Introduction

Lesson 8 showed how the Cedar Valley team collects numbers through a survey, administrative records and a chart review. This lesson turns to the half of the study in which the data are words. Interviews and focus groups are the most common ways of collecting qualitative data in health research, and they are often the first methods students use in a research position or a community needs assessment.

1.1 What Interviews and Focus Groups Produce

A qualitative interview is a conversation with a purpose, in which a researcher asks open questions and follows up on the answers so that a participant can describe experiences, views and meanings in their own words. The data are what the participant says, how they say it, and what the researcher observes about the setting and the exchange. Those data suit questions about how and why: how people experience a condition or a service, how a process unfolds over time, what an event meant to the people involved, and why people act as they do in a particular setting.

Steinar Kvale offered two images of the interviewer that help clarify what the method assumes (Kvale, 1996; Brinkmann & Kvale, 2015). In the first, the interviewer is a miner who digs for facts that already sit in the participant's head, waiting to be extracted. In the second, the interviewer is a traveller who walks alongside the participant, asks questions along the way, and returns with an account that both of them shaped. Most qualitative health researchers work closer to the traveller image. The participant's account is produced in a particular conversation with a particular interviewer, so how the questions are asked, who asks them, and where the conversation happens all become part of the data. This is why later sections of this lesson spend time on question wording, interviewer conduct and field notes.

1.2 How Much Structure? Three Kinds of Interview

Interviews differ in how much the researcher decides in advance. A structured interview asks every participant the same questions in the same order, often with fixed answer options. It is a survey read aloud, and it belongs with the quantitative methods of Lesson 8. A semi-structured interview uses an interview guide: a planned set of open questions and follow-up prompts that the interviewer may ask in a flexible order, adding questions when the participant raises something unexpected. An unstructured interview starts from one broad topic or opening request (for example, "Tell me the story of your move") and lets the participant lead, with the interviewer asking few planned questions. Unstructured interviews are used in ethnography and in life-history and narrative research.

The Cedar Valley team chose semi-structured interviews. With 24 interviews across four communities and two interviewers (the graduate research assistant and the community research associate), every interview must cover the same main topics so that accounts can be compared, while participants remain free to raise things the researchers had not anticipated.

1.3 Three Methods and the Questions Each Answers

In-depth interviews

An in-depth interview is a one-to-one semi-structured or unstructured interview that explores a participant's own experiences, views and reasoning in detail. Interviews in health research commonly last 45 to 90 minutes. Because the conversation is private, in-depth interviews suit sensitive topics (loneliness, grief, money, sexual health, experiences of discrimination) and questions about individual histories, such as how a person's social life changed after a move or a bereavement. Each interview produces a rich account from one person, and the analysis compares accounts across people. The 24 Cedar Valley interviews are in-depth interviews with older adults who live alone.

Key informant interviews

A key informant interview is an interview with a person chosen because their role gives them specialized knowledge of a community, program or system (Marshall, 1996). The interviewer asks mainly about what the informant knows and observes, rather than about the informant's personal experience. Key informant interviews are often shorter and more focused than in-depth interviews, and they are common in needs assessments, program evaluations and the early, formative stage of a study. Their main limitation is that each informant sees the community from one position, so a clinic manager and a seniors' centre volunteer may describe the same town quite differently.

Before writing its interview guide, the Cedar Valley team held four key informant interviews: with the health authority's regional lead for seniors' health, a community connector in Riverside, a primary care nurse in North Bench, and, under the team's partnership agreement, the community health director at the Cedar Valley First Nations Health Centre. These conversations told the team where older adults gather, which roads close in winter, where internet service is poor, and how to approach potential participants respectfully. They shaped the guide described in Section 2.

Focus groups

A focus group is a facilitated group discussion, usually with about five to ten participants who share a relevant characteristic, led by a moderator who asks a small number of open questions. Focus groups grew out of the "focused interview" that Robert Merton and colleagues developed in the 1940s to study audience reactions to wartime radio and film (Merton & Kendall, 1946). They were taken up by market researchers and later by health researchers. Krueger and Casey (2015) suggest that groups of about five to eight participants suit most noncommercial topics. Sessions commonly run 90 minutes to two hours.

