Getting Real About Critical Realist Interviewing: Five Principles to Guide Practice

Abstract

Critical realism refers to a broad project to realize a post-positivist social science. At its best, it responds to the challenges thrown down by social constructionist critiques of positivist science, while also allowing us to make interrogable claims about reality—a priority for many working to expose and eradicate structural oppression. In the social sciences, interviews remain one of the foremost methods through which researchers generate data to inform our understanding of reality. In this article, however, we argue that for critical realism to deliver on its promise as a philosophy of science for critical social scientists, we need theoretically sound guidance on what a critical realist approach to research interviewing might look like. Currently, this guidance is lacking. Through a systematic analysis of prominent qualitative research interviewing textbooks, we found that critical realism is ignored, mischaracterized, and underdeveloped. In response, we offer five principles, rooted in critical realism’s key tenets, that can guide researchers as they design, conduct, and evaluate critical realist interview studies. These principles are: (1) craft interview protocols to generate data that can inform answers to ontological research questions; (2) keep in view the interview as social practice throughout the study; (3) treat interview data as both interactively achieved co-constructions and as verifiable evidence for real phenomena; (4) be guided and informed by an aim to reduce suffering and promote social justice; and (5) demonstrate reflexivity as ongoing self-awareness.

Critical realism is a relatively new paradigm, gaining traction in fields such as education (Currie & Kelly, 2022), health research (Alderson, 2021), and journalism (Wright, 2011). It appeals to many post-positivist researchers who, informed by social constructionism, take seriously that researchers are situated and partial knowers, whose claims are inescapably mediated by available conceptual resources. Equally, however, these same researchers are committed to making ontological claims about what happened, what works, etc. While critical realism initially lacked methodological elaboration, efforts have been made to rectify that gap (e.g., Allana & Clark, 2018; Cabote et al., 2024; Mercier et al., 2023). Yet, despite interviewing being a key data generation practice in many qualitative methods, we have struggled to find broadly applicable guidance on how to conduct critical realist interviewing.

Our purpose here is to help novice researchers and those teaching qualitative methodology understand what might be distinctive in critical realist interviewing. Increasingly, researchers recognize the importance of what Braun and Clarke (2022) call “theoretical knowingness” as a criterion by which we assess the worth of qualitative studies. Theoretical knowingness means “having an understanding of the philosophical and theoretical assumptions that might be embedded in particular approaches or techniques, and striving to use such approaches and techniques with an awareness of those assumptions” (p. 268). In what follows, we: (a) briefly discuss four key tenets of critical realism as a philosophical stance, drawing links between each tenet and the practice of interviewing; (b) show that most of the popular qualitative research interviewing textbooks tend to ignore critical realism and, more generally, ontology; (c) argue that some critical realist methodological discussions rest on implicit theorizations of research interviews that are inconsistent with better accounts of their real nature; and (d) articulate five methodological principles, underpinned by critical realism, for designing and evaluating qualitative research interviewing studies.

Four Key Tenets of Critical Realism for Interviewing

We follow Maxwell (2012) and Archer et al. (2016) in using critical realism to refer to a broad project to “develop a properly post-positivist social science” (Archer et al., 2016, p. 4). They note the wide range of social theorists who have contributed to this project and assert that “there is not one unitary framework, set of beliefs, methodology, or dogma that unites critical realists as a whole” (p. 4). Informed by the work of Archer et al. (2016) to identify the common commitments of diverse scholars working under the broad umbrella of critical realism, in this section we discuss four key tenets that form the foundation of the framework for critical realist interviewing we offer here: ontological realism, epistemological relativism, judgmental rationality, and axiological criticality. As this article speaks primarily to those with some prior interest in critical realism, we do not aim to provide an in-depth discussion of each tenet. Rather, we provide a brief overview of how we understand the tenets and begin to tease out their implications for research interviewing.

Ontological Realism

Critical realists maintain that there is a reality that exists largely independently of our experience of it. Things like educational policies, institutional anti-Indigenous racism, and researcher-participant relationships do exist and can have causal effects on events in the social world, usually regardless of whether we are aware of them or what we think about them (Bhaskar, 2008). Critical realists are interested in the real nature of things–their layered structures, and the inherent properties and latent powers and liabilities that emerge at each layer of a particular structure (Mingers & Standing, 2017). They are also interested in understanding the causal mechanisms, actualized through the complex interaction of things (their properties and latent powers/liabilities) under particular conditions, through which phenomena occur in the open social world (Sayer, 2000).

Given critical realism’s axiological criticality (discussed below), its ontological realism and commitment to (fallibly) grasping it are important, especially when components of a process or emergent phenomenon can be critiqued as impediments to human flourishing. This aspect of critical realism differs from very strong forms of poststructuralism and social constructionism, which might deny or be agnostic about objective or non-discursive reality (Sayer, 2000). Critical realists understand that interviews bear traces of reality—albeit traces with varying degrees of fidelity—that can be used to get an always-mediated handle on reality (Sealey, 2013). Though we may find marked differences in how two people perceive and recollect the same event (see epistemological relativism, below), we think it is also important to acknowledge the extent to which such accounts often corroborate each other and point to a single, albeit emergent and complex, reality.

