotions can foster better teacher-student relationships, promote creativity in teaching, and strengthen students’ learning motivation; inappropriately managed negative emotions can have adverse effects on these aspects (Hargreaves, 1998 ). Although teachers’ emotion regulation has been widely examined (e.g., Taxer & Frenzel, 2015 ; Yin, 2015 , 2016a , 2016b ; Yin et al.,  2018 ) most studies, influenced by the concept of emotional labor, have mainly focused on two types of emotion regulation strategies: deep acting (the act of internalizing a desired emotion, matching expressed emotion with felt emotion) and surface acting (the act of altering emotional expression without regulating inner feelings) (Grandey, 2000 ; Hochschild, 1983 ). Comparatively, Gross’s ( 1998 ) process model of emotion regulation provides a more nuanced framework to examine teachers’ employment of a wider range of emotion regulation strategies. According to Gross ( 1998 , 2015 ), emotion regulation could be achieved through two main approaches: the antecedent-focused and response-focused approach. The former entails strategies that seek to avoid or regulate emotions by modifying the factors triggering emotion generation, which include situation selection, situation modification, attention deployment, and cognitive changes. The latter modifies an individual’s expressions and responses after the emotions have fully manifested, directly influencing physiological, experiential, or behavioral responses.

In recent years, the predominant focus of studies, guided by Gross’s ( 1998 ) process model, has been on investigating the motivations, strategies, and outcomes of teachers’ intrapersonal emotion regulation (e.g., Taxer & Gross, 2018 ; To & Yin, 2021 ; Xu, 2021 ). Teachers’ motivations for emotional regulation stem from their diverse teaching goals, including managing the impressions that various parties have of them, adapting to intensive educational reforms for survival, and enhancing students’ concentration levels (Hosotani, 2011 ; Xu, 2021 ). As for emotion regulation strategies, the existing literature has mainly been conducted under Gross’s ( 2015 ) model, and revealed a series of antecedent-focused (e.g., situation selection, attention deployment, and cognitive change) and response-focused strategies (e.g., suppression, relaxation, and avoidance) to cope with the ambivalent demands and enormous workload faced by teachers. Remarkably, certain strategies that reflect the unique nature of teachers’ work, such as genuine expression (Yin, 2015 ; Yin, 2016a , 2016b ) and interpersonal strategies (To & Yin, 2021 ), have been identified. Regarding outcomes of emotion regulation, genuine expression of emotion and cognitive appraisal strategies were found helpful to improve the effectiveness of classroom teaching and to maintain a balance between teachers’ professional and personal dimensions of their identities (Yin, 2016a , 2016b ). In contrast, suppressing, pretending, and restraining emotions may cause emotional dissonance and less received social support (Yin, 2015 ).

Emotion regulation and research supervision

In graduate education, supervisors’ emotional experiences are triggered by the complexity and high demands of research supervision (Han & Xu, 2023 ). The conflicting roles of taking responsibility for both supporter and supervisor simultaneously, the contradiction between supervisors’ high expectations of students’ learning autonomy and graduate students’ unsatisfactory performance, and the blurred boundaries between supervisory relationship and friendship (Han & Xu, 2023 ; Parker-Jenkins, 2018 ) are major challenges encountered by research supervisors. These challenges lead to various emotional experiences on the part of supervisors, including positive emotions, such as joy and love (Halse & Malfroy, 2010 ), and more prevalent negative emotions, such as anger, and disappointment (Sambrook et al., 2008 ). Given the diverse range of emotions that emerge during the supervision process, it is necessary for supervisors to employ various emotion regulation strategies to accomplish effective research supervision.

According to literature, emotion regulation is strongly associated with research supervision in three areas. First, effective research supervision requires a constructive and supportive supervisory relationship, which is facilitated by supervisors’ emotion regulation. As poorly managed supervision relationships contribute to low academic completion rates, supervisors are required to establish a respectful and caring relationship with their students (Halse & Malfroy, 2010 ). However, creating and maintaining such relationships can be challenging. Specifically, during the interactions with graduate students, supervisors are expected to offer emotional supports, including encouragement, motivation, and recognition based on students’ individual needs while ensuring that any critical feedback is delivered constructively (Lee, 2008 ). However, excessive emotional engagement or close relationships with students may hinder their ability to provide constructive criticism (Lee, 2008 ). As such, supervisors must strike a balance between offering emotional support and providing constructive feedback, thereby developing a successful educational partnership with their students.

Second, the emotional support provided by supervisors plays a positive role in facilitating graduate students’ research productivity and emotional well-being (Han & Wang, 2024 ; Wollast et al., 2023 ). In terms of research success, supervisors who encourage critical thinking and support constructive controversies tend to produce higher achievement and retention rates than those who adopt a directive and authoritarian approach (Johnson, 2001 ). Furthermore, emotional support from supervisors has been linked to higher levels of research self-efficacy and emotional well-being among graduate students (Diekman et al., 2011 ). Specifically, structure and autonomy support strongly influence graduate students’ feelings and expectations about their future academic success. Thus, in academic settings, supervisors should adopt effective emotion regulation strategies, offering constructive feedback, close guidance, and attentiveness to maintain graduate students’ motivation and mental well-being.

Third, effective emotion regulation is also critical for the well-being of research supervisors themselves. When faced with repeated frustrating events such as a lack of student progress and demanding requirements in accountability-based supervisory contexts, supervisors may experience feelings of exhaustion, particularly when they perceive their supportive efforts as being ineffective (Xu, 2021 ). Failing to regulate these negative emotions with effective strategies can lead to the accumulation and intensification of undesirable feelings, resulting in detrimental effects on supervisors’ well-being and job satisfaction, which may ultimately lead to their emotional burnout and disengagement (To & Yin, 2021 ).

