Social Encounters and the Worlds Beyond: Putting Situationalism to Work for Qualitative Interviews

Abstract

In Goffman's terms, qualitative interviews are social encounters with their own realities. Hence, the ‘situational critique’ holds that interviews cannot produce knowledge about the world beyond these encounters, and that other methods, ethnography in particular, render lived life more accurately. The situational critique cannot be dismissed; yet interviewing remains an indispensable sociological tool. This paper demonstrates the value that situationalism holds for interviewing. We examine seemingly contradictory findings from interview studies of middle-class identity (cultural hierarchies and/or egalitarianism?). We then render these contradictions comprehensible by interpreting data excerpts through ‘methodological situationalism’: Goffman's theories of interaction order, ritual, and frontstage/backstage. In ‘situationalist interviewing,’ we suggest that sociologists be attentive to the ‘imagined audiences’ and ‘imagined communities’. These are key to identifying the situations, interaction orders, and cultural repertoires that lie beyond the interview encounter, but to which it refers. In sum, we argue for greater situational awareness among sociologists who must rely on interviews. We also discuss techniques and measures that can facilitate situational awareness. A promise of situational interviewing is that it helps us make sense of contradictions, ambiguities, and disagreements within and between interviews.

1. Introduction

Interactionist insights about social situations are sometimes presented as insurmountable problems for qualitative interviewing (Atkinson 2014; Dingwall 1997; Jerolmack and Khan 2014a; Trouille and Tavory 2019). In Goffman's terms ([1967]1982), an interview is a social encounter with its own reality. Interviewers and interviewees need to agree upon the definition of the situation and establish a working consensus ([1959]1987:21). As the meaning produced in an interview is thus inseparable from the interview itself, the situational critique is that interviews fail to generate knowledge about the world beyond this encounter. This relates also to the ‘attitude-behavior-correspondence’ (ABC) problem (cf. Jerolmack and Khan 2014a); i.e., that people may not do what they say they do. However, ethnography, typically held up as solution, does not necessarily escape the situational critique either, and is often not feasible, due to lack of access or prime scenes of interaction, or even preferable, due to lack of breadth. ‘For good reason,’ Lamont and Swidler state, ‘interviewing is sociology's standard workhorse method’ (2014:158).

This paper argues that situationalism holds undetected value for interviewing. We show this by examining interview data from studies of middle-class culture, where the authors reported seemingly contradictory findings. Such contradictions might raise doubts about interviewing. Relying on ‘methodological situationalism’ (Knorr-Cetina [1981]2015; Goffman [1967]1982), we make sense of these contradictions. Drawing attention to the productive potential of situationalism, we suggest methodologies to aid sound situational interpretations of interview data. This endeavor is both an opportunity and a necessity, as debates about interviews correspond to certain methodological pitfalls. First is a tendency to discard the interview entirely, which is untenable since interviewing is an indispensable sociological tool. Second is to tacitly neglect problems like ABC and how accounts in an interview can provide data on anything beyond it. This is also unfortunate; the situational critique cannot be dismissed. These pitfalls in turn prevent insights: With the first pitfall, interactionists, who have the most relevant theoretical expertise have not brought that expertise to the table. With the second, interviewers who could have made good use of this expertise have not invited it.

In methodological debates on the situational critique and the value of interviews (e.g. Cerulo 2014; Jerolmack and Khan 2014a; Lamont and Swidler 2014; Maynard 2014; Vaisey 2014), there has been little elaboration on the implications for interviews (apart from rejecting their value), that we treat them as social situations with properties of their own; and whether and how interactionism/situationalism can help us make sense of interview data. In short, there has been little critical reflection about whether and how we can put situationalism to work for qualitative interviews.

A situational approach especially helps us make sense of contradictions within and across interviews. As we will demonstrate, such contradictions often correspond to contrasting situations in the lives of our interviewees, situations that engage different audiences, communities, and repertoires. These are interpretations not easily arrived at without explicitly building on situationalism/interactionism. While decidedly not solving all problems posed by the situational critique, and while acknowledging the challenges, thinking systematically in situational terms helps alleviate problems, and improves our understanding of the ontology of interviews.

Roadmap

Section Two recaps the ethnography vs. interview debate. In Section Three, we highlight interviews as encounters by comparing similar interview studies that reached conflicting conclusions about the boundary-work of the Norwegian middle class (Jarness 2013; 2017; Jarness and Friedman 2017; Sakslind and Skarpenes 2014; Skarpenes 2007; Skarpenes and Sakslind 2010). Whereas Ove Skarpenes concludes that egalitarianism curbs cultural distinctions, Vegard Jarness presents egalitarianism as more of a ‘cloak’ for elitism.