Epistemological Relativism

Like social constructionism, critical realism accepts that our efforts to apprehend reality will always be shaped and limited by myriad factors that make our experience of reality fundamentally relative (Sayer, 2000). These factors include the kinds of technology we have available to us in a particular epoch, the sociocultural mores that might shape our interests in particular objects of study and how we go about studying them, and our own subjectivities and the particular social vistas these enable. No account of the social world—whether a co-constructed narrative worked up through the course of a research interview or a scientific theory—will ever be a true reflection of ontological reality (Graeber, 2015, p. 27). At the same time, we concur with Al-Amoudi and Wilmott’s (2011) critique that the invocation of critical realism’s avowed epistemological relativism in many empirical studies often “amounts to little more than lip service to a stance that, in practice, is mobilised to brush off [critical realism’s] critics” (p. 39). As we assert below, a critical realist approach to research interviewing must take this tenet seriously through a consistent understanding of what interview data are.

Judgmental Rationality

In contrast to social constructionist or interpretivist approaches to social science (e.g., Charmaz, 2014), critical realism’s epistemological relativism does not lead to judgmental relativism, or the belief that, because our efforts to grasp reality are inherently fallible, we cannot adjudicate between competing accounts. On the contrary, critical realists hold that some accounts are better than others. While the criteria for weighing up the veracity of one account against another are far from decided and the process undoubtedly complex, we agree with Archer et al. (2016) when they assert that it “is possible for social science to refine and improve its knowledge about the real world over time, and to make claims about reality which are relatively justified, while still being historical, contingent, and changing” (p. 7). Regarding data generated through interviews, we might consider a range of factors when attempting to ascertain the veracity of a particular account, including (but not limited to) the internal consistency of the account, its external consistency with accounts produced by others of the same phenomenon, and the likely shaping forces of the researcher-participant relationship and the dynamic interactions between their respective multidimensional social locations.

Axiological Criticality

One of the tenets that differentiates critical realism from some forms of post-positivism is its criticality. What critical realists mean by critical, however, ranges from a reduction of illusion to freedom from unwanted constraints to a stronger standpoint of critiquing injustices and avoidable suffering (Sayer, 2011). This stronger standpoint of anti-oppression has been taken up by critical realists invested in understanding social injustices and intervening to ameliorate them. Wright (2011) states that critical realism “is critical of social structures and practices which produce or perpetuate suffering or other social ills—including those which are caused by the circulation of false or inadequate accounts of reality” (p. 161). This has several implications for the practice of research interviewing. First, critical realists are often (but not always) conducting interviews under the methodological assumption that interviews can generate fallible data that can inform our understanding of a real social ill. Second, during any research interview, participants and researchers alike may draw on problematic macro-level discourses in the construction of an inadequate account of reality (Parr, 2015). Based on the argument that “(other things being equal) it is better to believe what is true than what is false” (Collier, 1994, p. 172), there is a directive implicit in critical realism to challenge inadequate accounts. Third, the research interview itself can be a practice that produces or perpetuates suffering, so critical realists are compelled to develop and exercise a keen reflexive ethical literacy (Hultgren et al., 2016).

Critical Realism: Ignored, Mischaracterized, and Underdeveloped in the Qualitative Research Interviewing Literature

Both of us have engaged extensively with textbooks on qualitative research interviewing (QRI) in recent years, Kelly (a late-career researcher) because she has been teaching a course on this topic and Moore (an early-career researcher) as he prepared and conducted research for his PhD dissertation. We were looking for guidance on the implications of critical realism for the theory and practice of interviewing in social science research. Our impression was that most textbooks to date have ignored critical realism. To ensure that we weren’t missing something, we conducted a systematic search of books published in English between 2010 and 2022 using WorldCat, reportedly the world’s largest bibliographic database, run by the non-profit organization OCLC. We used the following search terms, which yielded 205 results: Interviewing OR Interviews (in Subject) AND Social Sciences ‐‐ Research ‐‐ Methodology OR Education ‐‐ Research ‐‐ Methodology OR Qualitative (in Subject) AND Interview* (in Title). We paid attention to texts with multiple editions, which displayed as separate results, made note of the top 10, then did a separate search on just the author/s and title to obtain the number of libraries worldwide that carried the textbook. By this measure, we present the 10 most widely available (and presumably used) textbooks on qualitative research interviewing in Table 1.

Table 1.

Top 10 Qualitative Research Interviewing Textbooks by Number of Libraries Worldwide.

