The Interpretive Interview. An Interview Form Centring on Research Participants’ Constructions

Abstract

One might be tempted to ask: Is discussing qualitative interviewing with yet another name tag on it once again old wine in a new bottle? In this contribution, I argue that describing our interviewing and data analysis still needs some reflection: It is a fine cuvée (ending the vinophile metaphor here) that you might want to get a taste of herewith: Discussing qualitative interviewing deepens the understanding of the kind of interviewing Kathy Charmaz proposed while building explicitly on its traditions within social scientific hermeneutics. Charmaz was kind enough to refer to my work – so it is my honour to proceed with my thinking following our time together in Vienna. In this paper, I will focus on the way of doing qualitative interviews that in the end leads to a data quality that, to the largest extent possible, fits our exigencies in interpretively oriented research. Put in a nutshell, we are seeking an interview form that puts the interview partners at the core of the communicative act and in the drivers’ seats of the interview situation. We thus end up with interview material that is rich in stories told and connections made at the free choice of the participants. This stands in contrast to the idea that qualitative interviews should always be based on interview guidelines. The latter can be clearly differentiated from the interpretive interview style proposed here – and is per se not a bad choice as long as it fits the overall methodology and the ensuing research interests.

Introduction

“An interested interviewer can engage interviewees in conversations that reveal their narrative constructions rather than gloss over them. Scheibelhofer’s initial approach contains a cautionary tail for interviewers: Do not assume that your research participants would describe themselves, their actions, and situations in the same way you do. Instead, use – and explore – the interview participants’ terms. (…) Participants’ terms often convey tacit discourses, stake claims about personal identity, and reveal distinctions they make between themselves and others.” Kathy Charmaz (2014)

In writing this manuscript ¹ and rereading Kathy Charmaz’ reflections on how to conduct ‘intensive interviews’ (Charmaz, 1991; 2014), I stumbled again across this passage in which she was kind enough to cite my work. Indeed, the lines she wrote go to the core of what I now am trying to clarify even more so than in the past: The need for an interview form that puts the participants in the drivers’ seat of our exchanges so that we can work with interview material that contains the results of meaning-making processes of the participants themselves. Unlike most scholars in grounded theory (GT), Kathy Charmaz was very concerned with interviewing and saw the need, in particular, to clarify the ways she carried out interviews (Charmaz, 1991; 2014): Intensive interviewing is an open-ended form of interview that Charmaz characterized as “in-depth” (Charmaz, 2014, p. 85). By means of her examples, it becomes evident that through the researcher’s questions, ad hoc theoretization is aimed at together with the interviewed person. Otherwise, scholars in GT have only given scant attention to data qualities. Charmaz saw the reason for this disregard in one of Barney Glaser’s famous sayings: “In essence, grounded theory is a method of data analysis with the intent of constructing theory.” Until recently (Charmaz, 2014; 2015; Charmaz & Belgrave, 2002 2002; Clarke, 2005; Mauthner & Doucet, 2008; Scheibelhofer, 2008), grounded theorists rarely attended to data collection in detail – maybe also induced by another famous saying of Glaser, namely: “All is data” (2001: 145). “These grounded theorists argue that the quality and quantity of data is not problematic if the analyst achieves ‘saturation’ of categories. Yet they do not delineate useful criteria for what should constitute either viable categories or saturation (Froschauer & Lueger, 2020). Consequently, some grounded theory studies skimp on data collection and tout description as theory” (Charmaz & Henwood, 2017, p. 243).