What makes a focus group distinct is the interaction among participants. Jenny Kitzinger (1994, 1995) argued that the way participants question, agree with, tease and contradict one another is itself data, because it shows the shared language, norms and disagreements of a group. A focus group of family caregivers, for example, may reveal that caregivers share an unspoken rule about not asking neighbours for help, which no single caregiver would think to mention in an individual interview. Focus groups also suit questions about reactions to an idea, a service or a draft resource, because participants build on one another's comments.

Focus groups are less suited to detailed individual histories, because each person speaks for only part of the session, and to highly sensitive personal topics, because the researcher cannot guarantee that other participants will keep what they hear private. Groups are usually composed of people who share a key characteristic, a practice called segmentation, so that participants feel able to speak freely. The Cedar Valley team holds separate groups for older adults, family caregivers, and clinic staff with community connectors, because caregivers may hesitate to criticize clinic services in front of clinic staff.

1.4 Matching the Method to the Question

The choice among methods follows from the question. Four considerations usually decide it: whether the question concerns individual experience or shared views and norms, how sensitive the topic is, who holds the knowledge the team needs, and what is practical for the participants (travel distances, hearing, mobility, work schedules and internet service). The table below shows how the Cedar Valley team matched its questions to methods.

1.5 How Many Interviews or Groups?

Quantitative studies calculate a sample size from a formula (HSCI 341 Lesson 2). Qualitative studies decide the number of interviews or groups by judgement, guided by two ideas. The first is saturation: the point at which new interviews add little new information relevant to the question. In one well-known study of 60 interviews, most of the codes the team eventually used had appeared within the first twelve interviews (Guest et al., 2006), although the number needed depends heavily on how varied the participants are and how broad the question is.

The second is information power (Malterud et al., 2016). The more relevant information a sample holds, the fewer participants a study needs. A study needs fewer participants when its aim is narrow, when its participants have very specific experience of the topic, when it draws on established theory, when the interviews themselves are rich, and when the analysis examines a few cases in depth rather than comparing many.

Section 2 turns from choosing a method to writing the guide that the Cedar Valley interviewers will carry into each conversation.

Introduction

Section 1 explained why the fictional Cedar Valley team chose semi-structured interviews and focus groups. This section shows how the team wrote, tested and revised its interview guide, and ends with the complete guide, which can serve as a template for other interview studies.

2.1 What an Interview Guide Does

An interview guide is the written plan for a semi-structured interview. It contains an opening script, a short list of open questions in a suggested order, follow-up probes and prompts for each question, a closing, and brief instructions for the interviewer. The guide serves three purposes. It keeps interviews comparable when several interviewers conduct them, it reminds the interviewer of every topic the research question requires, and it gives the research ethics board (Lesson 5) a clear picture of what participants will be asked. The guide is a plan for a conversation, so the interviewer may change the order of questions, skip a question the participant has already answered, and follow a new topic the participant raises.

Hanna Kallio and colleagues reviewed published advice on developing semi-structured interview guides and described five phases: identifying whether a semi-structured interview suits the study, using previous knowledge (the literature, experts and preliminary work) to identify topics, drafting a preliminary guide, pilot testing it, and presenting the complete guide in the study's reports (Kallio et al., 2016). The rest of this section follows those phases for the Cedar Valley guide.

2.2 From the Research Question to the Topics

A guide starts from the research question, which is usually too broad to ask a participant directly. Nobody can answer "How do you experience social connection after a move to a smaller town?" in one go. The team therefore breaks the question into sub-questions and writes one or more guide questions for each, using a question matrix: a table that links each sub-question to the guide questions that address it. The matrix shows at a glance whether any part of the research question has no guide question, and whether any guide question serves no part of the research question and could be cut.

The topics came from the literature the team summarized in Lessons 2 and 3, from the four key informant interviews (which pointed to winter roads, transport and internet service), and from the advisory group of six older adults, who asked that the guide cover everyday encounters in shops and clinics as well as family visits.

Question 8 links the interviews to the quantitative strand, whose PECO question asks whether loneliness is associated with emergency department visits. If interview participants describe going to the emergency department because they had nobody to call at night, that account helps the team interpret the linked administrative data.