Author/s	Year	Edition	Short title	No. of libraries worldwide
Seidman	2013	4th	Interviewing as Qualitative Research	482
	2019	5th		149
King & Horrocks/King et al.	2010	1st	Interviews in Qualitative Research	405
	2019	2nd		167
Rubin & Rubin	2012	3rd	Qualitative Interviewing	538
Kvale & Brinkmann	2015	3rd	InterViews	495
Alvesson	2011	1st	Interpreting Interviews	385
Roulston	2010	1st	Reflective Interviewing	334
Brinkmann	2013	1st	Qualitative Interviewing	296
	2022	2nd		9
Galletta	2013	1st	Mastering the Semi-Structured Interview and Beyond	244
Josselson	2013	1st	Interviewing for Qualitative Inquiry	241
Cassell	2015	1st	Conducting Research Interviews for Business and Management Students	208

Of the 10 textbooks, only two contain brief mentions of critical realism. King et al. (2019) include critical realism under the larger umbrella of “limited realism” as part of their general consideration of philosophical assumptions (about epistemology and ontology) and their implications for designing QRI studies. In their schematic treatment, King and colleagues equate critical realism’s ontology with neo-positivism’s, giving no indication of the conceptual resources that critical realism offers about the complex and heterogeneous nature of the social world (Archer et al., 2016). In his introduction, Alvesson (2011, p. 5) quotes leading proponents of critical realism in what Smith and Elger (2014) characterize as a “cavalier” and misleading way (p. 122). The most common pattern in the textbooks we reviewed was the inclusion of brief (Cassell, 2015; Rubin & Rubin, 2012; Seidman, 2019) or more extensive (Alvesson, 2011; Brinkmann, 2022; Kvale & Brinkmann, 2015; Roulston, 2010) mention of competing epistemologies, with little to no mention of ontology. Two books did not discuss epistemology or ontology in relation to QRI in any detail (Galletta, 2013; Josselson, 2013).

Our above analysis suggests a dearth of discussion in the general research interviewing literature on the implications of the ontological realism of critical realism and other forms of post-positivism. Social scientists might then look to avowedly critical realist work for guidance on how to conduct critical realist interviews. After all, following Bhaskar (1998), a critical realist social science holds that “the study of any social practice must start with the agents’ conceptions of it” (Collier, 1994, p. 167), and the interview is a commonly deployed method to generate such data (Price & Martin, 2018). Very often, however, we find that interviews are described in a way that belies the complexity revealed by the many dedicated books on research interviews, including those discussed in the preceding section. For example, in a recent edited collection, Working with Critical Realism: Stories of Methodological Encounters (Maisuria & Banfield, 2023), interviews were sometimes presented matter-of-factly in the simple past tense as methodological procedures used to “collect” data. As we will go on to demonstrate, interviewing is a complex practice, and the idea that interviews can be used to collect pre-existing or unmediated data is at odds with both the inherently social nature of all research interviews and the fact that the data they generate are co-constructed by the interviewer and interviewees in interaction.

Even when critical realists do acknowledge the interviewer/interviewee relationship, we often do not see that reflected in either the presentation or analysis of the data. For example, Lauzier-Jobin and Houle (2021), citing Maxwell (2012) and others, note that in realist interviews “the relationship between the interviewer and the interviewee is more proximal, interactive, and aimed at learning and sense making” (p. 2442). While acknowledging the unseen constraints that may have influenced the eventual write-up of the study, the interview data that form the basis of their description of social support processes that influence mental health recovery are, nevertheless, all presented as block quotes from the participants, and the interactive nature of the data is neither represented nor discussed.

We may more reasonably expect the emerging body of work that does feature extended methodological discussions of critical realist interviewing to provide robust and widely applicable guidelines. We have experienced challenges, however, in applying the ideas generated through such discussions to our own work due to what we feel are erroneous understandings of the methodological implications of critical realism.

Like us, McLachlan and Garcia (2015) were concerned with answering the question, “Is there a concrete way in which ‘to be’ critical realist in interview research?” (p. 199), and we are in complete agreement with their assertion that, because of its nature, “we must consider both the substantive and constructed aspects of the interview experience” (p. 199). Unfortunately, they came to the conclusion that their use of what they describe as critical realist interview practice “proved both difficult and inadequate” (p. 207), and they declare their project to develop “a critical realist approach to qualitative interviewing” (p. 207) a failure, retreating to a “social constructionist” approach. We are not surprised, however, that they failed to see the value of critical realist interviewing because their research questions—concerning the changing sense people made of gendered familial identities in the wake of austerity-related redundancies—were purely hermeneutical and pursued no ontological knowledge. To our minds, any critical realist approach to interviewing must bear in mind that “at the heart of critical realism is realism about ontology” (Archer et al., 2016, p. 4) and thus eventually must offer answers to ontological questions.

A recent article by Brönnimann (2022), published in the Journal of Critical Realism, offers guidance on “how to phrase critical realist interview questions in applied social science research” (p. 1). The author critiques other critical realist interview studies that employ question stems such as “Do you have an opinion [on] …” and “What role do you think individuals like yourself have [in] …” for their allegedly “non-realist ways of phrasing questions” (p. 8). Brönnimann rightly notes that “using these question forms allows the interviewee to construct an idea in the moment of the interview” (p. 8). He is misguided, however, in assuming that such an idea could have no potential relationship to reality. Furthermore, his proposed solution of using questions that are seemingly more objective (such as “How did the mill closure(s) affect you and your family, and the community?” [Brönnimann, 2022, p. 8, quoting Crosby, 2013, p. 267]) fail to offer anything other than a similarly inherently partial account co-constructed through the interaction of an interview. In this respect, it does not matter how one phrases an interview question; critical realism’s epistemological relativism means the answer will always have an asymptotic relationship to the truth.