We treat Skarpenes’ and Jarness’ contrasting conclusions as a puzzle (Abbott 2004; Tavory and Timmermans 2014; Vassenden 2018), which a situationalist approach renders comprehensible. Disagreements do not mean that interviewing is a flawed method. Rather, their contrasting conclusions stem from their interviews (as we interpret it) being different situations, which reflected contrasting situations in the interviewees’ lives outside of the interview. Participants hence evoked different cultural repertoires. We substantiate this interpretation through situationalist analysis of interview data from our own studies of class and culture (Vassenden and Jonvik 2019), and from findings from the wider literature on class and culture in Norway.

Section Four links the empirical examination to theory that treats social situations as the core units of analysis. We engage Erving Goffman's sociology ([1959]1987); [1967]1982; 1983), especially his notions of interaction order, rituals, impression management, and ‘backstage’/’frontstage.’ We connect this to a recent argument from Tavory (2020), that as refracted contexts, interviews offer a ‘window’ through which we can cautiously infer to ‘Landscapes of Meaning’ (Reed 2011); i.e., to imaginations, meaning-making and repertoires. This is a promising path for interviewing, which we combine with distinctive situationalism. This lays a foundation for developing methodologies that can treat interviews as encounters as a promise to sociology.

The task for qualitative interviewers, we contend, is to reflect critically upon how the interviewer and the interviewee co-define the situation. In short, we need to improve our situational awareness. Systematically considering how a specific definition of the interview situation makes other situations in the lives and imaginations of our interviewees relevant, is key to understanding the repertoires they mobilize. Here, we propose that researchers be acutely attentive to the ‘imagined audiences’/‘imagined communities’ ¹ at play across interviews, and at different moments within interviews. Section Five discusses the practical implications of situational interviewing, with hands-on advice on how to achieve situational awareness.

2. Interviews vs. Ethnography?

Textbook Stuff, yet Unresolved

The situationalist critique of interviews is found in methodology debates in journals and in standard textbooks. A notable debate was initiated by the article ‘Talk Is Cheap: Ethnography and the Attitudinal Fallacy’, in which Jerolmack and Khan (2014a) (J-K) proclaimed ethnography as superior to interviews in its ability to capture actual behavior, and not simply people's attitudes and accounts of their behavior. From the basic sociological insight that ‘(…) meaning and action are collectively negotiated and context-dependent’, J-K argued that interviews, being accounts and not behavior, reveal little about how people actually live (2014a:178;181). J-K's article appeared in a special section of Sociological Methods & Research, which included comments and criticisms from DiMaggio, Cerulo, Maynard, and Vaisey. A central theme in these debates, explicated by J-K, and present in the comments (especially Maynard's), is the interactionist notions of the social situation and the interaction order, with inspiration from Blumer (1969), Goffman (e.g., 1983) and Garfinkel (1967). In short, ‘signification and action (…) “belong” to situations as much as individuals’ (J-K 2014a:181). J-K argued that this premise means that interviews, being situations in themselves and thus local orders, fail to generate knowledge about people's meanings and actions outside of the interview (e.g., p. 181).

The saying vs. doing problem is not new. As J-K show, it has been acknowledged since at least the 1930s. The most famous example is LaPiere traveling around the US with a Chinese couple. Although the group was accepted at 250 of 251 hotels, more than 90% of those hotel proprietors claimed in follow-up questionnaires that they would not admit Chinese guests (1934:234; J-K 2014b:182).

J-K's also claim that sociologists’ attention to the ABC problem was greater in the 1950s than today (2014b:183), which is more dubious. The ABC problem is, after all, why field-experiments on ethnic discrimination in the housing and labor markets are considered more robust and trustworthy than interviews or surveys (Pager and Quillian 2005; Quillian and Midtbøen 2021). The latter example also illustrates that ‘social desirability’ in ethnic discrimination has been almost reversed since the 1930s; from articulated racism being comme il faut with LaPiere's hotel proprietors, even imperative to attract white guests, to widespread norms of tolerance in the 21^st century (cf. Tavory 2020).

As inspiring as such methodological debates can be, they are mired in an unproductive opposition of interviews vs. ethnography. And surely, experiments are not always suitable nor provide the needed data on people's experiences and perspectives. Despite the relevance of J-K's critique of sociologists’ negligence of the ABC-problem and the situational critique, and the fruitfulness of the subsequent articles (cf. DiMaggio 2014, etc.), participants in these debates fail to ask or discuss how we can turn these insights into improvements of the ways we conduct and analyze interviews. This is where our paper makes its contribution.

If a manuscript for a textbook on qualitative methods failed to explain the difference between saying and doing, no sociology publishing house should accept it. Nor should it find its way onto the syllabus for any postgraduate sociology course on qualitative methods. The doing vs. saying problem is textbook stuff, a claim that is confirmed by a review of textbooks in qualitative methods.