2.3 The Shape of a Guide

Most guides follow the same four-part shape. The opening explains the study, confirms consent and permission to record, and tells the participant what to expect. A warm-up question asks for easy, descriptive information so the participant becomes comfortable talking. The key questions address the research question and take most of the time, moving from concrete, descriptive questions to questions about feelings and meanings, with the most sensitive questions placed in the middle once rapport has developed. The closing invites anything the participant wants to add, answers their questions, and ends on a forward-looking note so that the participant does not leave the interview immediately after discussing a painful topic.

2.4 Writing Good Questions

Good interview questions are open, so that they invite a description and cannot be answered with one word. Each asks about one thing at a time, uses the words participants use, and avoids signalling the answer the interviewer expects. Questions about concrete experience usually work better than abstract questions, because people find it easier to describe what happened than to state a general opinion. James Spradley called broad descriptive questions such as "Could you walk me through a typical week?" grand tour questions (Spradley, 1979). The tabs below show five common problems and how the Cedar Valley team revised each one.

2.5 Probes and Prompts

A probe is a follow-up question or action that asks the participant to say more about something they have already said. Probes are mostly generic ("Can you tell me more about that?") or built from the participant's own words, and much of an interviewer's skill lies in choosing the right probe at the right moment. A prompt is a planned reminder of a topic that the interviewer raises only if the participant does not mention it unprompted. The prompts under Question 3 of the Cedar Valley guide include telephone and video calls. If a participant describes their week without mentioning telephone calls, the interviewer asks about them; if the participant has already described their Sunday calls with a daughter, the interviewer does not ask again.

2.6 The Focus Group Question Route

Richard Krueger and Mary Anne Casey (2015) call the focus group guide a question route and describe five kinds of question, set out in the table below. A question route contains fewer key questions than an interview guide, because several people must respond to each question and to one another. Short activities, such as listing ideas on cards or ranking options, give quieter participants a way into the discussion. Section 3 presents the complete Cedar Valley caregiver focus group guide.

2.7 Piloting the Guide

A pilot interview is a trial run of the guide with someone similar to the intended participants, conducted to find out whether the questions are understood, whether they produce the kind of talk the study needs, and how long the interview takes. Kallio et al. (2016) describe three kinds of pilot testing: internal testing within the research team, review by experts or members of the participant group, and field testing in real interview conditions. Data from pilot interviews are usually kept out of the analysis, and participants in a field test should give consent in the same way as study participants.

2.8 The Complete Cedar Valley Interview Guide

The guide below is the version approved by the research ethics board and used in the 24 interviews. Words in square brackets are filled in for each interview. Open each part in turn.

Section 3 follows the guide into the field, where the interviewer must use it while listening closely to the person in front of them.

Introduction

The quality of qualitative data depends on what happens in the room as much as on the guide: whether the participant feels at ease, whether the interviewer listens and follows up, and whether the interviewer records what the audio cannot capture. This section covers preparation, interviewing, focus group moderation, field notes and reflexive memos, using the fictional Cedar Valley study throughout. HSCI 841 Lesson 4 develops each of these skills at graduate depth.

3.1 Before the Session

Preparation begins with the setting. The location should be private, quiet, accessible and convenient for the participant. For many older adults that means their own home; others prefer a private room at a library or community centre, where they will not be overheard by family members. Interviewers working alone in participants' homes follow a check-in procedure, in which a named team member knows the address and expected finishing time and receives a call when the interviewer leaves. Interviews with members of First Nations communities in Cedar Valley are arranged as the partnership agreement with the Cedar Valley First Nations Health Centre specifies, including who conducts them and how community protocols are followed (Lesson 4).

Consent follows the process described in Lesson 5, with written consent or recorded verbal consent given before the interview starts. The interviewer brings two recorders, because a recorder that fails cannot be replaced after the conversation is over. For older participants the team also considers hearing (a quiet room, with the interviewer seated face to face), vision (the large-print topic card), fatigue (an offered break at the halfway point) and the participant's preferred time of day.