Lastly, QRI has received considerable attention in the realist arm of program evaluation (Manzano, 2016; Mukumbang et al., 2020; Pawson, 1996; Pawson & Tilley, 1997), and efforts have been made to elaborate a “realist interview technique” (Mukumbang et al., 2020, p. 485). One notable contribution of this body of work is the attention paid to considering and respecting the different kinds of knowledge accrued through experience and brought by each social actor—including the researcher—to the interview. This acknowledgement of differential expertise can become a useful tool to adjudicate between competing accounts. We agree with Smith and Elger (2014), however, when they note that because of realist evaluation’s tight focus on relatively bounded contexts “where subjects are ‘processed’ through programmes” (p. 122), critical realists pursuing very different kinds of research questions in fuzzier or more open contexts are likely to struggle if they were to apply the same techniques. For example, Moore’s study (2022, detailed below) was an exploratory investigation through which he attempted to develop explanatory theory regarding the nature and causes of an until-then unnamed phenomenon driving linguistic distancing behaviors observed in some Japanese plurilinguals; there was no clearly defined context, no program to evaluate, and no clearly defined organizational roles with areas of assumed expertise, etc. In contrast, the five principles we offer below transcend the particulars of any particular social science field to be applicable to any critical realist study involving QRI, regardless of whether one’s research object is a program, a psychosocial process, or a social movement. We submit that they also apply to all kinds of interviews, including data-driven interviews, typical at the outset of a study, in which the researcher’s goal is to grasp a sense of the research object through the empirical accounts of those who experience it, as well as theory-driven interviews in which the researcher’s goal is to test and refine the explanatory power of their theory with participants (Mukumbang et al., 2020), because, following critical realism’s epistemological relativism, theories are fundamentally also fallible accounts of reality.

Our principles also place more emphasis on the inherent and inescapable sociality of QRI and the data it produces than we have found in the treatment of QRI in the realist evaluation literature. For example, Mukumbang et al.’s (2020) conceptualization of the “information management process” (p. 507) in theory-gleaning, -refining, and -consolidating interviews—characteristic of a realist evaluation study—portrays a rather straightforward flow of information between the researcher and the participant; the participant freely shares their ideas regarding the context and components of the program before later—and with the same implied freedom—validating, modifying, or even rejecting the researcher’s theory. We would argue that this conceptualization does not reflect the ontological reality of the researcher/participant relationship and the undetermined yet very real ways this can shape the data generated through any interview (cf. Maxwell, 2012). Indeed, in their discussion of the challenges they faced when implementing their approach to realist interviewing, the authors noticed that when they taught their developing theories to the participants, the participants “would tend simply to agree to what [they] presented” (pp. 508–509), rather than modifying or rejecting the researchers’ ideas. To us, this suggests that on this occasion, the participants were orienting to the researchers as “experts” and deferring to their assumed expertise. Conversely, participants for whom researchers represent out-of-touch elitism or untrustworthy snoops might be inclined to outwardly reject ideas with which they privately agree. As we will go on to show, clues as to how participants and researchers are understanding and orienting towards one another can often be found by paying attention to the here-and-now of the interview interaction. This underscores our point that any critical realist approach to interviewing must be founded on better theorizations of the real structures and properties of interviews and the data they produce.

Five Methodological Principles for Critical Realist Interviewing

Critical realism, as we have used the term thus far, refers to a metatheoretical stance composed of ontological realism, a moderate social constructionism, judgmental rationality, and a critical ethical standpoint. Critical realists have begun to think through and articulate what this philosophy of science means for methodology, but, as Smith and Elger (2014) argue, there is no “uniform view on the implications of their philosophical stance for social research interviewing” (p. 123). Thus, rather than presenting a prescriptive recipe for doing critical realist interviewing, here we offer five broad methodological principles that can inform the design, conduct, and evaluation of a critical realist study that employs interviews. The principles are not intended to be exhaustive, but rather reflect what we think is most distinctive about or essential to a critical realist approach to interviewing. That said, the interviewing theorists we draw on to support the principles are not always critical realists; instead, we are informed by the work of scholars working in various paradigms who have, in our view, offered theoretical insight into the real nature of the research interview that is persuasive to us as critical realists. For each principle, we offer some concrete examples to demonstrate how it can guide actual practice.

Principle 1: Craft Interview Protocols to Generate Data that Can Inform Answers to Ontological Research Questions

As noted above, a distinguishing feature of critical realist social science is its commitment to achieving a better grasp of the reality of the social world. Thus, while critical realists might also ask epistemological questions (i.e., questions about how people make sense of things), a critical realist study should—at least in part—be motivated by one or more ontological questions (Fryer, 2022). By ontological questions, we mean questions that concern the existence and/or nature of real things and phenomena and the causal mechanisms they form through which, when actualized, other things and phenomena emerge. We note here that because “human reasons for actions can be the causes of those actions” (Collier, 2005, p. 332), we might ask ontological questions about semiotic processes when they are understood to form parts of a causal mechanism. Furthermore, interview protocols (and other data generation tools) should be carefully planned so as to produce the data needed to develop a persuasive theoretical account of the nature of things in the social world and the complexity of the generative mechanisms they form under particular conditions.