So far, the very limited research on research supervisors’ emotion regulation in medical and scientific disciplines found that although supervisors use instructional strategy modification (e.g., directly pointing out students’ writing deficiencies), cognitive change (e.g., reappraising the relationship between students’ underachievement and their supervision), and response regulation (e.g., lowering their voice to calm themselves) to deal with negative emotions (Han & Xu, 2023 ), they still have difficulties in stepping out of negative emotions (Sambrook et al., 2008 ). Meanwhile, supervisors from different disciplines may use different emotion regulation strategies due to disciplinary differences in occupational challenges, societal expectations, and specific work environments (Veniger & Kočar, 2018 ). Therefore, it is necessary for researchers to investigate the emotion regulation of supervisors with different disciplinary backgrounds.

Based on the literature, underpinned by Gross’s ( 2015 ) process model, the present qualitative multi-case study aims to investigate the emotion regulation strategies employed by research supervisors from different disciplinary backgrounds. Specifically, the study seeks to answer this core research question: What strategies do research supervisors use to regulate their emotions during the supervision process?

As the in-depth understanding of supervisors’ emotion regulation strategies relies on the narratives of their journey of research supervision, we used narrative inquiry to explore supervisors’ lived experiences in supervising graduate students. Narrative inquiry emphasizes the co-construction of specific experiences by the researcher and participants (Friedensen et al., 2024 ; Riessman, 2008 ), which allows us to co-construct the meaning of emotion regulation with participants through qualitative data including interviews, observations, and documents.

Research context: Emphasizing the accountability of research supervision

The Chinese research supervision system has its roots in the nineteenth century, evolving alongside the development of graduate education (Xie & Zhu, 2008 ). Within this system, research supervisors play a crucial role in research-based master’s and doctoral education. In 1961, a supervisor accountability system was formalized, placing the responsibility on supervisors for overseeing students in research projects, journal publications, and dissertation completion. Under the guidance of supervisors, students engage in specialized courses, master the latest advancements in a specific field, and conduct research (Peng, 2015 ).

In recent years, with the rapid growth of graduate education in China, both supervisors and graduate students have expressed concerns about the quality of research supervision (Xu & Liu, 2023 ). Thus, national policies have been introduced to stipulate supervisors’ responsibilities and enhance the overall supervision quality, with a particular emphasis on the accountability of research supervisors. In 2020, the Accountability Measures for Educational Supervision, released by China’s Ministry of Education ( 2020 ), outlined a code of conduct for supervisors, emphasizing that supervisors bear the primary responsibility for cultivating postgraduate students. Specifically, supervisors are held accountable for various aspects of graduate students’ academic progress, including the quality of dissertations, academic conduct, and the appropriate utilization of research funds. Failure to fulfill these responsibilities may result in serious consequences, such as disqualification from supervising students or the revocation of teaching credentials.

Participants

To explore a wide range of emotional experiences and emotion regulation strategies that arise when supervising students at various stages of their academic journey, participants were purposively selected based on the following three criteria: (1) doctoral supervisors with the qualifications to oversee research-based master’s students and PhD candidates were considered, which allows us to gain insights into their emotions in supervising students at different academic stages; (2) supervisors with a minimum of 5 years of supervision were selected, as their long-term experience would provide a comprehensive understanding of the depth and evolution of emotion regulation strategies; (3) supervisors of both hard and soft disciplines were involved, as disciplinary features may significantly shape supervisors’ styles, potentially leading to their diverse emotions and emotion regulation strategies. Finally, six doctoral supervisors from four universities in China agreed to participate in the study voluntarily and were informed of the research purpose and ethical principles before the study. Table 1 provides a summary of the demographic information for all participants.

Data collection

The positionality statement is essential as the authors’ roles may influence the data collection process. Specifically, two authors are doctoral supervisors with rich experience in research supervision, and one author is a doctoral student. Participants for this study were recruited from the authors’ colleagues or recommendations from friends. In the spirit of self-reflexivity, we acknowledge our positions in research supervision and recognize that our relationships with participants may impact our collection and interpretations of the data. However, the authors had attempted to minimize the possible influence through continuous reflection, crosscheck, and discussions during the data analysis and interpretation.

To produce convincing qualitative accounts, collecting data from multiple sources including semi-structured interviews, observations, and documentation was employed in the current study from November 2022 to April 2023.

The primary source of data was individual interviews with each participant. To gather participants’ narratives of critical events in their research supervision, an interview protocol was designed according to our research purpose, but the interview questions were sufficiently flexible to enable the interviewer to adapt the content according to the specific interview situation. The interviews lasted between 120 and 150 min, during which the participants were asked to describe critical events in their research supervision, their emotional experiences, and whether and how they regulated their emotions. Follow-up questions were asked to gain a more profound understanding of their emotion regulation strategies when they provided surprising and ambiguous responses. Sample interview questions included “What emotions do you typically experience as a research supervisor?” and “Do you regulate your emotions induced by research supervision? If so, how?” All interview questions were presented in Chinese, the participants’ first language, and were audio-recorded and transcribed verbatim.