3.2 Conducting an Interview

Rapport is the trust and ease between interviewer and participant that allows a candid conversation. It builds from small things: arriving on time, a few minutes of ordinary conversation, a clear explanation of the study, and evident interest in what the participant says. Active listening means attending closely to what the participant says and showing it, through eye contact, brief encouragers such as a nod, and follow-up questions that use the participant's own words. An interviewer who is listening actively hears the phrase that deserves a probe ("she was the one who found me this place") and asks about it before moving to the next question in the guide.

Interviewers aim to stay neutral. They avoid praising or disagreeing with answers, avoid giving advice during the interview, and avoid filling the conversation with their own experiences. If a participant asks for advice, the interviewer can offer to talk about it once the recording has stopped and can point to the resource sheet. The table below lists habits that new interviewers commonly fall into.

3.3 Moderating a Focus Group

A focus group is usually run by two people. The moderator welcomes participants, sets the ground rules, asks the questions, keeps the discussion moving and makes sure that everyone has a chance to speak. The assistant moderator manages the recorders and refreshments, deals with latecomers, takes notes on the discussion and on non-verbal reactions, and records the order in which people speak, which helps the transcriber attribute each comment to the right person. At the end of the session the assistant moderator may give a brief oral summary for participants to confirm. In the Cedar Valley caregiver group, the community research associate moderates and the graduate research assistant serves as assistant moderator.

Moderation depends on a few techniques. The opening question goes round the circle so that every participant speaks in the first five minutes, which makes it easier for quieter people to speak later. The moderator invites differences of view ("Has anyone had a different experience?"), because a group can drift toward agreement with the first confident speaker. Participants are encouraged to talk to one another rather than only to the moderator. Because focus group members hear one another, the moderator asks participants to keep what is said in the room private, while acknowledging that the team cannot guarantee that everyone will do so. The flip cards show responses to common challenges.

The complete Cedar Valley caregiver focus group guide

The guide below is for the family caregiver focus group: seven caregivers of older adults who live alone, meeting for 90 minutes in a community centre meeting room in Cedar City on a weekday evening. It follows the question route described in Section 2.6. Open each part in turn.

3.4 Field Notes

Field notes are the researcher's written record of what happened during and around a session that the audio recording cannot capture: the setting, who was present, non-verbal cues, interruptions, what happened before and after the recorder ran, and the researcher's first impressions (Phillippi & Lauderdale, 2018). Interviewers jot brief notes during or immediately after a session and write them up fully the same day, while memory is fresh. Many teams use a structured contact summary, a one-page form completed after each session that records the main points, what was new, and what to follow up next time (Miles et al., 2014). The template below is the one the Cedar Valley team uses.

3.5 Reflexivity and Reflexive Memos

Reflexivity is the researcher's ongoing, critical attention to how their own background, assumptions, position and actions shape the data they collect and the interpretations they make (Finlay, 2002). Because the interviewer helps produce the account, as Section 1 explained, their age, gender, culture, profession and relationship to the community all affect what participants say. A researcher's positionality is their social position relative to the participants and the topic. The Cedar Valley community research associate grew up in North Bench and is known to many residents. That insider position helps with recruitment and trust, and it may also make some participants more guarded about local matters. The graduate research assistant is in her twenties, grew up in a large city, and moved to the region for her studies, which gives her an outsider's view of small-town life and some shared experience of moving to a new place.

A reflexive memo is a short, dated note in which the researcher records their reactions, assumptions and decisions during data collection and analysis. Memos are kept in a research journal, discussed in team meetings and later support the methods section of a report. The consolidated criteria for reporting qualitative research (COREQ) ask authors to report the interviewers' characteristics and their relationship with participants, which reflexive memos help document (Tong et al., 2007). Lesson 12 returns to COREQ.

Section 4 follows the recording out of the room: how it is stored, how it becomes a transcript, and what choices the team makes along the way.

Introduction

Every Cedar Valley interview and focus group ends as an audio file, and most qualitative analysis works from a written transcript of that file. This section follows a recording from the interview room to a checked, de-identified transcript. It shows one short excerpt from the fictional study in three transcription styles, and it ends with the team's transcription plan, which is a worked example of such a plan.

