Trust,nuance,and care: Advantages and challenges of repeat qualitative interviews

Repeat interviews as a research method

Repeat interviews are not an uncommon or new practice (Vincent, 2013), and their use is widespread. Qualitative longitudinal researchers interested in participants’ changes, continuities, and even mundane experiences over time, have reflected upon several methodological issues related to repeat, or “revisit,” interviewing, even if they do not use the words “repeat,” “revisit,” or “serial.” These researchers met participants repeatedly, in what they call “waves” or “rounds,” to interview them over a longer time span than what is typical of cross-sectional research (Edwards and Weller, 2012; Henderson et al., 2012) and based on this practice, discuss the applications, advantages, and challenges of what we call repeat interviewing. Some of these methodological evaluations are by-products of their practice of following participants for an extended period of time (see Coltart and Henwood, 2012; Elder, 1998; Waldinger and Schulz, 2023), but a growing body of literature focuses specifically on the practice of what others call recursive interviewing (Edwards and Holland, 2013; Neale 2019, 2021; Thomson, 2009).

Edwards and Holland (2013: 7) argue that the decision to conduct repeat interviews should go hand-in-hand with defining the sample: “for each member of the sample you will have multiple interviews, and considerably more data than you would have from a one-off, snapshot study [so] you will need to consider the number of participants’ to avoid drowning in data.” Edwards and Holland further find that the time between sessions is a key benefit of repeat interviewing: it gives participants “more time to think, and to reflect on the memories elicited” (51). Similarly, Henderson and colleagues (2012: 25) assert that “each new interview brings the previous one to life in a powerful and illuminating way” (25), thereby increasing “reflexivity from researchers and participants” (29). And Neale (2021: 329) credits “recursive interviewing” with bringing descriptions interviewees made in previous sessions to the table to be updated and reconsidered. Thomson (2009: 35) merits reinterviewing as a method that “both privileges and decentres the individual” by making “visible contradictions between successive accounts” (14). Perera (2020; see also 2015) argues that conducting various interview sessions allows tracking participant reflexivity. Cycles of interviews and follow-up interviews make it possible to see if participants’ changes of position are temporary or permanent.

Regarding ethics, Rooney (2015) found that a series of interviews increases participants’ ability to evaluate their consent to the project. In single interviews, participants may feel that their opportunity to withdraw from the project disappears once the researcher departs, whereas repeated contact between interviewer and interviewee helps ensure that “ethical issues remain active during longitudinal qualitative research” (82). Similarly, Sörensson and Kalman (2018) noted that extended contact and repeated encounters with participants not only improve data quality but also is a way to care for the participants. As they said of their research: “The impression was that the research subjects felt confident in the situation” (718).

These methodological observations concerning what we call repeat interviewing in qualitative longitudinal research deal mostly with prospective applications—the “purest form” of longitudinal research that “follow[s] the same people in ‘real time’ capturing changes and continuities as they occur” (Neale, 2019: 5; see also Neale, 2021). Fewer observations and reflections on repeat interviewing exist based on projects that are longitudinal but not prospective or not longitudinal at all. Within this line of research, Vincent (2013: 342) found that repeat interviews support the “aim of providing ‘thick descriptions’ and holistic understandings.” Read (2018) assigned five qualities to what he called serial interviews: producing more complex and complete data; being able to elicit information from different angles and cross-check it; increasing trust between interviewer and participant; capturing longitudinal changes; and having more opportunities to observe participants interacting with their surroundings. Roos (2022: 429) argues that “repeated interviews” bring a more “realistic picture” and that interviews become more elaborate as sessions go by.