For example, Moore’s (2022) doctoral study used critical realist grounded theory method to develop explanatory theory regarding the real nature and causes of a process through which some Japanese-English late plurilinguals (Japanese people who grew up monolingually in Japanese and only became users of English after turning 12) come to distance themselves from their first language, Japanese. The process, ultimately termed first language dissociation, was theorized as being enduring and primarily psychosocial in nature; enduring in the sense that it played out across longer spatiotemporal scales than any single interactive event and psychosocial in the sense that the primary causes were best understood as emerging at the strata of the psychological and the social (rather than, say, the somatic or biochemical). After gaining written informed consent, Moore used interviews and other data generation methods such as language use journals and past social media posts to work with each participant to co-construct their linguistic biography, textured with their empirical, partial accounts of pivotal actual events that shaped their affective relationships to English and Japanese, such as experiences of being bullied in school or growing up gay in a Japanese society they perceived to be heteronormative. He then compared and contrasted those diverse accounts to theorize a shared configuration of real causal factors leading to the emergence of first language dissociation among Japanese-English late plurilinguals. These causes included the onset of learning English as an additional language, significant experiences of intersubjective conflict mediated in Japanese, and the effects of contrasting ideological discourses regarding Japanese and English languages and cultures on Japanese-English late plurilinguals’ beliefs about those respective linguacultural complexes.

Principle 2: Keep in View the Interview as Social Practice Throughout the Study

A common product of critical realism’s ontological realism is a commitment to trying to grasp the real nature of the things. That is, critical realists are particular about the structures of things, the properties that such structures grant, and the latent powers/liabilities they may have relative to other things that may be actualized under specific conditions (Mingers & Standing, 2017). Thus, we assert that a critical realist approach to interviewing depends on better theoretical understandings of the complex nature of interviews themselves. One particular aspect of their nature that has important consequences for critical realist interviewers is their inherent social character, or sociality.

Interviews do not sit apart from the complex, open social world—they are social events within that same world. As Fairclough et al. (2004) note, “an interview is a particular form of communication … that both creates a particular kind of social encounter and is itself socially-structured, for example by conventions of propriety, privacy and disclosure, by particular distributions of resources, material and cognitive” (p. 27). This understanding can be contrasted with a neo-positivist conceptualization of the interview as a scientific extraction in which the—in our view, futile—goal is to mimic the logics of a closed experiment in the natural sciences to “collect” apparently unsullied data about the participant’s experience and evaluation of the social world.

When we accept that all interviews are inescapably social encounters in which both the interviewee and the researcher are active participants, it forces us to recognize the ontological reality of the contributions that the researcher makes to the generated data. Indeed, Maxwell (2012) argues that from a critical realist perspective, the shaping influence of the researcher’s personal characteristics is “indisputable” (p. 96) and identifies the researcher/participant relationship as one of the main mechanisms through which this influence is exerted. Both what gets said and how it gets said will be shaped in non-determined ways by the unfolding intersubjectivity of the interview, as different facets of each interlocutor’s subjectivity (e.g., man, researcher, Chilean, Indigenous, disabled, wealthy, etc.), whether real or assumed, become meaningful to the interlocutors as the encounter progresses.

It is for these reasons that it makes little sense to us to talk about data “collection” in a critical realist interview study. Rather, we find the metaphor of data “generation” helps us to keep the intersubjective nature of the research interview, the contributions of the researcher, and the analytical and ethical implications of these facts in mind at all stages of a study.

Once we acknowledge the sociality of research interviews, what does this then mean in practice? In Moore’s (2022) doctoral study, he carefully considered the implications of his role within the social encounter of each interview. This included considering the meanings that aspects of his subjectivity, such as his status as a non-Japanese person or as a man, might achieve during his interviews with Japanese-English late plurilinguals. For example, would his embodied subjectivity as a man make some women participants feel uneasy at the prospect of a one-to-one in-person interview held in a necessarily quiet and private space? This was one reason that participants were offered the choice between in-person and videotelephone modes for their main interviews; videotelephone interviews afforded both privacy as well as relative safety. Because the early evidence he had regarding the possible causes of first language dissociation indicated that it might include the experience of trauma, Moore prepared various protocols and resources to be deployed should one of the participants have displayed signs of psychological distress during an interview.

As Moore processed, analyzed, and shared data to support his eventual theory, he took care to be transparent about his own contributions. The transcript excerpt presented below is adapted from Moore’s unpublished dissertation and is a rendering of the interaction that occurred in a videotelephone interview just after one participant, Musumi (pseudonym), had explained that one of the aspects of Japanese that made her feel uncomfortable using it was her sense that she had to be “more indirect about [her] opinions or feelings,” because to be direct “wouldn’t be culturally appropriate” (see Appendix A for transcription key).

Excerpt 1

Preserved in this excerpt is evidence of the intersubjectivity of this encounter. For example, the researcher and the participant know that both of them are users of English and Japanese, so they are able to switch between those linguistic resources at various points without the need for repair (Lines 6–7; Line 31). Elsewhere, such as when Musumi tells Moore that “Kyoto is very indirect for example” (Line 16), Musumi seems to adopt the role of a cultural informant, perhaps orienting to Moore’s apparent status as a non-Japanese person (if Musumi had thought the interviewer were also Japanese, it is probable that she would have included a “you know” or something similar to acknowledge the likelihood that this common trope was shared knowledge).