Textbooks, however, typically evade addressing the problem, and tend to be caught in discussions along an ‘objectivist-constructivist continuum’ (cf. Miller and Glassner 2016:52). One example is Silverman's three views on the interview: positivist, emotionalist, constructionist (2001:87). According to Silverman, positivists conceive of interview data as ‘facts about behavior and attitudes’; emotionalists as ‘authentic experiences’; and constructionists as ‘mutually constructed’. Positivism and emotionalism are seductive, but illusory, he claims. In constructionism, the interview is nothing more than ‘situational speech’ co-produced by interviewer and interviewee. Atkinson (2014) aligns with what Hammersley (2017) calls ‘the radical critique of the interview’ (cf. Dingwall 1997). Again, the fact that interviews are social encounters is cast only as a problem of generating knowledge about the world beyond (cf. Dingwall 1997), not as an issue that we must address head-on, but which we can build upon to improve interviewing methods (see however Holstein and Gubrium 2016; Miller and Glassner 2016).

Other textbooks make similar distinctions. Kvale and Brinkmann (2009:47–48) present two metaphors: the interviewer as miner or as traveler. The miner digs for the interviewee's ‘true meaning’, which exists before and outside of the interview. The traveler, in contrast, undertakes a journey of co-production of meaning, alongside the interviewee. Kvale and Brinkman considers ‘true meaning’ as illusory; a modernist perception of stable, incontestable stocks of meaning. Holstein and Gubrium (1995) present yet another version of the dichotomy: the interviewee as vessel of knowledge; and the interviewer and the interviewee as co-constitutive forces in meaning-making. Holstein and Gubrium, too, reject the first, and like Silverman and Kvale and Brinkmann take a constructionist stance, treating interviews as ‘products of social interaction’.

These are necessary corrections to naïve positivism. Nevertheless, researchers who want to say something about the world outside of the interview—and if not, why do we keep interviewing people?—may find themselves asking ‘then what?’ If we agree upon ‘the interactional basis of interviewing’ (Holstein and Gubrium 2016:68–70), but still believe we ought to make inferences about the world outside, and that such inference is possible, then what do we need to understand about interviews, and how do we proceed? Can we put situationalism to work for interviews? If so, how?

Many textbooks present problems like the ABC but offer no tenable solutions. In contrast, empirical publications based on interviewing, as J-K correctly note, as well as practical pieces on interviewing (e.g., Hermanowicz 2002), too often fail to engage with the problems. Articles like J-K's are thus timely reminders about unresolved methodological issues in sociology. However, the implied solution (Atkinson 2014; J-K, 2014b), that we rely solely on ethnography, is untenable.

Ethnography not a Catch-all Solution

Unreflective use of interview data, and subsequent overreliance on interviewing is arguably problematic. However, there are clear and strong limitations to ethnography too. This is also textbook knowledge, but it bears repeating (cf. J-K).

Ethnography's most important limitation is feasibility. Often, the alternative to an interview study is not an ethnographic study, but no study at all. The issue of access is significant; often, communities can be researched only through interviews; not because ethnography would not be well-suited, but because access is impossible. Reasons include secrecy/confidentiality, notably in business, the military, politics, or public administration. Even with access secured, research must be realistic. The empirical examples in this paper pertain to social class and culture. To understand how classed culture matters, one may need to investigate the multitude of people's daily interactions with different others. As ethnography more than interviews relates to specific sites, interviews may be necessary if one needs data on interactions in a multitude of sites, not least when one studies group relations. Regarding J-K's ‘Talk is Cheap’ article, Khan's ethnography of the elite school of St. Paul's (2011) was possible because of the overlap between the institutional/interactional boundaries and the students’ lives. The school was where their lives took place. While confined spaces like boarding schools, prisons and asylums are convenient for ethnography, we cannot restrict sociological investigations to total institutions or to social phenomena with one prime site of interaction (cf. below). Much of modern life takes place in less delimited contexts; like for the middle-class people in the studies we review. Indeed, if Khan was to conduct some follow-up of his study, and investigate the participants as adults, could he actually do ethnography? What would be the institutional context? Perhaps he ought to interview them instead.

There is also the obvious issue of time of events. If you are interested in life-courses spanning decades, or events/contexts in the not-too-distant past, ethnography is impossible. Perhaps you need to research some extremist groups from an earlier era that have since disbanded, like the self-professed Satanist church-burners from the Norwegian Black Metal music scene in the early 1990s. Ethnography is simply not an option. However, if you have access to key actors, you will interview them.

Ethnography or interviewing can also be a choice between depth and breadth in data. Although a key advantage of qualitative methods is the possibility of pursuing issues in depth, breadth cannot always be traded away. Interviewing also facilitates systematic comparative designs. Trouille and Tavory make good arguments for ethnographers to shadow participants across situations, such as ‘shoring up intersituational claims’ (2019:537). Their reasoning resembles J-K's critique of interviewing; meanings change across situations. Referring to J-K, Trouille and Tavory state that ‘(t)he problem is that interviews or survey situations are specific, experientially distant, and well bounded. To generalize from one situation (and a relatively odd and reflexive one, at that) to other situations and arenas requires a leap of faith that is often unwarranted’ (2019:537). Interestingly, they turn the same criticism back on ethnographers, whom Trouille and Tavory claim, too often rely on one main site for their ethnography, and ‘(…) quite often extrapolate from one set of situations to a very different set of contexts’ (2019:538). In other words, the situationalist critique also concerns ethnography.