In sum, qualitative researchers have documented three areas in which repeat interviews enhance research: data quality and data transformation through reflexivity, researcher–participant rapport, and ethical protection of participants, although the focus has overwhelmingly been on data quality. In the analysis of interviews and log entries from our study, we noticed similar benefits. Importantly, however, our research design diverges from the prospective longitudinal studies that provide most of the knowledge on repeat interviews. We conducted a high-intensity, retrospective project that can be described as quasi-longitudinal. This research design “explores dynamic processes through hindsight, a gaze backwards in time from the vantage point of the present day” (Neale, 2019: 49). Some of the insights about repeat interviewing gleaned from prospective designs are applicable for retrospective ones but take on another dimension when the application is retrospective and involves several interviews done over a short period of time.

With this article we argue for and illuminate the use of repeat interviews in retrospective, quasi-longitudinal and high-intensity interview projects. We base our analysis on our experiences in the CRIMLA project, where a team of over 20 researchers conducted repeat interviews with hundreds of incarcerated individuals. More concretely, we discuss three advantages (trust, nuance, and care) and three challenges (cost, risk of losing participants, and emotional impact) of repeat interviews.

Participants, recruitment, and consent

From January 2022 to August 2023, a research team, including the authors of this article, interviewed over 350 incarcerated persons in seven Latin American countries. The participants were selected from among those sentenced for specific types of offenses: drug trafficking, kidnapping, murder, sexual offences, and violent theft. Each inmate was interviewed several times with some days, but preferably at least one week, between sessions. Each session lasted from 1.5 to 2.5 h. Interviews were based on an extensive interview guide emphasizing life phases and life trajectories. This included questions about family context, childhood, youth, adulthood, crime, drug use, violence, detention, legal process, life in prison, and perceptions of victims. The interview guide followed a life-course and life-story design while leaving plenty of room for researchers and participants to pursue their own interests. We made a point of letting participants tell their stories in their own way, and interviewers were free to probe topics of special interest not covered by the interview guide.

The procedure of selecting participants differed from country to country and even from prison to prison depending on the institutional possibilities. Our research team visited 29 prisons: three in Argentina, ten in Bolivia, five in Brazil, five in Chile, two in Colombia, two in Honduras, and three in Mexico. Of these, 5 were low security, 16 were medium security, and 8 were maximum security. The variation in how prisons were organized makes it difficult to generalize about Latin American prisons beyond their overcrowding, underfunding, and prisoner co-governance (Darke et al., 2021; Sozzo, 2022). Variety in prisons also meant that recruitment procedures differed. The penitentiaries of some countries, such as Colombia, provided us with confidential digital lists of all prisoners. From these we randomly chose participants to fill our quotas. In other countries, such as Brazil, researchers were given access to confidential physical prison archives, and researchers studied prisoner files one by one to identify participants. Finally, other countries, such as Bolivia, do not have a reliable system of prison registration. In those countries, researchers identified participants more organically by talking to prisoners and staff.

All inmates we approached received an oral and written explanation of the project and their rights in Spanish, Portuguese, or English. These explanations included the purpose of the project (learning about their lives), the range of questions in the interview (family context and childhood, life phases from childhood to imprisonment, circumstances of crime, arrest, legal process, life in prison, and victims), their freedom to decide whether to participate, the confidentiality of the conversation and the privacy measures taken by the project, and the lack of concrete benefits and drawbacks of participating. Written and oral explanations also included information about the number of interview sessions needed (three) to cover the extensive interview guide. Participants were not told about what we suspected could be advantages and challenges of repeat interviewing.

Consent to participate can be seen as problematic in institutions such as prisons. On the one hand, people might want to be nice or avoid confrontation and discomfort with researchers and authorities by agreeing to participate. On the other hand, and particularly when control over prison social life is knowingly in the hands of the prisoners as it is in Latin America, participants might hesitate to share details with outsiders out of fear of sanctions for “snitching.” To protect participants from coercion, we requested, and were allowed, to approach potential participants directly. To diminish risk for participants, we shared information about the project both with those who were interviewed and those who were not asked to be interviewed. Several people who were asked felt free to say no, and some who agreed to participate asked questions about the nature of the research before agreeing to be interviewed, enhancing their consent to participate. The three sessions also made it possible to monitor the threat of sanctions by other prisoners, which in almost all cases (except one we detail below) was not a problem. While many participants were hesitant and tentative during the first minutes of the interview, most expressed appreciation for the experience of being interviewed (Di Marco and Sandberg, 2023).