Principle 3: Treat Interview Data as Both Interactively Achieved Co-Constructions and as Verifiable Evidence for Real Phenomena

The fact that research interviews are a form of social practice also has consequences in terms of how we understand the very data that are generated through them. A number of scholars have noted the dual “analytic status of interview data” (Rapley, 2004, p. 3) as, on the one hand, a resource, where interviewees speak to their experiences on a given topic, and, on the other hand, as a co-construction, where the interaction between the interviewer and respondent as well as conventions of the interview and its wider context shape what gets said. This distinction, in turn, reflects a difference in how interviewing is commonly conceived in the literature. Talmy (2010) captures this in his contrast between the “interview as research instrument” (underpinned by neo-positivism) versus the “interview as social practice” (underpinned by social constructionism). In the former, interview data are reports, “which reveal truths and facts, and/or the attitudes, beliefs, and interior, mental states of self-disclosing respondents” (p. 132). In the latter, interview data are accounts “of truths, facts, attitudes, beliefs, interior, mental states, etc., coconstructed between interviewer and interviewee” (p. 132).

Critical realists argue for conceiving of interview data as both, reflecting their commitments to epistemological relativism and ontological realism. In fact, Sealey (2013) argues:

The distinctiveness of a realist approach is that interviews, for example, are understood not simply as discursive versions of the social world, intersubjectively “constructed” by researcher and interviewee, but as accounts, however partial, of actual experiences, of situations which exist independently of participants’ perceptions of them. (p. 4)

Keeping this dual nature of interview data in mind is crucial, yet some work that is included in the wider category of realism (e.g., realist evaluation) seems to downplay that interview data are, in part, socially constructed. For example, Manzano (2016) misquotes a key theorist of critical realism when she states, “Maxwell (2012) explains how in realist studies data collected through qualitative means are not considered constructions. Data are instead considered ‘evidence for real phenomena and processes’ (p. 103)” (p. 344). In fact, in the first part of Maxwell’s original sentence, he argues that qualitative (interview) “data are usefully seen … as ‘texts’ to be interpreted” and “as the constructions of participants” as well “as evidence for real phenomena and processes (including mental phenomena and processes) that are not available for direct observation” (2012, p. 103, italics in original).

If critical realists ignore that interview data are partly socially constructed, then it is easy to see how their work gets conflated with neo-positivism and a referential approach to language, conceived as a “neutral medium that reflects or corresponds to objective or subjective reality” (Talmy, 2010, p. 131). Yet it is equally important to appreciate that critical realists use interview data “to make inferences about …[real] phenomena, which can then be tested against additional data” (Maxwell, 2012, p. 103, italics in original). This approach contrasts sharply with strong social constructionism, where it is asserted that “we cannot step out of language” (Braun & Clarke, 2022, p. 180, quoting McNamee), thereby reducing meaning “to intra-discursive relations in abstraction from reference and practice” (Sayer, 2000, p. 32). Critical realists argue, instead, that “meaning is a product of both intra-discursive and referential relations” (Sayer, 2000, p. 33).

What would adherence to Principle 3 look like in practice? Moore (2022) explicitly understood and analyzed his interview data as both co-constructed interactions and as evidence of real experiences and processes. As noted above, Moore attempted to preserve the co-constructed nature of interview data as he made decisions regarding how to transcribe the audio recordings, analyze those transcriptions, and represent the excerpts that were eventually selected as evidence for particular theoretical points. These practices enabled Moore to make transparent how his interaction with a particular respondent produced a certain “trajectory of talk and how specific versions of reality are co-constructed” (Rapley, 2004, p. 7) and for others to “work out the interactional accomplishment of the account” (Silverman, 2017, p. 151). For example, how would a reader interpret Musumi’s revised account of the relative levels of directness she perceived across various social contexts if, instead of the questions that open the excerpt (Lines 1–2), Moore had asked, “But surely, don’t you think there are variations in directness between different regions and cities within Japan?” Combined with the researcher/participant dynamic in which researchers often enjoy a relatively powerful position, we think this much more leading phrasing would cause many to doubt the sincerity of Musumi’s revised account. Critical realists need to show more of the co-constructedness of their interview data and, when necessary, discuss how this co-constructedness has shaped the resultant data and the conclusions they have drawn from it.

As discussed above, although critical realists must acknowledge the co-constructedness of interview data, that co-constructedness does not negate the possibility of also approaching interview data as interrogable evidence about social reality (Smith & Elger, 2014). As Moore (2022) iteratively generated and analyzed his data, two competing accounts regarding the nature of the Japanese language became apparent. On one hand, multiple participants asserted that the “indirect” nature of Japanese was one of the reasons they wished to avoid it. On the other hand, as a critical applied linguist and an English-Japanese plurilingual himself, Moore felt there was very little evidence to support the notion that one language was more indirect than the other and that the idea of linguistic indirectness itself was theoretically shaky. Given that, by the time he interviewed Musumi, Moore had already identified such linguaculture beliefs as part of the generative mechanism of first language dissociation, he was at that point interested in interrogating their veracity as theoretical accounts of the nature of the English and Japanese languages. Like Musumi (Lines 4–8, 10–11), many participants traced this less satisfactory account back to things others had told them about the supposed nature of English and Japanese. As many of those others happened to be English teachers, Moore was able to identify language classrooms as the very real sites in which such unsound linguaculture ideologies are inculcated.