Juliet Corbin stated that Anselm Strauss himself also favoured open-ended interviewing over more structured forms such as guideline interviews (together with Morse Corbin herself refers to this kind of interviewing also as ‘unstructured interactive interview’, see Corbin & Morse, 2003; Morse & Clark, 2019):

“The question may be revised as the research goes along, but there has to be some initial statement in order to write a proposal. Also, for a proposal a researcher may have put together a preliminary interview or observational guide. However, Strauss generally advised doing open-ended interviews because they produced denser and richer data.” (Corbin, 2021, p. 31, p. 31)

Such open-ended interview styles have been described in detail already. In order to situate what I here propose to call the interpretive interview among the various forms of qualitative interviews (as I have done elsewhere in detail including many examples from my own empirical research, see Scheibelhofer, 2008), it seems helpful to differentiate between interview forms that rely on the idea of asking questions that are of interest to the researcher (topical or guideline interviews) and those that rely on theories of the narrative.

When collecting data via interviews for constructivist GT (CGT) analyses, we also carry out interviews that are more structured (e.g., expert interviews), but foremost, we talk to people who are involved in certain social experiences or situations. Especially in the phases of initial coding (Charmaz, 2014), we aim to gather rich data which reflect less the researchers’ structuring and meaning-making – in contrast to guideline-based interviews where the researcher introduces the topics assembled in advance based on literature reviews. Rather, we seek complex accounts or narratives, in which the interviewed persons may decide which issues need to be brought up, at what point of time during the conversation and connecting the stories among one another as is deemed necessary within the interview situation (see for a similar argumentation Lueger & Vettori, 2014).

During our discussions when Charmaz taught a couple of times CGT classes in Vienna, she was intrigued by the German-speaking developments of interviewing in the interpretive social sciences: She saw a great potential for GT scholars in learning more about objective hermeneutics (Oevermann et al., 1979; 2021) and social scientific hermeneutics (Soeffner, 2004) – in general, what today is commonly discussed as the hermeneutic sociology of knowledge as well as communicative constructivism (Knoblauch, 2019).

Scholars in GT and social scientific hermeneutics have for many decades been in close exchange over their take on social theory: Already Anselm Strauss and Hans-Georg Soeffner – the latter being another regular visiting professor at our Viennese Department of Sociology – had extensive discussions about the similarities and differences in their understanding of interpretive sociology and the ensuing methodologies. They met in person during their cross-Atlantic research stays (Reichertz, 2009): One of Strauss’s visits to Germany took place in the early 1980s, leading to encounters with numerous scholars of the hermeneutic research field, including Soeffner, and members of Richard Grathoff’s and Fritz Schütze’s teams (Reichertz, 2009). An ongoing exchange of ideas as well as friendship developed between Soeffner and Strauss, which lasted until the latter’s death (Soeffner, 2018).

The exchanges between these two scholars also had a well-founded philosophical and social theory basis – most of all in Mead’s writings: Pragmatism played a crucial role in the works of both (Corbin, 2021; Soeffner, 2014). The key importance of Mead’s thinking to Strauss is also – and this is the point at which I come back to the main argument for an interpretive interview form – the reason why Strauss perceived intepretations as central to theory-building in GT: “Theory is not the formulation of some discovered aspect of a pre-existing reality ‘out there’ […] Theories are interpretations made from perspectives as adopted or researched by researchers” (Strauss & Corbin, 1994, p. 279; Strübing, 2006). This passage cannot only be read as taking a clear stance in the discussion over emergence versus forging in CGT (Bryant, 2003; Glaser, 2002), but in fact as a clear standpoint on Strauss’ understanding of theories: They are interpretations and thus need not and even cannot be discovered as the researcher is an active and necessary agent in their making. If we need interpretation from a specific stance in order to form theories, then we need data that allow for such interpretation of what we are interested in: the research participants’ points of view of their everyday life realities. Following this thought, reconstructing meaning is one of the main goals in both GT and hermeneutic approaches. In both fields, then, we need ways to carry out interviews, which allow us to gather data that is structured by the interviewed individual and, to the least possible extent, by the interviewer.