Research team and data

The research team was large: In addition to us, the project leader and coordinator, there were four interviewers in Argentina, two in Bolivia, seven in Brazil, one in Chile, three in Colombia, two in Mexico and five in Honduras. Fourteen team members self-identified as female and ten as male. The seven local researchers in Brazil were native Portuguese speakers; local researchers in the other countries were native Spanish speakers; the project leader's native tongue is Norwegian but he understands Spanish and conducted interviews in English. The project coordinator is a native Spanish speaker and conducted interviews in Spanish and Portuguese.

The team members were master's candidates (most in their mid- or late 20s, with experience from prison work or psychology practices), PhD candidates, and professors; their backgrounds were criminology, law, sociology, and psychology. The 24 interviewers across the seven countries worked as a team. They were connected not only by their shared responsibilities and a commitment to the project, but most importantly through regular interaction with other team members across the seven countries. Most team members and the project coordinator met regularly to exchange information, share experiences, pose questions, and receive feedback. Meetings included researchers from one or several countries, depending on the need and purpose of the meeting. Approximately every two months researchers met in a plenary session to share insights. Team members were also connected through WhatsApp groups that functioned as forums to pose methodological questions and exchange ideas.

We implemented various ethical procedures and safeguards to protect the integrity of the research team. Most of these procedures were described in the instructions for interviewers written by the project leader and shared with the team members before the fieldwork started:

Do not make yourself or the participants uncomfortable. Do not do an interview if you are unsure or feel unsafe about anything regarding the interview-setting, the participant or something else. If you feel really uncomfortable or unsafe in an interview, try to find a natural way to end it. In these cases, we will not use the interview and try to find another participant instead […]If you, or participants, during the course of the research require to talk with a psychologist or some other professional the CRIMLA project will facilitate this. You can at any time resign from this job, for personal or other reasons.

Researchers were expected to record field observations and experiences immediately after an interview in a research log. They were not expected to track changes in the interviewee nor the progression in their rapport with the interviewee over the course of the interviews. Yet, both elements were prominent when we analyzed the logs.

Before the interviews began, interviewers were told that we would analyze their logs to evaluate the project's methodology and publish on it. In the process of writing this article, we asked them to consent to our use of specific entries. The findings we present below build on their research logs as well as extracts from the interviews.

The research project was hosted by the University of Oslo, Norway. Therefore, we sought and obtained authorization from the Norwegian Agency for Shared Services in Education and Research (Sikt) to collect and store life stories. We also received authorization from local ethics committees in the seven Latin American countries where we conducted fieldwork. As opposed to our usual practice of using pseudonyms, in this article, we have chosen to provide the real first names of the researchers. We did this to emphasize the methodological dimension and to credit the team members—something they appreciated.

Trust

The success of an interview depends on the interviewer building rapport and a positive personal relationship with the interviewee (Mason, 2002). Repeat interviews meant having time to allow the personal relationship, on which the quality of an interview depends, to mature and develop (see Fujii, 2017; Read, 2018). Liza, the researcher in Chile, noted this early on in her field log:

At first, she was untrusting, not sure of giving an interview, she says she does not like talking about her stuff. At the end [of the first interview], she said she was willing to meet again, that it was fine, that maybe next time she will say why she is there [in prison]. [When we ended the second interview] she also asked when I would come again, like asking for company. She said next time she would tell me more about the murder she committed. I promised to come next time with newspapers and magazines. In the third meeting she received me happily; she knew I would come, and she waited for me. She asked about my son, part of the brief conversation we had before we began.