While Trouille and Tavory make a thoughtful argument similar to ours, their suggestion to follow ‘[…] interlocutors across multiple settings and situations’ (2019:536) certainly cannot answer all sociological research questions, not least because of feasibility, access, and breadth. Their argument is illustrated with an example from Trouille's study of Latino working-class men who socialize by playing soccer in Los Angeles parks. This was a five-year ethnography, where shadowing in new situations began after two years in the main site. To say the least, such prolonged fieldwork is not always an option.

Khan might need to confine himself to interviews should he conduct a follow-up ethnography of St. Paul (2011). He might consider shadowing a couple of his interlocutors across work, administrative boards, politics, civil associations, family-life, or friendship networks, but breadth would be sacrificed for depth. His participants would long since have scattered across cities, neighborhoods, work-places, associations, families and social networks.

However, weighing ethnography and interviewing against each other would do a disservice to both. For interviewing, it is more fruitful to acknowledge and address the situational critique and then discuss whether and how situationalist interpretations of interviews can aid sociological research.

Tavory's contribution (2020) is exemplary in this regard. He proposes a typology of inference from interviews. This follows from his claim that sociologists need to acknowledge different types of inference. Using the metaphor of a window between interviews and contexts elsewhere, Tavory identifies three types of interview contexts: open, closed, and refracted. Open contexts can be cautiously inferred from, if we consider biases of desirability, agency, and hindsight (pp. 3–4). Typical open contexts are processual accounts (where, what, who, etc.), and symbolic situations known to be similar to real life, like voting. Closed contexts, which do not point beyond the interview, are rarer. Tavory is unclear about the difference between contexts that are de facto closed off from life outside, and those that researchers treat as closed because of research interest (as ethnomethodologists do), but the category is useful. The open and the closed are fairly straightforward; we can relatively easy decide how they do and do not point beyond the interviews, respectively. In contrast, the most intriguing category in Tavory's typology is also the most difficult: Refracted contexts is an optical metaphor; a window refracts light as it streams into the interview situation. Refracted contexts are ‘promissory aspects of talk’, and (from Reed 2011) ‘landscapes of meaning’. Tavory states that researchers should approach refracted contexts with the assumption that ‘(…) the representations and narratives that we get at through interviews are refracted indications of how interviewees represent their world in other contexts’ (p. 8).

Tavory's typology is especially useful for researchers who are about to start analyzing interview transcripts. It also helps us situate our contribution; among the contexts, the empirical examples that we discuss relate to the refracted contexts and ‘landscapes of meaning’ (i.e., cultural repertoires; Swidler, 1986; Lamont and Thévenot, 2000).

To make good use of refracted contexts, a reflexive approach to how the interviewer and the interviewee co-define the situation is crucial. Rather than, like the methods textbooks, being satisfied with the acknowledgment that the interview situation is constructed, proper interviewing requires meticulous analysis (‘close reading’) of the interaction order. Moreover, we should attend to how these situations reflect situations outside of the interview. This is particularly relevant for understanding contradictory findings (which, with situationalism can be much less contradictory). This, we propose, involves careful reflection about ‘imagined audiences’ and ‘imagined communities’. To show how and why these issues matter, we turn to our empirical puzzle.

3. Puzzling Findings, Faulty Interviews?

Contrasting Findings on the Middle Class and Cultural Boundaries

The last 15 years have seen a proliferation of qualitative papers on culture and class in Norwegian sociology, inspired by Bourdieu ([1979]2010), and/or Lamont (1992). These papers typically revolve around egalitarianism and whether and how cultural hierarchies define middle-class lives in egalitarian societies. Norway is a small country of 5.4 million, known for its redistributive social-democratic welfare state, marked by universalism and de-commodification (Esping-Andersen 1990). Economic inequality is decidedly rising, but is lower than in most European countries, not to mention the US—at least in terms of income inequality.

We can conceive of egalitarian disapproval of hierarchies as national legacy (Sakslind and Skarpenes 2014; Skarpenes and Sakslind 2010 on Norway) or as a broader phenomenon to do with temporal shifts (Wouters 2007). Regardless, egalitarian norms imply that hierarchization is hard to justify.

Norwegian interview studies include Jarness (2013), Jarness and Friedman (2017); Jarness, Pedersen and Flemmen (2019), Ljunggren (2017); Skarpenes (2007), Skarpenes and Sakslind (2010; 2014), and Vassenden and Jonvik (2019). Explicitly or implicitly, all dwell on whether egalitarianism restrains cultural boundaries (Lamont 1992) or distinctions (Bourdieu [1979]2010). The authors seem to agree that egalitarianism is a ‘common good’ (Boltanski and Thévenot 1999), a shared morality vis-à-vis which actors, especially in public life, justify their opinions, and practices. This shared backdrop notwithstanding, significant discrepancies between their findings seem at first glance puzzling. With J-K's line of argument, these discrepancies could indicate that interviewing is a faulty method.