Principle 4: Be Guided and Informed by an Aim to Reduce Suffering and Promote Social Justice

Many critical realists, including key theorist Roy Bhaskar, have taken inspiration from Marx’s assertion that the point of social inquiry is to change the world. “Bhaskar argued that if we can identify the underlying mechanisms that are producing injustice or suffering, then we can show under what conditions it could be otherwise” (Clegg, 2016, p. 507). While we agree that critical realist studies do not necessarily have to be driven by critical research questions (Moore’s research questions were not explicitly critical), such a critical standpoint puts a concern for values (e.g., anti-oppression, human flourishing) front and center, informing one’s entire research program and approach to ethics (how we treat people). Like research in general, interviewing studies never consist of a neutral set of practices; understanding this requires an acknowledgment of positionality and power relations.

Critical realism’s endorsement of values as indispensable to social inquiry distinguishes it from both neo-positivism and certain forms of realism that aim for value neutrality-as-objectivity (e.g., Hammersley, 2009; Alderson’s [2021] critique of realist evaluation). As Sayer (2000) notes, value neutrality is not the same as objectivity in an epistemological sense, defined, instead, as the practical adequacy of knowledge claims (i.e., not merely a matter of opinion and based on the preponderance of evidence and the best arguments available). All things equal, researchers can work from a set of values and produce knowledge that is more practically adequate than those claiming to be value neutral, provided, of course, that they avoid dogmatism.

At the same time, strong social constructionists sometimes fail to challenge neo-positivists’ conflation of objectivity-as-value-neutrality with objectivity-as-the-search-for-practically-adequate-knowledge and thus reject the notion of objectivity altogether (Sayer, 2000). They ignore a third (ontological) meaning of objectivity, that is, that “objects can have qualities which exist independently of our consciousness of them” (and about which we can be mistaken); this tenet of realism elicits an acknowledgement of the “extra-discursive nature of body and world” (Sayer, 2000, pp. 60–61).

In contrast, strong social constructionist work can downplay “complex embodiment” (Siebers, 2008) as well as ignore the shaping impacts of unequally distributed, scarce material resources. In a critique of poststructuralist Rom Harré’s discursive psychology, for example, Archer (2000) insists on the importance of an extra-discursive, material dimension to social life:

… generically, resource distributions hinder the aspirations of some and help in the attainment of others, whether those involved discuss this state of affairs or not. Even if they do, their causal influence does not hang upon correct conversational diagnosis: our “life chances” do not hinge upon our knowledge of them, for discrimination or privilege operate independently of their discursive detection. (p. 90)

What would adherence to Principle 4 look like in practice? Researchers should demonstrate sensitivity to power and status differences at all stages of the study. While many qualitative interviewing studies will enact practices to minimize harm to participants, the tenets of critical realism suggest that this needs to go further, taking account of the circulation of interview data beyond the immediate context of their production (Briggs, 2007). To return to our example in Excerpt 1, Moore (2022) traced the notion of Japanese as an indirect language—as espoused by Musumi and other participants—to its roots in auto-Orientalism (Befu, 2001), a form of internalized racism. Given that Moore thought it was a potentially harmful ideology, he felt compelled to invite participants like Musumi to question the idea that English is a more direct language than Japanese. As noted above, Moore also eventually traced this ideologically distorted belief largely to English language classes underpinned by problematic theories of language. This created a goal for further possible interventions: challenging such ideological distortions through better teacher education for language teachers.

Notably, Moore did not always feel it was ethical to query participants’ potentially harmful beliefs, determining, for example in the case of Ikumi (another participant), that her belief in Japanese exceptionalism “formed part of a conceptual system connected to [her] wellbeing” (Moore, 2022, p. 258). These ethical judgment calls require ongoing self-monitoring, a topic we take up in the next section.

Principle 5: Demonstrate Reflexivity as Ongoing Self-Awareness

The call for reflexivity is widespread among qualitative researchers (e.g., Alvesson & Sköldberg, 2009; Davies, 2008; Finlay, 2012; Roulston, 2010). As we argued earlier, researchers should demonstrate reflexivity from the very beginning of their study in terms of “theoretical knowingness” (Braun & Clarke, 2022) about the philosophical underpinning of their approach and associated conception of the interview. A critical realist approach to qualitative interviewing requires an awareness of our positionality and its influences on how we perceive and think in order to help guard against distortion and misrepresenting the objects of our study. Since knowledge is inevitably situated and thus partial and provisional, reflexivity is essential to achieving practically adequate accounts (Sayer, 2000). As Sweet (2020) observes, “Reflexivity asks us to approach our work with epistemological unease because we are always at risk of reproducing categories that reify power” (p. 924) or reproduce stereotypes.

By reflexivity, we mean ongoing self-awareness. We draw from Archer’s critical realist theorization of reflexivity as mediating between sociocultural conditioning and human agency, between a person’s positionality and their choices and actions, albeit not under conditions of their choosing. In short, reflexivity refers to our continual inner dialogue; it is “the mental commentary which precedes, attends, and reflects upon our actions” (Archer, 2000, p. 319; for further discussion, see Kelly & Currie, 2021).