Several meetings facilitate a relationship between interviewer and interviewee that is different than when the interaction is limited to one exchange; trust and resonance usually develop through a process as interviewer and interviewee get to know each other. While it can be argued that people might trust an interviewer with their stories because it is liberating to talk about their lives to a stranger they will never meet again (Stark, 2016), the experience of our researchers was that repeat interviews increased trust and rapport. In a short time period, interviewer and interviewee developed a bond of trust, but researchers remained “outsiders” in most senses and retained the benefit of being a “fresh face” that such status entailed (Bucerius, 2013).

The increasing trust and rapport Liza noted is evident in the transcripts of the interviews as well. In the first and second interviews, she gave the interviewee space and an opportunity to talk about the murder she was convicted of. Here in the second interview, the participant is hesitant to talk about it: Liza:

Do you want to say a little about that or do you prefer that we move on to another topic? Because you told me that it didn’t bring you very good memories.

In the third interview, Liza gave again the participant another opportunity to talk about the murder. This time, the tone and interaction were very different: Liza:

You told me the other time, ‘well maybe later I’ll tell you why I’m here’.

We found the same pattern of increased rapport in most of the interviews our team members conducted across countries. Some participants were open and willing to share from the first moment, as María Victoria from Colombia wrote in her field log: “From the first to the third session, she was willing to tell stories. She liked telling stories, mainly those involving her children. She was very expressive, mainly with her body language. She laughed and screamed a lot—always happily.” On the surface, it might seem that speaking about one's children requires less trust than about a murder you committed, but it was common for participants to be hesitant at first and open up on gradually, even if it was about their children. The experience of Henrique, an interviewer in Brazil, was illustrative. In the first session, in a flat, monotonous voice, a woman interviewee told Henrique she had four children. In the second session, she mentioned her children again, this time on the verge of tears. She explained that she spends her days in prison counting the hours until she will reunite with them. When the longing is too intense, pictures of her children provide solace. In the third session she excitedly exclaimed, “If I knew you would come today, I would have brought the pictures!” After the interview ended, she asked Henrique to walk her back to her cell. She wanted to show him the pictures. Henrique waited outside her cell beside a guard until she returned with a school notebook clenched in her arms. Her finger ran over the images inside the book as she connected each child's face with the stories told in the three interviews. When she finished, she leaned in for a hug.

The researchers agreed that growing rapport was one of the most striking experiences from their fieldwork. Sveinung's notes from Mexico characterize the first interview with a prisoner as consisting of a lot of questions from the participant about the project, unfocused and rapid answers on the part of the participant, and a constant change of topic. In the second interview, however, the participant started by saying “well, tell me what this is really for, what do you want?” When told that the aim was to get the “story of your life, from the beginning to the end” the participant took more than two hours to provide a coherent story, key event after key event, with minimal interruption. Sveinung noted after this session that “it strikes me how different these interviews would have been if we only did one of them.”

It was common for participants to state early on that their closest family members did not have anything to do with their getting involved in criminal activity. Still, little by little, it often became clear in interviews that their family's role was more complex than support unrelated to crime. Subsequent interviews helped in establishing the trust needed to elicit more accurate descriptions. We have examples of this from all seven countries in the study, but an interview by Verónica in México is illustrative. In the first session, she asked the participant about her family:

Interviewee: My family is composed mainly of my mother, my precious mother. I love her, she is now aging. She suffers a bit from backache and that hurts in my heart because I cannot be beside her and give her a glass of water. Me, for taking the wrong path in life. She was never a bad example for me, never, she always wanted to educate me. She gave me an infinity of advice.

Vincent (2013: 341) describes the key benefit of repeat interviews as being “the quality of the relationship that develop[s] with participants over time.” Increasing trust is especially important when interviewing marginalized, vulnerable, and stigmatized groups who, confronted by systemic injustice, often develop an attitude of skepticism. Walton and colleagues (2022: 7) mention that a “cultural connection” can facilitate resonance and make participants feel safe. Garton and Copland (2010: 548) document how a prior acquaintance between the interviewer and interviewee is helpful. While it is not always possible or even desirable for a researcher to share a personal, social, or cultural background with the participants, repeat interviews allow for personal connections and a mutual set of references that increase trust and rapport.