Skarpenes and colleagues conclude that Norwegian egalitarianism results in few cultural distinctions or little disdain from the middle class towards other social groups (Skarpenes 2007; Skarpenes and Sakslind 2010). This is disputed, though. Several sociologists argued that Skarpenes’ findings only reflected his interviewees’ masking of cultural hierarchies (Skogen et al. 2008; cf. Andersen and Mangset 2012). Other researchers, Jarness among them, add nuance; while recognizing the importance of egalitarianism, for class identities, they claim that egalitarianism is coupled with class condescension (Jarness and Friedman 2017), or class clustering that happens hidden from view (Vassenden and Jonvik 2019).

Jarness's study is on cultural identifications in Stavanger. With a research design inspired by Bourdieu and Lamont, he interviewed 46 people from the cultural elite, economic elite, intermediate, and working class. Jarness’ publications from this study are on symbolic boundaries within the middle class (2017), modes of cultural consumption (2015), and ‘strategic’ downplaying of cultural boundaries (Jarness and Friedman 2017). How middle-class Norwegians relate to ‘the honorable cultural ideal of egalitarianism’ (Jarness and Friedman 2017:21), informs much of his writing. Nonetheless, Jarness reports elitism and clear downwards boundaries, especially from within the ‘cultural fraction’ of the middle class. This excerpt from Jarness’ interview with a male journalist in his early 40s, is illustrative (2013: 212; shorter version in 2018: 513).

(V: You mentioned elitism earlier?) I’m elitist. (V: Yeah?) And a snob. […] I’m marked by having spent considerable time at the university and frequenting what the Progress Party [V-M: right-wing populist] would classify as the cultural elite. […] I demand that the people I’m surrounded by are conscious about themselves, their surroundings and society at large. I expect them to be able to discuss politics, and Bob Dylan's discography, for that matter. […] If people aren’t interested in such matters… I mean, what's the classical definition of an idiot? It's people who don’t care about politics. And it doesn’t have to be… I mean, I don’t follow election campaigns that closely myself. […] But I’m socially minded in general. And if I am supposed to communicate with others, I expect to talk about more interesting stuff than gardening, doing up houses, and what the kids said at the kindergarten yesterday. I mean, I want to talk about more than just private stuff.

Obviously, Jarness is a skillful interviewer, capable of establishing trust. The journalist, to whom intellectual matters are important, appears enlivened by the conversation. Here is a display of emotions: like annoyance with people who are uninformed about politics and culture and from whom small talk is all one can expect. Knowing the local context (Vassenden and Jonvik 2019 and Jonvik 2015 are also from Stavanger), we surmise that one reason the journalist mentions such small talk about funny things that kids say, gardening and home renovations, and why he is annoyed with these people, is that he ‘has’ to interact with them, like with neighbors, acquaintances, relatives, through local school, etc. We also sense another emotion in the excerpt: amusement. The journalist seems to have a bit of fun deriding these people.

Skarpenes and colleagues make contrasting claims. Modeling their study on Lamont (1992), they conducted 113 interviews of upper-middle class people. Below are examples of data excerpts from this study (2007:553; our translations). The excerpts inform us about non-elitist and anti-hierarchical upper middle-class people, who as reported by Skarpenes, repeatedly refrained from making distinctions, unlike Jarness’ journalist. Three people, interviewed individually, are asked what being intellectual means to them:

That is something negative, bookworms, it gives negative feelings. Professor-like, in that direction. (Male, 50–60, law/economics)

The term itself implies that some are higher-ranking than others. It is exclusionary. (Female, 30–40, natural sciences)

Extremely tiresome notion. That is a cultural elite that are cultural snobs, those who have been in academia too long and don’t know reality. They are condescending towards those who for instance like Idol [cf. American Idol]. (Female, 40–50, social sciences/humanities).

On the surface, these are entirely different statements from Jarness’ interview excerpt; from the ‘elitist’ and ‘snobbish’ journalist. They are no less emotional. We do not sense amusement, but certainly annoyance: they are indignant with intellectualism and cultural hierarchization, that is, people who shamelessly disregard being down-to-earth and disdain those who like popular culture. The emotional displays suggest that Skarpenes too facilitated trust in his interviews. They also show that these issues matter to his informants.