Observation was used to complement the data obtained from interviews. Before the observation, all supervisors and their students were informed about the research purpose and ethical principles. Then non-participant observation during their group and individual meetings proceeded only with their voluntary participation. Supervisors’ supervisory methods, activities, meeting atmosphere, and emotions of meeting members were recorded to supplement and validate the data collected through the interview. A short follow-up interview was then conducted with supervisors, focusing on their reflections on emotional events that occurred during the observed group and individual meetings.

Documentation was also used as a supplementary method. With the consent of the participants and their students, supervisors’ annotations and feedback on graduate students’ manuscripts, unofficial posts about supervision on social media (e.g., WeChat moments sharing), and chat logs between supervisors and students were collected to obtain additional information about the participants’ emotional experiences and supervisory practices. Table 2 presents the interview durations, the total minutes recorded during observations, the length of follow-up interviews, and the specific number and types of documents reviewed by both supervisors and students.

Data analysis

The analysis involved a three-level coding process (Yin, 2016a , 2016b ). First, interview transcripts were repeatedly read to label data excerpts that addressed the research questions. Initial codes were based on participants’ original perspectives and then iteratively refined and combined. Second, the coding system was organized according to Gross’s ( 2015 ) process model of emotional regulation, which distinguishes between antecedent-focused and response-focused strategies. Meanwhile, the study also remained open to other emotion regulation strategies that were evident in the empirical data. Third, the coding system was distilled to capture the nature of the identified strategies, resulting in three types of emotion regulation strategies. During the analysis process, the data were classified and organized using the NVivo software.

To strengthen the credibility of the data analysis, the interview transcripts were carefully examined multiple times to ensure that the data were accurately reflected in the coding scheme. Moreover, the coding scheme was collaboratively developed by the authors, and any discrepancies in classification were thoroughly deliberated to achieve mutual agreement. The final coding system, along with sub-categories and patterns, is presented in Table 3 .

In sum, seven emotion regulation strategies in research supervision emerged from the empirical data, which can be grouped into two categories, namely, antecedent-focused strategies and response-focused strategies.

Antecedent-focused strategies

Supervisors used antecedent-focused strategies to regulate the external situation and their internal cognition before the emotions were generated.

Prevention involves the prediction and avoidance of situations that may lead to undesirable emotional experiences during supervision prior to the generation of emotions. Prevention strategies were frequently utilized in the graduate student recruitment process and early stages of supervision, as a means of avoiding undesirable situations. On the former occasion, supervisors identified multiple recruitment indicators, such as research experiences and GPA, to avoid supervisory situations that may lead to negative emotions. This is commonly related to their former supervisory experience: “It was frustrating to supervise a student who was not invested in her work, so I have to implement a rigorous recruitment process to prioritize candidates who are truly interested in research, rather than rashly recruiting students” (P1-interview).

Supervisors remain vigilant once a supervisory relationship was established, as they are required by accountability-based policies to be responsible for students’ research performance and safety. Many supervisors stressed the significance of “establishing rules and regulations” (P4-interview) in the early stages of supervision to avoid infuriation and disappointment with students’ academic misconduct. Therefore, establishing an academic code of conduct is an effective prevention strategy for supervisors: “I’m frustrated by academic misconduct among students, as discovering data falsification in student-published articles holds me accountable, risking serious consequences for my academic career. So I frequently emphasize the need for high academic honesty and integrity standards” (P2-interview, observation).

Another concern that worried supervisors, especially those of science and technology, is student safety: “Whenever I hear about a laboratory explosion that causes student injuries, it makes me very nervous” (P3-interview, documentation). It is crucial for the institutions and supervisors to establish comprehensive laboratory safety rules and educate students on safety protocols before conducting experiments: “I told my graduate students: Failure to obey laboratory rules and lack of safety awareness can lead to immediate accidents that not only affect yourself but also pose a risk to other students” (P3-interview).

Intervention

Intervention is the most commonly employed strategy by supervisors to enhance the effectiveness of their supervision once a supervisory relationship is established. They employed various intervention strategies to improve students’ academic attitude and develop their academic ability.

Specifically, supervisors improved their students’ engagement and altered procrastination either by scaffolding their research or enforcing discipline and prohibitions. On the one hand, our participants acknowledged the importance of instructional scaffolding in the supervisory process.

We need to cultivate students’ interest so that they can actively engage in research. For instance, I often demonstrate interesting phenomena between the English and Chinese languages to generate my students’ curiosity. Then I am delighted to see their willingness to immerse themselves in linguistic research. (P5-interview)

On the other hand, some supervisors emphasized the enforcement of discipline in supervision. One supervisor expressed disappointment and dissatisfaction with the lackadaisical research atmosphere within the entire research group. In response, she implemented strict discipline and prohibitions to restrict students from engaging in activities unrelated to research in the office (P2-observation).

Finding a student watching a movie in the office angered me as it may disturb other students trying to focus on their studies. So, activities like watching movies and listening to music are not allowed in our office. By rigorously enforcing these rules, our research group was able to collaborate more effectively and ultimately achieve satisfactory results. (P2-interview)

Furthermore, intervention strategies were also used to enhance graduate students’ academic competency. Modifying supervisory activities was considered as a useful method. One supervisor shared: “We used to read literature in our group meeting together, but it was not effective. I felt frustrated and decided to change our meeting activities this semester.” As a result, the supervisor organized students to provide feedback on each other’s manuscripts in weekly group meetings, because “it was very effective in improving their writing abilities” (P1-interview, observation).