4.1 Recording and Handling Audio

Participants must agree to be recorded, and the consent form says who will hear the recording, how it will be transcribed, where it will be stored and when it will be destroyed (Lesson 5). If a participant agrees to an interview but declines recording, the interviewer takes detailed notes instead and writes them up immediately afterward. Recordings are identifiable personal information, because a voice can identify a person and because participants often mention names and places. The team therefore names files with study codes only (for example, CV-INT-14_A.m4a for the main recorder and CV-INT-14_B.m4a for the backup), transfers them the same day to encrypted university storage, checks that the copy plays, and then deletes them from the recorders. Video interviews produce video files that show faces and homes; if the analysis needs only the words, the team keeps the audio and deletes the video.

4.2 Transcription Involves Research Decisions

Transcription is the conversion of recorded speech into written text. It can look like clerical work, but every transcript reflects decisions about what to keep. Spoken language is full of hesitations, false starts, overlapping talk, laughter and changes of tone, and a transcriber must decide which of these to write down and how. Oliver and colleagues (2005) described a range from naturalized transcription, which records every utterance in as much detail as possible, to denaturalized transcription, which removes features such as stutters, pauses and involuntary sounds so that the content reads clearly. The right point on that range depends on the analysis. A study of how people take turns in conversation needs every pause; a study of what older adults say helps them stay connected needs mainly the words. Teams agree on a written transcription protocol before transcription begins, so that every transcript follows the same rules (McLellan et al., 2003).

Each transcript begins with a header that records the study, the interview or group ID, the date, the type of location (home, community centre or telephone), the interviewer, the length of the recording, the convention used, who transcribed it and when, who checked it and when, and the date de-identification was completed. Speaker labels identify each turn (I for the interviewer and P14 for participant 14, or M and P1 to P7 in a focus group), and timestamps link the text to the audio, so that an analyst can return to the recording to hear tone or check a doubtful word.

Conversation analysts use a far more detailed system developed by Gail Jefferson, which marks overlaps, stress, pitch and the length of pauses to a tenth of a second (Jefferson, 2004). HSCI 841 introduces it; most health research does not need it.

4.3 Verbatim, Intelligent Verbatim and Clean Verbatim

Transcription services and textbooks describe three common levels of detail. Their names vary: some services use "intelligent verbatim" and "clean verbatim" as synonyms, and others define them differently from the definitions below. A team should therefore write down exactly what its convention keeps and removes, whatever label it uses. This course uses the following definitions.

Verbatim transcription (also called full or true verbatim) records every word and sound as spoken: fillers such as "um", "uh", "you know" and "like", false starts, repeated words, stutters, the interviewer's brief responses such as "mm-hmm", pauses with their length, laughter, sighs, interruptions and overlapping speech, with grammar exactly as spoken. Intelligent verbatim removes fillers, false starts, stutters and the interviewer's brief responses, but keeps the speaker's own words and grammar, including non-standard forms, and marks noticeable pauses, laughter and other sounds. Clean verbatim goes one step further: it also corrects slips of grammar, removes interruptions unrelated to the content, and keeps bracketed notes only where they change how the words should be read. Clean verbatim keeps the speaker's word choice, local expressions and repetition used for emphasis.

One excerpt in three forms

Comparing the versions shows what each one gains and loses. The verbatim version is the most faithful, but the fillers make the content slower to read, and it takes longest to produce. The intelligent verbatim version reads easily while keeping P14's own grammar, which some researchers regard as part of her voice. The clean version is the easiest to read and to quote, but it loses the sigh after "Quiet" and the interviewer's laugh, which showed the interviewer joining in with the care-home joke, a moment the interviewer's reflexive memo examined. The clean version keeps the laugh after "care home by Christmas", because without it a reader could take the remark as a plain complaint about her daughter. It also keeps "nobody knew me from Adam" and "a week, a whole week", because the idiom and the emphatic repetition are part of what she said.