Partial Truths, but General Claims

The easiest interpretations of these contrasting findings would be that one or both authors conducted bad interviews, or that they sampled differently. We believe neither is correct. Rather, if Jarness had interviewed one of the above, the answers would not be so condemning of intellectuals. Conversely, if Skarpenes had interviewed the journalist, he would not have expressed his elitism so emphatically. Most likely, these were different situations, framed by different types of interviewing, that opened up to different accounts, and different parts of the lives of their interviewees. Both Skarpenes and Jarness claim to describe middle-class identity, but their class portraits are conflicting. We suggest that rather than one or both being wrong, Skarpenes and Jarness teased out different (but real) aspects of middle-class identity—that is, partial truths.

Later, we analyze the social situations of each of the two authors’ interviews. Before doing that, we substantiate why we consider both Skarpenes’ analysis of egalitarian sentiments and Jarness’ analysis of distinctions to be significant truths about the Norwegian middle class. This pertains to a social duality shown in several Norwegian studies of social class, and supported and contextualized by historical and statistical information. Norwegian egalitarianism, of which the generous welfare state is an emblematic current expression, is a well-established historical backdrop, dating back at least to the 19^th century, with freeholding farmers and the absence of aristocracy (see Myhre 2018). The public unity school (enhetsskolen) both embodies and produces egalitarianism. It enrolls 96 percent ² of Norwegian pupils, who complete the same nationally standardized ten years of schooling, with the same curriculum. Since the 19^th century, its premise has been integration across social divides (Slagstad 1998). The other side of this duality is that, irrespective of egalitarianism, cultural consumption in Norway follows the same social divides as in other countries (Mangset and Hylland 2017), as does recruitment to higher education (Andersen and Hansen 2012; Heggen, Helland and Lauglo 2013).

Next is an excerpt from our own interview study on the same topic (Vassenden and Jonvik 2019). Curiously, several interviewees (from the same city as Jarness’ interviewees, Stavanger; N = 39) express both intellectual/cultural interests, like in Jarness’ study, and anti-intellectualism, like in Skarpenes’. However, as our example ‘Dag’ suggests (below), encapsulating broader tendencies in the study (see Jonvik 2015), egalitarianism and cultural capital belong to different social situations in middle-class life. (The following three paragraphs appear in Vassenden and Jonvik 2019:48–49.)

Dag is a chartered engineer, and decidedly part of Stavanger's economic elite. He holds an executive position in the petroleum industry. He describes his friends and what they have in common: higher education and intellectual curiosity. Still, he is uncomfortable speaking freely about such qualities, and does not easily use words like ‘intellectual’.

(Dag:)

My social circle (…) we never use the word intellectual, because we know … it’s not understood, not accepted. When I use it now, to you: I guess it has to do with … what kind of conversations you can be in and feel comfortable and good (…)

(Interviewer:)

Why is that word not used freely? (…)

(Dag:)

I almost don’t use the word academic either. Except when you sit here now. Well, it probably has a bit to do with that we (Norwegians) are taught, in good social-democratic spirit, not to emphasize differences (…) Because it creates distance (…) That is part of Norwegian culture—and something that I hold dear.

Intellectual capacity matters when Dag makes friends. He feels comfortable when people are on the same ‘intellectual wavelength’. That, however, is not easily articulated. Reformulated, Dag says ‘it’s important that my friends are intellectual, but don’t tell anyone’.

Unlike Jarness’ and Skarpenes’ implicit claims to uncover a middle-class identity as concerned mainly with cultural distinction or egalitarianism, both are important to Dag’s identity. Thus, not only can one find different types of boundary work among different middle-class people; we also find contrasting boundary work within the same interview with a single person.

Other authors report similar findings. Mangset's comparative study of elite bureaucrats in Norway, the UK and France (Mangset 2015; 2018) witnessed similar tensions between elitist awareness and egalitarian sentiments, as with ‘Dag.’ Discomfort about elite status was especially pronounced among Norwegian interviewees. Jarness himself makes numerous notations of downplaying of cultural capital in cross-class encounters (Jarness and Friedman 2017), although he portrays it as largely strategic in a Bourdieuian sense, as concealment of class condescension (e.g., p. 17; 21). Further, in a study of the Norwegian cultural elite, Ljunggren (2017) showed how its members de-emphasize information about their elite status when in cross-class encounters (cf. Gullestad 1989). Both egalitarianism and elitism seem to matter to their identities, but they balance displays of each. In encounters with non-elites, they downplay elite tokens like a professor's title (cf. Vassenden and Rusnes 2022) Ljunggren's interviewees do this to avoid making both their interactants and themselves uncomfortable in the situation (p. 569). This, apropos, supports our own Goffmanian interpretations (Vassenden and Jonvik 2019; Vassenden and Rusnes 2022); that downplaying cultural capital is a situational matter of ‘[…] tacit cooperation in face-saving’ (Goffman 1982[1967]:29), safeguarding the ‘sacred self’ of interactants (ibid).