Interestingly, some supervisors opted to micromanage students’ research processes when they were disappointed with their research performance

At first, I encouraged students to independently identify research topics, but I later realized with disappointment that it was challenging for them to identify gaps in the existing literature. To make things more efficient, I started assigning research projects directly to help them complete their dissertation and meet the graduation requirements. (P5-interview)

Reinterpretation

Reinterpretation refers to the process of cognitively reappraising a supervisory situation from different perspectives to change its emotional impact. Supervising a graduate student who lacks interest in research was described as a “prolonged and painful undertaking” (P4-interview). However, one supervisor noted that: “Dwelling on negative emotions can be unproductive as it does not necessarily solve problems. Despite the challenging experience, I have gained valuable insights and will be better equipped to handle such situation” (P4-interview).

In addition to explaining the meaning of the situations from supervisors’ viewpoints, they reconsidered the events from graduate students’ perspectives to rationalize their unsatisfactory performance and procrastination. For example, supervisors understood students’ time arrangements when they procrastinated: “I used to become annoyed when students failed to submit assignments punctually… Now I know that students need a balance between work and rest. They need adequate time for rest” (P5-interview).

On occasion, supervisors reappraised the connection between students’ misbehaviors and the effort they invested from the perspective of the teacher-student relationship.

I felt angry when things happened, but I wouldn’t let that emotion affect my life. I see myself as a supervisor to students, not a parent, so I don’t hold high expectations for them. If students choose not to follow my guidance, it’s not my concern anymore. (P6-interview)

Reconcentration

Reconcentration is the strategy by which supervisors focus on another aspect of supervision or divert attention away from supervision with the intent of changing emotional consequences. Specifically, during the supervisory process, supervisors prepared themselves to be optimistic by reminding themselves of their students’ strengths: “I was anxious about a student who always made slow progress in research. But when I later realized that his incremental results were consistently good, indicating that he was very meticulous, I felt much better” (P2-interview, observation).

Apart from diverting attention during supervision in working environments, the participants highlighted the importance of balancing personal and professional life to manage negative emotions that may arise during supervision.

After giving birth, I realized that caring for a child demands a considerable amount of time and energy. Then I redirected my attention from supervising students to my family. Thankfully, my family provides a supportive environment, and the pleasant moments shared with my family members helped me overcome negative emotions associated with work. (P4-interview)

Detachment refers to the act of separating from or terminating the supervisory relationship to disengage from negative emotions. This strategy was often employed when intervention, reinterpretation, and reconcentration strategies were ineffective. When supervisors found that various proactive measures failed to resolve the challenges in research supervision, they experienced enduring feelings of helplessness, confusion, and distress. One supervisor expressed deep frustration, stating, “I’ve exhausted all efforts—careful communication with her and her parents, and providing my support during her experiments. Yet, she continued to resist making progress with her experiments and dissertation. I felt lost in supervising this student” (P4-interview). As a result, they have to release themselves from the emotionally harmful supervisory relationships.

Some supervisors chose to disengage, meaning they no longer actively push the student: “Continuing to push a student who refused to participate in research despite all my efforts would only increase my frustration. I have decided to let him go and will no longer push him” (P5-interview).

In some extreme cases that evoke negative emotions, supervisors even terminated the supervisory relationship.

Supervising this student was a painful experience as his inaction negatively affected the entire research team. Other students started following his behavior and avoided conducting experiments. It made me feel suffocated. I had to terminate my supervision to avoid any further negative impact on the team and myself… I felt relieved after he left. (P3-interview)

Response-focused strategies

Response-focused emotion regulation involves the use of strategies after an emotion has already been generated.

Suppression

Suppression involves consciously attempting to inhibit behavioral and verbal emotional responses. Although supervisors experienced negative moods during research supervision, some refrained from expressing these emotions to students. Certain supervisors believed that criticism hinders problem-solving. One participant explained, “While interacting with students, I found some are genuinely fearful of supervisor authority. In such cases, venting emotions on students only heightens their fear, makes them hesitant to express themselves or their confusion in research, and ultimately hinders their progress” (P1-interview). In addition, some supervisors believed that expressing anger or disappointment toward students could harm their self-efficacy in research. One supervisor stated, “Obtaining a master’s degree is a challenging journey, especially for novice researchers. Confidence is crucial for their success. As a supervisor, I refrain from expressing negative emotions as it can hurt students’ feelings and even damage their confidence” (P3-interview).

As mentioned by the supervisors above, expressing anger and disappointment to graduate students may not resolve issues but damage their self-efficacy. In challenging situations where negative emotions were hard to suppress, supervisors opted to temporarily suspend supervision activities or introduce new tasks to regain composure: “Sometimes revising students’ manuscripts can be a painful task. To avoid the risk of expressing negative emotions to them, I often temporarily suspend the revision. Sometimes I take a walk until I feel calmer and more collected” (P1-interview).

In supervision, expressing emotion is another effective strategy for regulating supervisors’ emotions. Although supervisors were aware that expressing negative emotions may sometimes negatively affect students’ feelings, the importance of their own emotional well-being was emphasized, as “expressing feelings helped me recover from negative moods faster” (P6-interview). However, supervisors had different expressive styles when interacting with their students.

Some supervisors expressed their anger and dissatisfaction to their students directly, through behavioral or verbal emotional responses. A supervisor recounted an incident, “During a phone call with her, I lost my temper because of her terrible attitude, and ended up throwing my phone” (P4-interview).