To underscore the central tenets of critical realism already discussed, in this section we highlight some key dimensions of reflexivity that can serve as prompts, namely: retroductive, epistemological, evidentiary, political, and ethical. These do not exhaust the possibilities for ongoing self-monitoring but can serve as a mental checklist, and they flow directly from the prior four tenets we have identified, which, in turn, have been inspired by a critical realist paradigm.

Retroductive reflexivity: critical realism’s concern for causal explanation, as we have seen, means asking ontological questions. Retroduction is a method of reasoning that involves asking: What must the world be like for this to occur? (Alderson, 2021, p. 55, citing Bhaskar). Thus, researchers ask themselves questions like: “What does the existence of this object/practice presuppose? What are its preconditions …? What is it about this object which enables it to do certain things: e.g. what is it about professional associations that makes them able to bid up the salaries of their members?” (Sayer, 2000, p. 16, italics in original). Critical realists seek answers to these sorts of questions as they analyze their interview data but also in their initial review of the relevant literature. As Sims-Schouten and Riley (2019) advise, early on “we turn to published research, policy documents, and sometimes gray literature with the intention of identifying some of the entities that may combine to form causal factors” (p. 1019). Indeed, Willis (2019) recommends adding retroductive questions to one’s interview schedule to enhance the inquiry into contextual constraints and enablements. Strategies to elicit retroductive thinking include “counterfactual thinking, social and thought experiments, studies of pathological cases, studying extreme cases and comparative case studies” (p. 455, citing Danermark et al., 2002). “The emphasis is on retroductive analysis with the research participants, instead of retroduction being something that is done for them by being applied to the data retrospectively by the researcher” (p. 454, italics in original). Willis argues that this collaborative approach can enhance the reflexivity of both interviewer and interviewee.

Epistemological reflexivity: critical realism’s epistemological relativism means researchers need to scrutinize their disciplinary and other conceptual resources as well as the common-sense assumptions underlying central categories used by participants (Al-Amoudi & Willmott, 2011, p. 35). In education, such central categories include student, teacher, curriculum, pedagogy, and age appropriateness. In addition, we should interrogate how we are implicated in the study we have designed (Davies, 2008; Maxwell, 2012). How does my multifaceted identity influence my evolving understanding, and am I open to surprise and how my initial assumptions might be wrong? How does my emotional investment in the topic shape how I interact with various participants? How might my affective reactions and those of participants serve as an important source of potential insight?

Evidentiary reflexivity: Regardless of whether the term validity is used, care must be taken to generate data that allow researchers to “develop and test [their] emerging understandings of the phenomena [they] are studying” (Maxwell, 2012, p. 103). Researchers need to keep in mind that they might be wrong or only have partial understanding. Can the evidence presented be interpreted in another way? How? What difference might an alternative interpretation make?

Ethical reflexivity: Ethical reflexivity demands that researchers think through the consequences of the many choices they make while designing, conducting, and evaluating their projects (Finlay, 2012). Why am I doing the research? For whose benefit? Do I see participants as a source of data, as experts on their lives, or as research collaborators? How do I compare in status with my participants, and how do I negotiate interviewer-interviewee power differentials? Under what conditions might participant confidentiality not be the most ethical option (Pickering & Kara, 2017)? How might I respond if an interviewee makes racist, sexist, homophobic, or other discriminatory comments? Blakely and Moles (2017) have highlighted the “reflexive identity work” that research interviews require, noting: “The interview is disruptive to the routine meaning making and situated action of daily life’ (p. 169). How might participating, then, in an interview potentially disrupt the longstanding biographical work of the interviewees, and with what consequences? Have I considered how my transcription conventions and data excerpts from the resulting interview transcripts might represent interviewees in the minds of various intended audiences (Oliver et al., 2005)? In interpreting my data, what tensions arise as I (perhaps implicitly) speak for or about participants?

Political reflexivity: In light of the inescapably value-laden nature of research, researchers must attend to the dynamics of power involved in any interview study, such as “the hierarchical, often charged relations between researcher and informants, the politics of interpretation and representation, and the social consequences of making claims on the basis of science” (DeVault & Gross, 2012, p. 206). What are my values as compared to the values of my interviewees, and how are these shaping the interview? What kind of politics, broadly understood, is promoted by my study? What kind of action or inaction does it encourage? Does the written account support or marginalize a particular group? With what potential consequences?

Conclusion

The five principles we have offered here are by no means the only ones that might guide those planning, conducting, or evaluating a critical realist study involving interviews. Indeed, a single article could never cover the consequences of critical realism for the innumerable facets of the complex practice of research interviewing. Among these facets, however, there are methodological priorities to which we think critical realists should attend. Through our discussion of critical realism’s four key tenets and what we consider to be the five most important principles that follow when those tenets are applied to the practice of interviewing, we have attempted to clarify for qualitative researchers what is most distinctive about or essential to critical realist interviewing.

Critical realism is a philosophy of science that allows us to make interrogable claims about reality, and interviews can generate data to inform such claims. To produce credible claims, however, we first need to get a handle on the truth of interviews themselves.