To reconcile the contradictions between Jarness’ and Skarpenes’ findings, we now turn to ‘methodological situationalism’ and Goffman's theories, for a close reading of the interview situations. We follow Tavory (2020) in understanding interview situations as ‘refracted contexts’ for inference about ‘landscapes of meaning’ (cf. Reed 2011), or cultural repertoires. As our discussion reveals how contradictions between interviews can correspond to contrasting social situations outside of the interview, we next discuss how different social situations evoke different repertoires.

4. Theories to Assist in Situational Interviewing

‘(M)ost micro-sociological approaches conceive of social situations as a reality sui generis which entails a dynamics and organization of its own that we cannot predict from knowledge of the attributes of single actors’ (Knorr-Cetina [1981]2015:9). Such views are referred to as ‘methodological situationalism’ (ibid.). This branch of sociology is very different from attitudinal research, and from Parsonian sociology of internalized norms. It also contrasts with Bourdieu's theory of practice (1977), where the embodied dispositions of the habitus predispose agents to act in (class-)specific ways. Sociologists who treat ‘(…) interaction in social situations as the relevant methodological “units”’ (Knorr-Cetina [1981]2015:8–9), study, with Goffman's words, not ‘men and their moments. Rather moments and their men’ ([1967]1982:3). ‘Encounters make their encountees,’ as Collins, an influential contemporary proponent of Goffmanian microsociology, puts it (2004:5).

Underpinning this idea of the methodological (and ontological) precedence of situations/encounters, is a particular view of the human self. Goffman ([1959]1987), like Mead before him (1934), conceived of the self as reflexively constructed in relation to others; never passively defined, the actor co-defines the situation and her role in it. Crucially, however, she does so as an interactant, that is, attuned to the other actor/s and audience/s present and involved, to ‘situational copresence’ (Collins 2004:23). Situational copresence ranges from mere ‘physical copresence’ like public pedestrian movements, to full-blown interaction rituals (ibid.).

Although our focus is not rituals as such, Goffman's ideas (transplanting Durkheim's) about encounters and conversations as rituals hold unexamined potential for thinking about interviews, and prove useful as we try to solve our ‘puzzles.’ Collins, in a Goffmanian vein, defines ritual as ‘(…) a mechanism of mutually focused emotion and attention producing a momentarily shared reality’ (2004:7, emphasis added). An interview is not naturally occurring talk, and typically does not charge the participants with the emotional energy of a successful ritual that leads to group formation (in Collins’ terms; ibid.) but is nevertheless a momentarily shared reality. Interviews can certainly involve emotional displays, as in the examples from Jarness and Skarpenes, which show that the topics mattered to their interviewees. When an interviewer and an interviewee meet for an appointment in a café or workplace, the mere physical copresence they might have just had in the lobby of the office building, perhaps unknowing of each other, becomes in Collins’ words, ‘(…) a full-scale encounter by becoming a focused interaction’ (ibid.).

Goffman conceived of conversations as small social systems ([1967]1982:113) that put demands on participants for agreement, for respecting the conversation as ‘(…) a reality that is at least temporarily believed in’ (Collins, ibid.). Like a natural conversation, an interview—especially a good one—is a shared reality, which researcher and participant produce together, and which both can come under pressure to honor.

Goffman analyses situations as rituals centered on the self. Conversation is itself a ritual. If the conversation is successful and as people become engrossed in it, it is enhanced into whatever tone of humour, anger, interest or anything else which might emerge with the flow of talk. The result of this conversational ritual is to create a little temporary cult, a shared reality consisting of whatever is being talked about. (…) (O)nce the conversational ritual is in full swing, it builds up its own pressures which control its participants. The topic has to be respected, at least temporarily believed in; it has become, for ever so short a time, a sacred object to be worshipped. (Collins 1994:72, emphasis added)

Let us translate to interviews. If we conceive of interviews as encounters that may (like conversations) become ‘small social systems’ (Goffman); temporary little ‘cults’ (here is Durkheim's legacy); and ‘realities temporarily believed in’ (Collins), the discrepancies among the reviewed middle-class studies pose less of a puzzle. We suggest that both Jarness’ and Skarpenes’ participants tuned into and honored momentarily shared realities: Their interviews were different social systems, and different realities momentarily shared. The ‘working consensus’ (Goffman [1959]1987:21) between Jarness and his ‘elitist’ journalist seems to have been: ‘Let us agree that as middle-class people, intellectual issues matter to us. Having agreed upon that, while still acknowledging that others will regard this negatively, let's have some fun at the expense of others’. Quite possibly, a ‘temporary little cult’ even developed around the fun of breaking an egalitarian ‘taboo’. If our previous surmise is correct, moreover, that Jarness’ journalist mentioned his annoyance with people who fall into small talk partly because he sometimes ‘has’ to interact with them, then playing with ‘forbidden feelings’ may also have been even more fun to him, making the ‘temporary little cult’ all the more appealing.