Interestingly, given that “graduate students are all adults” (P6-interview), some supervisors expressed their emotions more tactfully, taking care not to lose their temper and cause distress to their students. One supervisor “felt angry with a student’s poor writing.” However, instead of scolding the student directly, he made a joke during a one-to-one meeting, saying “It’s not that you wrote poorly. It’s that I am not clever enough to comprehend your writing.” The student laughed, and then the supervision was conducted in a relaxed atmosphere. The supervisor explained: “I do not hide my emotions but prefer to avoid losing my temper and instead use humor to guide my students better” (P5-interview, observation).

This study contributes to the existing literature on emotion regulation by providing detailed insights into how emotion regulation strategies were utilized by research supervisors. It also sheds light on the dilemmas supervisors encounter and the paradox between the context-dependent nature of research supervision and the accountability-based managerial context.

Supervisors’ dilemmas in research supervision

Our study demonstrated supervisors’ capacity to proactively employ diverse emotion regulation strategies when coping with difficulties in research supervision. It also revealed some paradoxical phenomena within the supervisors’ utilization of these emotion regulation strategies, highlighting the dilemmas they encountered in the context of research supervision.

In general, supervisors in our study demonstrated a higher tendency to employ antecedent-focused strategies for emotion regulation rather than response-focused strategies, which can alleviate their emotional burnout and enhance their well-being. Specifically, participants utilized intervention strategies as antecedent-focused strategies to improve the effectiveness of research supervision, rather than seeking consolation to alleviate generated emotions. Previous research has indicated that antecedent-focused strategies were associated with increased life satisfaction (Feinberg et al., 2012 ). By intervening in the emotion generation process at an early stage, these strategies can potentially alter the emotional trajectory, contributing to improved well-being among supervisors (Gross & John, 2003 ).

While supervisors displayed a strong inclination to utilize diverse strategies to enhance the effectiveness of their supervision, our findings unveiled two paradoxical phenomena in their emotion regulation strategies, indicating the dilemmas that supervisors faced in authentic supervisory situations. First, in antecedent-focused strategies aimed at modifying situations that may trigger negative emotions, numerous interventions and detachments highlighted the conflicts supervisors encountered as they strived to balance adequate assistance and excessive interference. Specifically, while participants in our study “inspired students through scaffolding” or “encouraged students’ autonomous learning,” they also “micromanaged students’ research process” or “enforced discipline” to enhance supervision efficiency. This pedagogical paradox concerning the choice between intervening and non-intervening approaches has generated ongoing debate in existing research (Janssen & Vuuren, 2021 ). Both approaches have the potential to evoke negative emotional experiences for supervisors and graduate students. Research found that a highly intervening approach has negative implications for both supervisors and graduate students (Lee, 2020 ). Students who have encountered autonomy-exploitative behavior from their supervisors, such as being restricted to specific research topics and methodologies, have reported experiencing negative emotions (Cheng & Leung, 2022 ). For supervisors, the burden of an intervening approach, the dissonance between supervisors’ expectations and students’ actual research progress, as well as students deviating from conventional practices (Han & Xu, 2023 ), all contribute to feelings of frustration, sadness, and exhaustion. Nevertheless, non-intervening approaches do not always fulfill the expectations of both parties either. Supervisors who encouraged graduate students’ autonomous action acknowledged the value of promoting their independent thinking, which has been identified as a significant predictor of students’ research self-efficacy (Gruzdev et al., 2020 ). However, students who initially expected their supervisors to play a leadership role felt dissatisfied and disappointed when supervisors were reluctant to offer explicit guidance (Janssen & Vuuren, 2021 ). This misunderstanding of supervisors’ intentions can ultimately generate negative effects on supervisors’ emotional experiences (Xu, 2021 ).

Another evident paradoxical phenomenon arises in the response-focused strategies employed after emotions have already been triggered. Although supervisors opted to suppress their negative emotional expression to safeguard the confidence and self-esteem of mature learners, there were instances when they outpoured their disappointment and anger to students, aiming to swiftly step out of their negative moods. The act of expressing and suppressing emotions highlights the dilemma of cultivating a mutually beneficial relationship that promotes emotional well-being for both supervisors and students. On the one hand, the existing literature emphasizes the importance of supervisors being sensitive to students’ emotional experiences (Bastalich, 2017 ). The inherent power imbalance in supervisor-student relationships may create a sense of student dependency on their supervisors (Friedensen et al., 2024 ; Janssen & Vuuren, 2021 ). Excessive criticism from supervisors can potentially lead to feelings of loss, and alienation throughout students’ academic journey, which highlights supervisors’ responsibility to manage their emotional criticism in supervisory interactions (Parker-Jenkins, 2018 ). On the other hand, although pursuing a research degree is a challenging journey for graduate students, it is important to acknowledge the vulnerability of research supervisors and their need for support (Parker-Jenkins, 2018 ). Power dynamics within supervisory relationships, particularly when students challenge or disregard supervisors’ advice, can lead to repression and disengagement for supervisors if negative emotions are not effectively regulated (Xu, 2021 ). Thus, recognizing supervisors’ needs and allowing for emotional expressions are also essential in developing a relationship that is mutually beneficial and conducive to the well-being of both parties (Parker-Jenkins, 2018 ).

The conflicts between research supervision and institutional policies

The dilemmas present in supervisors’ emotion regulation strategies inherently illustrate the context-dependent and non-standardized nature of research supervision. However, as modern higher education institutions move toward implementing accountability-based policies that aim to standardize and quantify research supervision (Jedemark & Londos, 2021 ), conflicts between the nature of supervision and these institutional policies not only place an emotional burden on supervisors, but also endanger the quality of graduate education.