Qualitative interviewing techniques with minimal structuring done by the interviewers is also the form that is preferred over topical interviewing in hermeneutic social scientific research, as these methods allow for the interviewed persons to narrate most freely and thus follow their own reasoning and meaning-making in so doing. I myself was trained by leading scholars in the field of hermeneutic social sciences, including Hans-Georg Soeffner, Ulrike Froschauer, Ulrich Oevermann, Anne Honer and Jo Reichertz (Pfadenhauer & Scheibelhofer, 2020). With this academic background, I brought my knowledge to the empirical work I began as a sociologist in GT in 1995 studying the social relations between corrections officers and inmates in a women’s prison (Scheibelhofer, 1996). An important aspect that I incorporated in this study (and the following) was the idea of elicitating narrations structured as widely as possible by the interview partners, as described by Fritz Schütze (1976, 1992).

Narrations and In-Depth Interviews

Interview methods can be based on narrations and storytelling done by the interviewee in order to minimize structuring on the part of the interviewer (Scheibelhofer, 2008). Such interview styles imply interviewing techniques that leave most of the structuring within the interview situation up to the interviewee (cf. For example Gubrium & Holstein, 1997, p. 153). Thus, privileging the voice of participants up to the point that they guide the interview is not a novel idea. Forms of narration-based interviewing techniques include the in-depth interview, the narrative interview and the ethnographic interview: In the course of in-depth interviews, the interviewer seeks to encourage free and open answers by analogy with everyday conversations ² (Johnson, 2002; the open-ended interview as it was coined by Lazarsfeld, 1944; Legard et al., 2003; Lofland & Lofland, 1995). In-depth interviews are meant to capture the respondents’ perceptions and perspectives such that the researcher can reconstruct meanings attributed to experiences and events. In the course of the exchange, the interviewer asks an initial open question and then uses different probes and other techniques to achieve a greater depth of answers. Using these techniques, in-depth interviews and ethnographic interviews often – but not necessarily – elicit narrations. Another interview method that is based on open-ended conversation is the ethnographic interview (Spradley, 1979). Spradley proposed descriptive questions in order to allow for openness for the interviewee’s ‘subjectivity’ that would be necessary in qualitative interviews. In order to deepen understanding and contrasting, Spradley introduces structural questions into his interview framework.

In contrast to the other forms of narration-based interviews, the narrative interview developed by Schütze (1977, 1992, 2003) ³ forges out the interviewee’s orientations by focussing on so-called “extempore narrative renderings” (for more recent versions of the narrative interview, see Eichsteller, 2018; Filipkowski, 2019; Fischer-Rosenthal & Rosenthal, 1997; Svašek & Domecka, 2020). The narrative interview is based on the sociolinguistic structural approach of Labov and Waletzky (1967), according to which the structure of the narration recapitulates the past structures of the original processes. The methodological implication for extempore narrative renderings is that they are therefore seen as “a powerful means of recollection. They tend to express the personal experiences of the informant as that human being who acted and suffered then, i.e. in those former days during which she or he was embroiled in affairs as they are told throughout the course of narration” (Schütze, 1987; 1992, p. 191, p. 191). ⁴ Consequently, the narrative interview is often used to study biographical processes. ⁵

The Interpretive Interview

My examples below of interpretive interviews are based on recent research as well as a research project that focused on the mobility patterns of Austrians who emigrated to New York City after 1965 (Scheibelhofer, 2001; 2008; 2009). At the beginning of this project, I assumed that classic concepts of migration may not have captured these individuals’ international mobility processes. Thus, it seemed appropriate to concentrate on reconstructing the interviewees’ main orientations that had led them to emigrate to and settle down in the United States. Using a biographical research approach to reconstruct migration patterns and orientations from the migrant point of view, the life histories of 26 persons were collected by conducting what in retrospect I would call interpretive interviews ⁶ with a focus on openness during the overall interview. This seemed appropriate because the study focused on the emigrants’ own constructions. The project was mainly based on data collected with interpretive interviews but also included participant observation and document analysis in the research process. All accessible material that could be used to obtain a dense picture of migration experiences from the individuals’ perspectives was included. The main result of this research project was that three key dimensions of social orientation with substantial effects on biographical patterns of mobility can be distinguished: the orientation towards personal relations, the orientation towards occupational matters, and the orientation towards values of self-fulfilment. ⁷