Nuance

Repeat interviews produce more data about each participant, increasing both depth and nuance in their portrayals. Importantly, nuance, in this case, is not only more information but also contrasting and even contradictory information that illuminates different aspects of the same issue. Decades of studies of social identities (e.g., Jenkins, 2014) and storytelling (e.g., Frank, 2010) have revealed that people's stories about themselves are far from simple. They have multiple identities, and the stories they tell change immensely depending upon both the factual and storytelling contexts. Narrative ambiguity, multiple identities, and multifaceted stories are therefore key to understanding cultures and societies.

Several encounters provided opportunities to draw out the complexity and nuances of stories. In another of Liza's interviews in Chile, she spoke with a participant who emphasized how he missed his son even though he was not the biological father of the child. In the first interview, he said that he had treated him as his own: Interviewee:

It's my son. … I suffer in silence, I cannot express it, I cannot express it because I was there, the 9 months, I saw him, I received him. During childbirth? I was there, those things are not forgotten, first time in my life, for me he is my son, although they tell me that he is huacho, cachai or not? ¹ Those things break my soul. I can’t say anything, nothing, do you understand me? … That kills me, that's my pain. Recently his house burned down, and I can’t help him.

In the third interview, Liza asked the participant what was important in life for him and provided a couple of suggestions. She got a different version of the participant's relationship with his son: Liza:

Your mom. Your son?

The participant went on to explain that the son did not mean much to him anyway. We interpreted this as a flexible story that changed depending upon the perspective and timing of the conversation. This is common of all narrative engagements and a reason why some research recommends ethnography and several interviews instead of single interviews (Fleetwood and Sandberg, 2021). The participant harbored conflicting, fluid, and constantly changing feelings about his son. Arguably, when the participant could no longer have contact with him, he reinterpreted the meaning of the child, from the first to the third interview, to make it easier to live with the absence. While there is a grey zone between these examples and our examples above on the effects of increasing trust for getting more information, the nuances we describe here are understood more as instances where the participants were in an inner and, as yet, unresolved dialogue with themselves regarding how to interpret important dimensions, people, and events in their lives.

Repeat interviewing is not a guarantee for increasing trust and rapport but can still elicit more and richer information. What we describe in this section about nuance differs from what we said regarding trust in that there is not simply a development or increase of rapport but, rather, that information and stories told from different angles add shades and layers of complexity to the life stories and thus reveal a fuller account of the participants’ lives. When researchers talk to people more than once, participants reveal more information, present various facets of themselves, and nuance their assertions. They present them in slightly different ways thus adding variety to the stories they tell. It is not necessarily that something false is replaced with something true, but complexity is added—capturing better the intricacy of life itself—as the examples in this section indicate.

In Colombia, María Victoria interviewed a woman who gave various versions of her relationship with her mother. In the first interview, she blamed her mother for her drug addiction and sexual orientation. She also highlighted all the negative aspects of their relationship. “I hated it when I had to go home,” declared the participant, highlighting her mother's anger: “She was stricter, with that anger, with that pain, all bitter, so we were the ones who sucked, like, all that bitterness up.” During the second and third sessions, however, the participant stated that her mother was “sweet and considerate” and that one of her biggest regrets was not having appreciated her more:

Because my childhood was really very beautiful, right now I appreciate it and I see everything. Because before, ‘Oh, mom, so annoying, mom this and that.’ But it was teaching you, habits, principles, uh, things that really serve you as an adult that at that time you don’t recognize.