The dilemmas observed in supervisors’ emotion regulation strategies highlight the divergent understandings between supervisors and graduate students regarding their respective responsibilities and the boundaries of the supervisor-student relationship. This divergence is influenced by context-dependent factors in research supervision, including the beliefs, motivations, and initiatives of the individuals involved (Denis et al., 2018 ). Due to the difficulty in achieving a perfect agreement on these context-dependent factors, it becomes challenging to establish a standard for what constitutes an ideal beneficial research supervision (Bøgelund, 2015 ). In authentic supervisory situations, the relationships between supervisors and graduate students can range from formal and distant to informal and intimate in both academic and social interactions (Parker-Jenkins, 2018 ). Therefore, research supervision is a highly context-dependent and non-standardized practice that relies on the capabilities of supervisors and students, which are shaped by their individual experiences and personalities.

This nature of research supervision underscores the significance of avoiding standardization and a “one size fits all” approach. However, as higher education institutions move toward a corporate managerial mode, research supervision is increasingly perceived as a service provided within a provider-consumer framework, and the fundamental aspects of research supervision are being reshaped to align with a culture of performance measurement, control, and accountability (Taylor et al., 2018 ). In modern academia, universities and institutions have established specific guidelines and protocols for research supervision, which require supervisors to follow diligently and take accountability in the supervision process (Figueira et al., 2018 ).

The presence of extensive external scrutiny or accountability ignored the context-dependent and non-standardized nature of research supervision, leading to adverse effects on both supervisors and graduate students. On the one hand, supervisors face significant pressure within an accountability-based context. They are expected to serve as facilitators of structured knowledge transmission, which is enforced through the demanding requirements and time-consuming tasks associated with supervisory practices (Halse, 2011 ). However, the distinctive characteristics of various disciplines and the interdependent relationship between the supervisory context and graduate students’ learning process are neglected (Liang et al., 2021 ). Such a narrow focus on knowledge transmission may pose potential threats to supervisors’ autonomy and academic freedom, generating their feelings of self-questioning, helplessness, and demotivation (Halse, 2011 ). Supervisors in our study reported many examples of emotion regulation strategies utilized to cope with performative and accountability pressures in their workplace. Specifically, the responsibility to ensure timely doctoral completions, prioritize students’ safety, and maintain accountability for those experiencing delays or violating research codes evoked feelings of nervousness, pressure, and insecurity among supervisors.

On the other hand, interventionist supervision within accountability-driven supervisory contexts is perceived as detrimental to students’ academic innovation (Bastalich, 2017 ). The prevailing environment of heightened performativity and accountability alters supervisors’ attitudes toward academic risk-taking, thereby influencing their supervisory practices (Figueira et al., 2018 ). For example, participants in our study utilized prevention and intervention strategies to mitigate potential negative occurrences. This included adopting a directive approach to supervise students’ work and dissuading them from undertaking risky or time-consuming methods to ensure timely completion. However, such micromanagement may stifle innovation, thereby inhibiting doctoral students’ development as independent researchers (Gruzdev et al., 2020 ). Providing pre-packaged research projects or excessive support may hinder students’ acquisition of essential knowledge, skills, and expertise required for their future pursuits, potentially obstructing their progress toward independent thinking (Gruzdev et al., 2020 ).

The conflicts between the prevailing shift from autonomy to accountability in higher education and the context-dependent and non-standardized nature of research supervision highlight the necessity for practice-informed evaluations for research supervision. This finding resonates with previous studies on policy-making in graduate education (Taylor et al., 2018 ), which emphasized the challenges of establishing evidence-based institutional policies to capture the intricate realities of supervision in practice.

Limitations

This study contributes to the understanding of research supervisors’ work by examining their emotion regulation strategies in authentic supervisory situations. However, certain limitations should be addressed for future research. First, the small sample size is a significant limitation, as only six supervisors participated. Future studies may increase the sample size and enhance diversity within the sample. Second, as our study only involved perspectives from research supervisors, future studies may consider incorporating the perceptions of both supervisors and graduate students and analyzing the level of convergence and divergence between the obtained results to enhance the validity of data collection.

Implications for practice

Despite being situated in China’s supervisory accountability system, our study holds broader implications in the global context. As the shift toward corporatized management models in higher education worldwide reshapes research supervision to align with performance measurement and accountability culture (Jedemark & Londos, 2021 ), our results offer implications for research supervision and policy-making beyond the Chinese context.

First, for research supervisors and graduate students, the intricate and dynamic nature of research supervision revealed in our study makes it challenging to offer direct recommendations for optimal emotion regulation strategies. Instead, supervisors are encouraged to adaptively employ a range of emotion regulation strategies in different supervisory situations to enhance their emotional well-being. Additionally, recognizing the context-dependent nature of research supervision, both research supervisors and graduate students are urged to take into account factors such as each other’s beliefs, motivations, and initiatives in their research and daily interactions.

Second, in light of the discrepancy between the current standardized accountability measures in higher education and the context-dependent nature of research supervision, it is imperative for universities and institutions to develop practice-based policies that are tailored to supervisors’ and students’ academic development, avoiding generic and assumed approaches. To effectively address the distinctive requirements of research supervision, policy-makers are strongly encouraged to implement multi-dimensional, discipline-oriented evaluation systems for supervisors in the future.

Data Availability

Data from this study cannot be shared publicly because participants may still be identifiable despite efforts to anonymise the data. Therefore, data will only be made available for researchers who meet criteria for access to confidential data.