This research project started out with an unstructured, broad interest in Austrian migrants who had settled in New York City. Guided by the main principles shared by all variants of GT – constant comparison, theoretical sampling, empirical work in research cycles – I started with a literature review and carried out my initial interviews with Austrians who lived in that city. After some weeks of coding and memo writing, I decided to gear my research towards the process of migration and its intertwining with individual biographies. Coding the first interviews had at this point already indicated that this group’s migration could only be understood in terms of processes: Reconstructing my interviewees’ biographies, it turned out that at the moment they had arrived in New York, they had not yet planned to settle down permanently. Thus, I decided that it was insufficient to concentrate on the events that had occurred before the interviewees had left Austria. Also, I had to come to an understanding why the interviewees would not perceive themselves as migrants – even after living in the USA for more than a decade.

Based on these considerations, I decided to do problem-centred interviews with a stronger emphasis on the narrative beginning than I had planned at the outset of my research, as the narrative part seemed especially pivotal due to my interest in constructions of migration from the interviewees’ points of view. In doing so, I initiated the interviews with phrases such as:

“Could you please tell me everything that was involved in your coming to New York and how your life has gone on since then? I will listen and make some notes and I will not interrupt you until you have finished. Please take as much time as you feel necessary and tell me all the details you remember that, in your opinion, are connected to your living in New York.”

According to Schütze (1977), the way the initial opening question ⁸ is formulated has an outstanding role to play within the narrative interview setting. This is because it is meant to stipulate the main narration involving utmost limitations to interviewers’ exercise of influence. The narration is then structured by the interviewee who can choose which story to tell and how to tell it so that the interviewer may collect meaningful extempore narrations.

Further Interviewing Techniques: Schütze’s Immanent and Exmanent Questions

Many researchers give preference to guideline interviews because they fear their research participants may not cover their research interests. It is not unusual for researchers to introduce topics one after another following the narrative beginning of an interview. The problem that should be taken into consideration is that this approach may lead to confusion within the interview situation, as the communicative roles are not consistent when switching from a narrative to a semi-structured interview part. In the following section, I will propose some thoughts as to how the narrative introductory part can be followed up by further questions of interest to the researcher without disturbing the initial narrative flux of the interview.

After the migrant interviewees had completed their first narrative accounts in my research project on Austrian migrants, I asked open-ended questions relating to topics that the interviewees had brought up initially but then had not further elaborated on specifically. These so-called “immanent” questions (following Schütze, 1977, p. 35) were based on the few notes I usually take during interviews and I ask them in the order in which the interviewees had brought them up. Also at this point, the interviewees would often go on to narrate. Some of the interviewees who had not given elaborate accounts of their biographies after my initial question provided detailed stories when I got back to the points they had already mentioned. (If interviewees start to repeat themselves when prompted to elaborate on certain aspects, the analysis may gain from such repetition, as one experience is virtually never narrated twice in the very same way. These differences make it possible to double-check the constructed hypothesis in later analysis. Therefore, these passages may indeed help to assess the quality of the interpretations drawn up to this moment of coding).

Only following these immanent questions did I introduce themes prepared beforehand – but only if they seemed suitable to me during the very communicative situation: These external questions ⁹ that I brought up were always formulated as open questions (e.g., regarding the interviewees’ levels of education, job biographies, ties with relatives or relationships with friends). These issues were brought up and combined with other themes that the interview partners may had already mentioned during the first part of the interview in order to avoid a rapid question-and-answer format.