The intervals between interview sessions made it possible to tweak or rework questions and interview strategies to specific contexts, and revisit prominent topics. This is also common in qualitative longitudinal research and allows, in subsequent interviews, depth and nuance of topics mentioned in the first interview (Tabib-Calif and Lomsky-Feder, 2021). In Colombia and Brazil, David, the project coordinator would often listen to or read through the first two sessions and then identify gaps in the information, elements of special interest, or difficulties in eliciting responses, making it possible to tailor the interview guides for subsequent interviews so that additional nuances and depth could be brought to the fore. As Vincent (2013: 341) notes, repeat interviews “allow the researcher to seek clarification or additional information about issues raised in earlier interviews.” In an interview where the participant provided only scant information about her father, the tailored guide included the following question aimed at deepening and providing more nuance in the answers: “You told me that the fondest memory you have of your dad is when he gave you a doll. Why is that memory so special? What did you feel in that moment?” The participant, seemingly prompted by this detailed question, provided a lengthy answer that augmented the monosyllabic responses she had previously provided:

They made me return that doll! At that time I was living with my stepdad and my stepdad did not allow me to be in contact with my dad. But as I got older they let me get closer to him. We are very similar, for the bad! [Laughter] But we have a very good friendship, he has always been my confidante.

Allowing days between sessions meant that both interviewer and interviewee had time to reflect on the topics discussed, remember things that were missed, and solicit fuller accounts. Repeat interviews allowed the interviewer and interviewee to go over episodes, thereby creating more comprehensive accounts of particular events. Most importantly, however, repeat interviews facilitated more nuanced and multi-layered versions of participants’ lives. The different versions might be influenced by the stories being told at the moment in which a “fact” or reflection is brought into the conversation, the temporality and turn-taking in the conversation (Sacks, 1992), and the mood of the participant (and researcher). Having several storytelling occasions increases the chances of getting multiple, varied, and nuanced life stories. These nuances and variety can happen in single interviews as well (e.g., Sandberg, 2010), but when extensive ethnography or prospective longitudinal studies are difficult, doing repeat interviews helps bring out the complexity of social life.

Care

Doing single interviews can be seen as taking advantage of people for research purposes. As Kapiszewski and Wood (2021: 951) write, “brief, limited interactions with any particular human participant” might send the message that interviewees are only “evidence.” At worst, researchers “extract” data and then disappear. Especially in life-history interviews, participants open up and reveal a lot to the researchers, which can be painful if they never meet again. Although returning for follow-up interviews can on occasion be an imposition on participants’ time, this is a powerful way to show that the researcher cares about the participants and sees them as more than data. Historically, marginalized groups may carry a wound created by the many projects in which their knowledge was taken by outsiders who profited from it and never returned to their communities (Goyes et al., 2021). While not solving this problem entirely, repeat interviews create a more permanent bond both in the interviews and in the halls of the establishment, signaling to participants that they are not seen by the researcher as mere data sources.

Most concretely, high-intensity repeat interviews offer opportunities for interviewers to provide information to participants and help them in other ways. Contrasted to long-term prospective research, where the interval between encounters with participants is long, the quick succession of meetings of repeat interviews makes it possible to show care to participants in concrete, pragmatic ways. Eduarda in Brazil brought cookies for a participant who stated during an interview that she was hungry every Friday because the timing of her job tasks made it difficult for her to claim her meals. Magaly in Bolivia shared her knowledge of tai chi as an instructor with prisoners. Repeat interviews allowed her to offer continued instruction. Magaly reported:

Before the first session, I saw the participant was upset. I asked him if he was in pain. Usually, they say it's the chest, the throat, or the stomach. In this case, it was the chest. I told him about the chest organs and their function. Then I asked him to make a fist as if to hit, and I showed him how I gently hit my chest. I asked him to do the same. After the second session, he asked if we could do the same exercise because his throat ached. In this case, I asked him whether he had something urgent to say but didn’t dare to say it out loud. I was there to listen, not for the project, but just to help. Then I asked him to breathe with me, to look at his belly, and to focus only on his breath.