Agudo, J. D. M. (Ed.). (2018). Emotions in second language teaching: Theory, research and teacher education . Springer.

Google Scholar  

Bastalich, W. (2017). Content and context in knowledge production: A critical review of doctoral supervision literature. Studies in Higher Education, 42 (7), 1145–1157.

Article   Google Scholar  

Bøgelund, P. (2015). How supervisors perceive PhD supervision-and how they practice it. International Journal of Doctoral Studies, 10 , 39–55.

Brotheridge, C. M., & Grandey, A. A. (2002). Emotional labor and burnout: Comparing two perspectives of “people work.” Journal of Vocational Behavior, 60 (1), 17–39.

Campos, J. J., Walle, E. A., Dahl, A., & Main, A. (2011). Reconceptualizing emotion regulation. Emotion Review, 3 (1), 26–35.

Chen, J., & Cheng, T. (2022). Review of research on teacher emotion during 1985–2019: A descriptive quantitative analysis of knowledge production trends. European Journal of Psychology of Education, 37 (2), 417–438.

Cheng, M. W. T., & Leung, M. L. (2022). “I’m not the only victim...” Student perceptions of exploitative supervision relation in doctoral degree. Higher Education, 84 (3), 523–540.

Denis, C., Colet, N. R., & Lison, C. (2018). Doctoral supervision in North America: Perception and challenges of supervisor and supervisee. Higher Education Studies, 9 (1), 30–39.

Diekman, A. B., Clark, E. K., Johnston, A. M., Brown, E. R., & Steinberg, M. (2011). Malleability in communal goals and beliefs influences attraction to stem careers: Evidence for a goal congruity perspective. Journal of Personality and Social Psychology, 101 (5), 902–918.

Feinberg, M., Willer, R., Antonenko, O., & John, O. P. (2012). Liberating reason from the passions: Overriding intuitionist moral judgments through emotion reappraisal. Psychological Science, 23 (7), 788–795.

Figueira, C., Theodorakopoulos, N., & Caselli, G. (2018). Unveiling faculty conceptions of academic risk taking: A phenomenographic study. Studies in Higher Education, 43 (8), 1307–1320.

Friedensen, R. E., Bettencourt, G. M., & Bartlett, M. L. (2024). Power-conscious ecosystems: Understanding how power dynamics in US doctoral advising shape students’ experiences. Higher Education, 87 (1), 149–164. https://doi.org/10.1007/s10734-023-00998-x

Grandey, A. A. (2000). Emotional regulation in the workplace: A new way to conceptualize emotional labor. Journal of Occupational Health Psychology, 5 (1), 95–110.

Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2 , 271–299.

Gross, J. J. (2015). Emotion regulation: Current status and future prospects. Psychological Inquiry, 26 (1), 1–26.

Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology, 85 (2), 348.

Gruzdev, I., Terentev, E., & Dzhafarova, Z. (2020). Superhero or hands-off supervisor? An empirical categorization of PhD supervision styles and student satisfaction in Russian universities. Higher Education, 79 (5), 773–789.

Gu, J., Levin, J. S., & Luo, Y. (2018). Reproducing “academic successors” or cultivating “versatile experts”: Influences of doctoral training on career expectations of Chinese PhD students. Higher Education, 76 (3), 427–447.

Halse, C. (2011). ‘Becoming a supervisor’: The impact of doctoral supervision on supervisors’ learning. Studies in Higher Education, 36 (5), 557–570.

Halse, C., & Malfroy, J. (2010). Retheorizing doctoral supervision as professional work. Studies in Higher Education, 35 (1), 79–92.

Han, Y., & Xu, Y. (2023). Emotional support from the perspective of extrinsic emotion regulation: Insights of computer science. Teaching in Higher Education, 28 (7), 1725–1743.

Han, J., & Wang, T. (2024). Exploring graduate students’ research characteristics, emotional exhaustion, mastery approach, and research career commitment: insights from the JD-R theory. Studies in Higher Education, 1–15.

Hargreaves, A. (1998). The emotional practice of teaching. Teaching and Teacher Education, 14 (8), 835–854.

Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling . University of California Press.

Hosotani, R. (2011). Emotional experience, expression, and regulation of high-quality Japanese elementary school teachers. Teaching and Teacher Education, 27 (6), 1039–1048.

Janssen, S., & Vuuren, M. (2021). Sensemaking in supervisor-doctoral student relationships: Revealing schemas on the fulfillment of basic psychological needs. Studies in Higher Education, 46 (12), 2738–2750.

Jedemark, M., & Londos, M. (2021). Four different assessment practices: How university teachers handle the field of tension between professional responsibility and professional accountability. Higher Education, 81 (6), 1293–1309.

Johnson, J. W. (2001). The relative importance of task and contextual performance dimensions to supervisor judgments of overall performance. Journal of Applied Psychology, 86 (5), 984–996.

Lee, A. (2008). How are doctoral students supervised? Concepts of doctoral research supervision. Studies in Higher Education, 33 (3), 267–281.

Lee, A. (2020). Successful research supervision: Advising students doing research (2nd ed.). Routledge.

Liang, W., Liu, S., & Zhao, C. (2021). Impact of student-supervisor relationship on postgraduate students’ subjective well-being: A study based on longitudinal data in China. Higher Education, 82 (2), 273–305.

Ministry of Education, PRC. (2020). The accountability measures for educational supervision [yanjiusheng zhidao xingwei zhunze].  http://www.moe.gov.cn/srcsite/A22/s7065/202011/t20201111_499442.html . Accessed 17 May 2024.