To be more precise with a view on the development of the interpretive interview in contrast to the narrative interview, it needs to be recognized that Schütze introduced external questions exclusively to elicit self-evaluations of past experiences and not as described here for the interpretive interview. Yet also in Schütze’s framework, these evaluation-inducing questions are still brought forward by the researcher and not by the interviewee, although he pointed to the fact that external questions should not be asked if they cannot be logically related to what had been said earlier.

While some researchers also opt for including confrontational probes in an attempt to bring interviewees to speak about the issues they had left out until the very end of the interview (Douglas, 1985, p. 138; Witzel, 2000; Witzel & Reiter, 2012), I continue to refrain from employing such techniques with extremely few exceptions. In terms of methodology, David Riesman pointed out the fact that a challenging interview style can, under certain circumstances, lead to a revelation of a respondent’s “real feelings” (1954, p. 504), such that more accurate interpretations of the empirical data are facilitated. One may thus argue that the utility of confrontational interviewing techniques actually depends on the topic of an interview and research context, and that its justification reflects the observation that bland styles of interviewing may produce bland responses. According to my view, however, such strategies jeopardize the interview setting as a whole: Confronting interview partners after an open interview beginning in most cases proves problematic. This is because such a communication style may lead the interview partners to feel uneasy about having a researcher whom they hardly know point out to them that their narration is not logically consistent – or even ask them to think it over again and perhaps correct what has been said. Besides ethical problems, there is also a methodological pitfall: The risk of channelling the communication event into a setting in which interview partners feel compelled to defend themselves. This is a thorny issue, as, while analyzing an interview, there can be no more than speculation about the social and structural implications of vindications that go beyond their situational meaning in the interview setting itself. Therefore, it seems sensible to me to ask confrontational questions only if the interviewer has an outstandingly well-established rapport with the interviewee so that such probing would not lead to the described interview dead ends. In the course of this exemplary migration project, I refrained from asking confrontational questions altogether.

Although it is not the issue of this paper (for more details see Scheibelhofer, 2020), I would like to include some brief remarks here regarding the analysis of interpretive interviews: Working exclusively with narrative interviews, interpreters would generally opt for analytical procedures that recur on linguistic assumptions while attempting to reconstruct typical biographical process structures (Schütze, 1992; 2003). Alternately, they would choose to focus on both the presentation of the narrated life and the process of telling the story, so that in the end, the relationship between these two elements of the narration can be analyzed (for example, see Fischer-Rosenthal & Rosenthal, 1997). On the other hand, social scientists interpreting semi-structured interviews done with the help of topic guides would usually go for strategies of content analysis. This approach aims to reduce the data without losing the crucial arguments that are relevant to research (Mayring, 2000). Yet compared to open or initial coding, content analysis adheres more to the written letter and relies less on the necessity to interpret, as discussed above.

Within my own research on migrants’ biographies as portrayed here, I attended to extensive open coding work as a foundation for interpretation (in order to clarify, as requested by one of the reviewers: the unit of analysis was thus not a story told or parts thereof but in line with the coding techniques of CGT). According to the well-known basics of GT, fieldwork and analysis are done interchangeably from the beginning on. Starting the research, the focus of my study had not yet been clearly set and I thus began by carrying out participant observation at expatriate meetings and at events I had been invited to through these meetings. The protocols of these meetings were the first basis for open coding. The results of these initial coding sessions helped me focus my research and do further empirical work based on the theoretical considerations brought up during this phase. As an example, coding the first interviews induced me to decide to look for interview partners who had not taken part in the emigrant circles where I had identified my initial interviewees. This redirection of research strategy was based on my hypothesis at the time that these circles reflected a specific segment – albeit in itself closely knit – of Austrians living in New York in the late 1990s.