In repeat interviews, participants can also get information about and evaluate their participation along the way, making consent to participate a long-term, informed, and reflexive decision. Sarah in Brazil recorded similar experiences about how a participant reflected on her feelings during the sessions. She wrote:

At the start of the second session, the participant reflected on how it felt for her to participate in the second session and why she decided to continue. She did the same evaluation at the start of the third session.

While it can be argued that meeting several times makes it more and more difficult to say no, in our experience, repeat interviews allowed informed consent to be a continuous and collaborative process, one that included checks at various stages of the research to ensure that the participant understood the project and wanted to continue and identifying where there were possibilities of negotiating the form of the interview (see Todd-Kvam and Goyes, 2023). Milena in Argentina spent time before each interview answering the participant's questions about the informed consent. [First session]:

Previously, the prisoner had various questions about the anonymity and the confidentiality of the information. I tried to answer the questions and put him at ease. [Second session]: As in the first interview, before recording, the participant had various questions about the goals of the project and how the information would be used. I tried to answer all the questions.

Piotr in Honduras also discussed how repeat interviews were helpful in clarifying roles, obligations, and information about the research project. Between the first and second interview, he was asked questions about the project by other prisoners, and he kept evaluating how to provide the most comprehensive information not only to the participant but also others around him. He then got a print copy of the informed consent form for the other prisoners, which he gave to the interviewee to pass around. At the end of the second interview, the participant raised the issue: Interviewee:

this guy that is in front, here, that one right here … [points to a prisoner] the other day, he kinda, he asked me like this; ‘hey, these gringos … do they help you?’ And I look at him, right, after y’all leave that day I was here. Hours later, he asked me. So, I was watching, he asking me that in front of some people, right? So, I asked him, straight away: ‘Can you ask that question more clear please? What exactly do you mean with that question?’

The second and third interviews can be used to clear up misunderstandings, help participants understand better what they are part of, and, in Piotr's case, assist a participant in communicating the nature of the project to other prisoners. In this sense, repeat interviews can be seen as providing ethical protection and care for the participants.

In repeat interviews, participants also have more opportunity to say things that are important for them, making their participation more agentive. It is difficult to remember all that is important when being “thrown into” an interview for the first time, and it is not surprising that frequently participants began the second or third interview by stating that they had something important to say. For example, Sveinung's second interview with a woman in a Mexican prison began with a reflection about the previous session: Sveinung:

So, my first question is like since last time, have you thought about things from the interview that you think would be important to understand your life or understand why you ended up in prison?

Repeat interviews make possible multiple attempts to understand the complexity of people's social identities and stories. While long-term prospective repeat interviewing is necessary to capture identity changes as people's selves evolve through time (Thomson, 2009), high-intensity repeat interviewing can be better at documenting synchronic conflict and contradiction in the self of the present. Repeat interviews combined with flexibility to pursue issues across interviews enable interviewers to see various accounts of the same event or issue. As Thomson (2009) explains, by revisiting topics and collecting accounts in various sessions, repeat interviews enable documenting the constant, and sometimes contradictory, work that is part of the construction of the self.

Repeat interviews allow for better care for participants in that they can follow up on concerns they have and make sure they understand what consent and participation means. As Presser (2004) emphasizes, interviews are opportunities for presenting a moral self and making self-claims: “The research encounter is a venue for doing social problems work and social problems resistance.” Constructing such self-stories can be useful for psychological well-being and for venting emotions (Beck, 2005; Copes et al., 2015; Rosenthal, 2016), and when the presentation of these selves takes place over time, participants have more control over them and they can become more meaningful.

Challenges

Compared to single-interview research designs, repeat interviews have several disadvantages including higher costs, the risk of losing participants, and the emotional impact of the termination of an extended, personal connection. The first obvious and main problem of repeat interviews is that it demands a lot of resources. We were able to conduct a project based on repeat interviews in seven countries with hundreds of participants only because we had sufficient funding. Researchers who cannot access this type of funding will be limited in applying this method (see Goyes & Skilbrei, 2023), and their choice will often be between interviewing a limited number of people in depth (repeat interviews) or including more participants but interviewing them only once (single interviews). With limited time and resources, it can be difficult to decide what is best: 15 repeat interviews or 45 single ones. The research context and questions will be the determinants.