Parker-Jenkins, M. (2018). Mind the gap: Developing the roles, expectations and boundaries in the doctoral supervisor–supervisee relationship. Studies in Higher Education, 43 (1), 57–71.

Peng, H. (2015). Assessing the quality of research supervision in mainland Chinese higher education. Quality in Higher Education, 21 (1), 89–100.

Riessman, C. K. (2008). Narrative methods for the human sciences . Sage.

Saleem, M. S., Isha, A. S., Awan, M. I., Yusop, Y. B., & Naji, G. M. (2022). Fostering academic engagement in post-graduate students: Assessing the role of positive emotions, positive psychology, and stress. Frontiers in Psychology, 13 , 920395.

Sambrook, S., Stewart, J., & Roberts, C. (2008). Doctoral supervision... A view from above, below and the middle! Journal of Further and Higher Education, 32 (1), 71–84.

Tao, Y., Liu, X., Hou, W., Niu, H., Wang, S., Ma, Z., & Zhang, L. (2022). The mediating role of emotion regulation strategies in the relationship between big five personality traits and anxiety and depression among Chinese firefighters. Frontiers in Public Health, 10 , 901686.

Taxer, J. L., & Frenzel, A. C. (2015). Facets of teachers’ emotional lives: A quantitative investigation of teachers’ genuine, faked, and hidden emotions. Teaching and Teacher Education, 49 , 78–88.

Taxer, J. L., & Gross, J. J. (2018). Emotion regulation in teachers: The “why” and “how”. Teaching and Teacher Education, 74 , 180–189.

Taylor, S., Kiley, M., & Humphrey, R. (2018). A handbook for doctoral supervisors (2nd ed.). Routledge.

To, K. H., & Yin, H. (2021). Being the weather gauge of mood: Demystifying the emotion regulation of kindergarten principals. The Asia-Pacific Education Researcher, 30 (4), 315–325.

Veniger, K. A., & Kočar, S. (2018). The impact of academic discipline on university teaching and pedagogical training courses. Croatian Journal of Education, 20 (4), 1261–1298.

Wollast, R., Aelenei, C., Chevalère, J., Van der Linden, N., Galand, B., Azzi, A., Frenay, M., & Klein, O. (2023). Facing the dropout crisis among PhD candidates: The role of supervisor support in emotional well-being and intended doctoral persistence among men and women. Studies in Higher Education, 48 (6), 813–828.

Xie, A. B., & Zhu, Y. (2008). Retrospect and prospect of Chinese degree and graduate education development in the past three decades. Degree and Graduate Education, 11 , 19–29.

Xie, F. (2021). A study on Chinese EFL teachers’ work engagement: The predictability power of emotion regulation and teacher resilience. Frontiers in Psychology, 12 , 735969.

Xu, Y. (2021). Unpacking the emotional dimension of doctoral supervision: Supervisors’ emotions and emotion regulation strategies. Frontiers in Psychology, 12 , 651859.

Xu, Y., & Liu, J. A. (2023). Exploring and understanding perceived relationships between doctoral students and their supervisors in China. Humanities and Social Sciences Communications, 10 (1), 1–10.

Yin, H. (2015). The effect of teachers’ emotional labour on teaching satisfaction: Moderation of emotional intelligence. Teachers and Teaching, 21 (7), 789–810.

Yin, H. (2016a). Knife-like mouth and tofu-like heart: Emotion regulation by Chinese teachers in classroom teaching. Social Psychology of Education, 19 (1), 1–22.

Yin, H., Huang, S., & Lv, L. (2018). A multilevel analysis of job characteristics, emotion regulation and teacher well-being: A job demands-resources model. Frontiers in Psychology, 9 , 2395.

Yin, R. K. (2016b). Qualitative research from start to finish (2nd ed.). Guilford Press.

Zembylas, M. (2021). The affective turn in educational theory. Oxford Research Encyclopedia of Education. https://doi.org/10.1093/acrefore/9780190264093.013.1272 . Accessed 17 May 2024.

Download references

Acknowledgements

The authors would like to thank the participants who made this publication possible.

This work was supported by the Project of Outstanding Young and Middle-aged Scholars of Shandong University, Shandong University Program of Graduate Education and Reform (grant number XYJG2023037) and the General Research Fund of Hong Kong SAR (grant number CUHK 14608922).

Author information

Authors and affiliations.

School of Foreign Languages and Literature, Shandong University, Jinan, 250100, Shandong, China

Jiying Han & Lei Jin

Department of Curriculum and Instruction, Faculty of Education, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong, SAR, China

Hongbiao Yin

You can also search for this author in PubMed   Google Scholar

Contributions

Jiying Han: writing—original draft preparation, writing—reviewing and editing; Lei Jin: writing—original draft preparation, formal analysis; Hongbiao Yin: conceptualization, validation, writing—reviewing and editing.

Corresponding author

Correspondence to Hongbiao Yin .

Ethics declarations

Conflict of interest.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Han, J., Jin, L. & Yin, H. Supervisors’ emotion regulation in research supervision: navigating dilemmas in an accountability-based context. High Educ (2024). https://doi.org/10.1007/s10734-024-01241-x

Download citation

Accepted : 13 May 2024

Published : 18 May 2024

DOI : https://doi.org/10.1007/s10734-024-01241-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Emotion regulation
• Research supervision
• Accountability
• Graduate education
• Find a journal
• Publish with us
• Track your research