The Cedar Valley team chose clean verbatim for its interview transcripts, with the rule that pauses, laughter and other sounds are kept wherever they change how the words should be read. The team's question concerns what participants experienced, and its planned analysis (Lesson 11) works with the content of what people say. The interviewers check each transcript against the audio, and the audio remains available throughout the analysis. A team planning to study how participants talk, rather than what they report, would need full verbatim with more detailed notation.

4.4 Automated Transcription and Privacy

Automated transcription uses speech recognition software to turn audio into a draft transcript, often with timestamps and an attempt to label speakers. Video meeting platforms, dedicated transcription services and open-source models that can run on a local computer all offer it. Automated transcription saves many hours, since manual transcription takes several hours for each hour of audio. Its drafts still need correction against the audio. Accuracy falls with background noise, overlapping speech, soft or hesitant voices, accents, local place names, words from Indigenous languages and medical terms, and error rates have been shown to differ across groups of speakers (Koenecke et al., 2020). Some speech recognition models have also been found to insert words or phrases that were never spoken, particularly during pauses (Koenecke et al., 2024). Many automated tools also drop fillers and smooth grammar by default, so a team that wants verbatim or intelligent verbatim transcripts must restore what the software removed.

The larger concern is privacy. Uploading an interview to a cloud service sends identifiable personal information to a company's servers, which may be outside Canada and may keep the audio or text, give staff access to it, or use it to improve the company's products. In British Columbia, public bodies such as universities and health authorities are subject to the Freedom of Information and Protection of Privacy Act, which sets rules on how they store and disclose personal information, and TCPS 2 Chapter 5 sets out researchers' duties of privacy and confidentiality. Automatic meeting transcripts and note-taking assistants can also record an online focus group without the team intending it, so moderators check these settings before every session. The questions below help a team decide whether a tool is acceptable.

Human transcribers raise their own privacy questions. A professional transcriber or research assistant signs a confidentiality agreement, receives files through a secure transfer method, and deletes them after delivery. Transcribers can also be affected by distressing material, so teams warn them about difficult content and offer a debrief.

4.5 De-identifying and Storing Transcripts

De-identification of a transcript replaces names and other identifying details with codes or bracketed descriptions such as "[daughter's name]". A separate linking key, a protected file that connects participant codes to names and contact details, is stored apart from the transcripts and held by one named team member (Lessons 5 and 9). In small communities, combinations of ordinary details can identify a person: a 78-year-old widow who moved to Kestrel Lake two years ago, stopped driving after cataract surgery and attends the library coffee morning may be recognizable to many residents. The Cedar Valley team therefore keeps community names in its analysis transcripts, because place matters to the research question, and generalizes them in reports ("a woman in her late seventies who moved to a small town in the region"). Transcripts and recordings from First Nations participants are handled as the data governance agreement with the Cedar Valley First Nations Health Centre specifies, including who may transcribe and access them and where they are stored (Lessons 4 and 5).

This plan, together with the guide in Section 2 and the recruitment plan from Lesson 7, makes up the data collection plan for the interview strand. COREQ, which Lesson 12 introduces, asks authors to report many of these decisions, including whether sessions were recorded, whether field notes were made, how long sessions lasted and whether transcripts were returned to participants (Tong et al., 2007).

Bringing It All Together

This lesson followed the qualitative strand of the fictional Cedar Valley Social Connection Study from the choice of method to a checked transcript. In-depth interviews suit personal and sensitive experiences, key informant interviews draw on people's specialized knowledge of a setting, and focus groups produce data from the interaction among participants. The choice among them follows from the research question, the sensitivity of the topic, who holds the knowledge the team needs, and what is practical for participants.

A semi-structured interview guide turns the research question into a small number of open, neutral questions with probes and prompts, linked to the question through a question matrix and tested in a pilot. In the field, rapport, active listening, a distress protocol and careful moderation shape the quality of the data, and field notes and reflexive memos record what the audio cannot. Transcription then turns the recording into text, and the choice among verbatim, intelligent verbatim and clean verbatim, the use of automated tools, and the handling of identifiable audio are research decisions that the team must document and justify.

The final reflection asks you to plan a complete small qualitative data collection study, drawing on all four sections.

Final Knowledge Assessment

These terms, tools and people appear in this lesson on collecting qualitative data through interviews and focus groups.