Moreover, if arranging the details of a single interview is challenging on occasion, coordinating three meetings triples the complexity. Our participants were in confinement, with sentences that were usually at least two years in length—both elements should seemingly reduce complexity. Yet, we struggled to arrange appointments and sometimes had a difficult time locating participants because they had been moved to another unit. In some cases, albeit rarely, participants decided that they did not want to do more interviews. Liza noted in her field log that the participant was very emotional (crying) in the first interview, while in the second she “tells about her fights, … is sympathetic, laughs a lot, speaks fast.” At the end of the session, she is still “clear that she wants to leave; she asks for the time and we wrap up. She did not want to attend the third session.”

Another pragmatic concern worth noting is that some journals have set a minimum number of participants on which an article must be based to be published. We have had papers based on repeat interviews with 12 participants rejected outright for this reason alone by research journals who publish qualitative studies. These well-regarded sociological and criminological journals do not consider the difference between repeat- and single-interview research designs. This might also reflect how others in less formal arenas see and value qualitative interview research: one might, for example, do 45 interviews but only get credit for 15. While this is inexplicable for most qualitative researchers and ethnographers, it might be something to consider, especially for junior scholars with limited resources and who are under pressure to publish, but also for journal editors and peer reviewers.

The most important challenge with repeat interviews, however, is the shadow side of what we described above as the advantage of developing close personal relationships: the weight of building a stronger bond. María Victoria emphasized that “you are emotionally invested in the interviewee, and it is a tiny mourning process when the sessions are over.” One participant, she wrote, “tried to find me every time to have a new session. After the three sessions finished, she expressed sadness about that.” Sveinung realized that he had not adequately informed one of his interviewees about the sequence of interviews, causing an awkward and sober ending to the third session. When the participant said that she was looking forward to the next meeting, he got a lump in his throat when he had to tell her this was the last one. He noted that he was “sad not to be able to see her again.” And as McCracken (1988: 26) warned, researchers must establish a social connection with participants, but “it is possible to go too far and allow the intimacy to obscure or complicate the task at hand.”

The shadow side to the increasing trust that develops between interviewer and interviewee is, therefore, all the more pronounced upon exiting: by getting people to trust and give more across sessions, it is possible to exploit participants for information. SmithBattle and colleagues (2018: 6) argue that some of “the risks associated with single in-depth interviews, such as an unintentional disclosure, are well known and may be intensified with repeated contacts.” After all, in most cases, researchers leave behind the field, the participants, and the relationships they have built. The research advances their careers, contributes to their reputations, and increases their salaries. Participants get little in return except hope that the system will change.

The challenge of breaking bonds with participants after having established a relationship with them is an established ethical issue in the qualitative research literature on methods (e.g., Taylor, 1991: 238). Under the term of fieldwork “exit,” this literature “has centred on the possible impacts for participants from what could be seen as a form of harm or exploitation” (Watts, 2008: 10). Morrison and colleagues (2012) show that the most common “exit strategies” include giving presents and letters of thanks at the end of the study, hosting social farewells, sharing celebratory meals, distributing certificates of participation, and when possible, inviting participants to keep in touch. In prison, however, many of these strategies are institutionally and practically impossible. Our approach was to engage in what Morrison and colleagues (2012: 424) term “meaningful dialogue around closure”: from the first contact to the last we explained the project's design to participants, and in the final meeting, researcher and participant debriefed. However, researchers and participants testified to the distress of parting ways.

Departure was, however, not always problematic. In some cases, both interviewers and interviewees felt relief when the three sessions were over. Piotr describes a third interview session like this:

I was under the impression that she has nothing new to say even when I was trying to ask some new questions and motivate her to give more extensive answers. She told me that everything she had to say, she had already said in the first two sessions.