Applying extensive coding as the main analytical effort with respect to the interviews reflects the methodological assumption that narratives are the communicative form within an interview setting, in which the interviewee is largely able to choose what to tell and how to tell it – quite unlike the semi-structured interview based on a topic guide. Such an understanding of the narrative part of the interview calls for reconstructive interpretation. In retrospect, extensive coding on the narrative parts of the interviews helped approximate the migrants’ main orientations. Open coding was done in the project both with the interview transcripts from the narratives and the interview parts that focused on the further questions I had introduced at the end of the interviews. However, the codes based on the narrations much more directly referred to the interviewee’s relevance systems than those that were reconstructed through the parts of the interview induced by the interviewer’s questions – and thus concerns. In my analytical work on migrants, I encountered problems dealing with such diverse data qualities as I became more and more interested in understanding the decisive moments within the migration biographies. Because of the different data qualities, I decided to give more importance to the codes based on the narration than those constructed upon the answers surrounding my exmanent questions in order to prevent the danger of imposing categories.

Concluding Remarks

Following the thinking of Kathy Charmaz, I emphasize the importance of qualitative interviewing techniques that serve to strive for the interviewed person to be in charge of the conversation. Whilst it may sometimes also be useful to carry out more structured forms of interviews (e.g., semi-structured ones), doing CGT with the help of qualitative interviews may most of the time be based on the narrative interview as described by Fritz Schütze. I refrain at this point from using such terms as “open-ended” or “in-depth” interviews, as they have become generic descriptions no longer in a position to delineate the character of what I here call the interpretive interview.

GT is an interpretive endeavour and it thus needs to build on data that mirrors the respondents’ views, perceptions and meaning-making processes. Only coding such material will help us gather a theoretically rich understanding of the social worlds we study.

In view of the large number of grounded theorists who opt for topical or guideline interviews (Corbin, 2021; Corbin & Morse, 2003), I thus argue to reinforce all aspects of the interview interaction that may enhance narrations. Thus, the first question is not the beginning of such an endeavour, but rather how we initiate the contacts, build trust with future interviewees and chit-chat with them before the actual interviews take place. In this sense, I refer to the works of Fritz Schütze and his narrative interview method. Of course, there is much more to be said about Schütze’s theory of narration, his analysis of the biographic interview, as well as the further developments of his method during the later years (for further details see e.g. Rosenthal, 1995; Schütze, 2007).

In our research project, the techniques of encouraging interviewees to elaborate on their experiences and perceptions were introduced in an attempt to gather data that is self-structured and little-influenced to the best possible extent by my interventions as an interviewer. Within the interview setting, this approach did not seem to prevent the interviewees from continuing after an initial invitation and then being asked immanent and external questions, the latter being based on prior coding and theoretical work. By means of this strategy – building on the narrative interview and extending the external questions to issues other than those the interviewees had already mentioned – we may also address problems linked with Schütze’s rather strict framework: This technique presupposes that all interview partners have rather high communicative competencies (Bates, 2004). If an interviewer meets an interviewee who is not used to narrate, the interview does not have to end at this point, but the researcher can go on and ask questions based on prior coding or reading. This way, data established from such interviewees may less likely be excluded from interpretation.

Implementing extensive coding as the main analytical effort with respect to the interviews reflects the methodological assumption that narratives are the communicative form in an interview setting, in which the interviewee is largely able to choose what to tell and how to tell it – quite unlike the semi-structured interview based on a topic guide. Such an understanding of the narrative part of the interview calls for a far-reaching understanding of the need to interpret. In the migration project, this interpretation was performed with the GT coding strategies, as presented in Anselm Strauss’ later work (Strauss, 1987). In retrospect, such extensive coding on the narrative parts of the interviews served to approximate the migrants’ main orientations. Open coding was done in the project both with the interview transcripts from the narratives and with the final interview parts, in which I introduced questions. However, the codes based on the narrations much more directly refer to the interviewees’ relevance systems than those that are reconstructed through the parts of the interview induced by the interviewer’s questions – and thus concerns. In my analytical work on migrants, I decided to deal with this issue by giving more importance to the codes based on the narration parts of the interviews.