Skarpenes and his interviewees honored a contrasting (momentarily) shared reality, more like: ‘Let us agree that as middle-class people, it matters that we get along with all people. Having agreed upon that, let's identify the threats to people getting along’, such as intellectualism. However, as Dag showed, one reality (or ‘cult’) does not exclude the other. Rather, people are more than capable of participating in both, but they cannot easily do both at the same time. Participants come to such agreements, and honor them, because they attune to each other's presence. Jarness’ journalist appeared enlivened by the topic of intellectuality in his life. This seems to have been Jarness’ and the journalist's ‘(…) little temporary cult’ (Collins 1994:72). In contrast, the object of worship in the competing ‘little temporary cult’ of Skarpenes’ interviews, was egalitarianism.

Beyond Frontstage and Backstage

We should credit Jarness for methodological reflexivity, and for a nuanced stance on interviewing (Jarness and Friedman 2017; Sølvberg and Jarness 2019). In a paper co-authored with Sølvberg, Jarness (2019) relies on Pugh; (2013) distinctions about types of information generated in interviews. Pugh suggested four types: honorable, visceral, schematic, and meta-feelings. Sølvberg and Jarness engage with the honorable and the visceral. Whereas a visceral narrative/self concerns practical consciousness (ibid.), i.e., ‘instinctive’ and innate, an ‘honorable’ narrative/self is reflective, and concerns how one wishes to appear to others. An honorable narrative mirrors social desirability. Sølvberg and Jarness reflect well upon how interviews differ regarding these two selves, and in line with our reasoning, suggest that the display of these selves mirrors tensions that people experience in their lives, and the social encounters with people similar to or different from themselves (ibid.; cf. Jarness and Friedman 2017:22). They further suggest that the ‘visceral’ self belongs to Goffmanian ([1959]1987) backstages, the emotional area of disgust and the like, and (implicitly) that the ‘honorable’ self is exhibited on frontstages. The latter is a ‘tolerant front’: people present themselves in a ‘socially desirable light’ (Sølvberg and Jarness 2019:2). To maintain a ‘tolerant front’, people show open-mindedness and acceptance of people as they are. Linking back to the journalist, then, Jarness implies that he got to the backstage and the journalist's visceral self, which is ‘truer’ than the honorable self. An implicit claim in Jarness’ authorship on middle-class identity is thus that people appear egalitarian, yet beneath, their identity formation rests on cultural distinctions.

However, what is backstage and frontstage in qualitative interviews is not straightforward, especially if we take situationalism seriously. As our own studies (Jonvik 2015; Vassenden and Jonvik 2019), and those of others show, both egalitarianism and cultural distinctions matter to middle-class people—the emotional displays among Skarpenes’ and Jarness’ interviewees, on separate notes, suggest likewise—but they depend on the social situation, with clear contrasts between e.g., inter- and intra-class encounters. (Jarness shows this himself, with Friedman, but interprets it as a Bourdieuian ‘strategy of condescension’; 2017:22.) However, one is not ‘truer’ than the other. Moreover, what constitutes the ‘socially desirable light’ (cf. J-K 2014a) will also vary. If the ‘working consensus’ (Goffman [1959]1987:21) highlights sameness (as in Skarpenes’ interviews), then the interviewee will likely present a ‘tolerant front’ (Sølvberg and Jarness 2019:9). In that case, being open-minded and non-judgmental—or egalitarian—about how other people live is frontstage and ‘socially desirable’. If, in contrast, the ‘working consensus’ relates to cultural competence, then Jarness’ journalist is in fact frontstage, even though he may be disclosing a well-kept secret (his snobbery), and even though he may see this as a welcome occasion to deride people he encounters who talk about mundane things. (When in such encounters, he almost certainly keeps such thoughts to himself, as facework; Goffman [1967]1982:29.) In that sense, a visceral narrative like from Jarness’ journalist may be ‘honorable’, or ‘socially desirable’, in this situation. Put differently, in cases like these, with competing values and ideals—tolerance vs. cultural competence—what is the ‘social desirability bias’ depends on what becomes the mutually shared reality. The journalist possibly presented a ‘culturally competent and playful front’, to honor the demands of the little social system (Goffman) that the interview had become. Based on our experience with interviews on this topic, this might well have been the case. That is not to say that this moment did not represent a truth. We believe that it did.

Our's interviewee Dag hesitated to use the word ‘intellectual’ even though he claimed to care about the intellectual capabilities of his friends. We suspect that if he had been interviewed by Jarness, he would appear more ‘intellectual’—to present a ‘knowledgeable front’ in the interview, and to honor that little social system. If Dag had been interviewed by Skarpenes, however, he would have been critical of the word ‘intellectual’, but it would probably be further from his mind to mention that he socializes mostly with people with ‘intellectual capabilities’. This would have been his response to the situation's ‘tolerant front’, his way of honoring that little social system. These facets—cultural capital (similarity to ‘people like me’) and egalitarianism (getting along with ‘all kinds of people’)—are important truths about the lives of many middle-class and elite people in Norway (and probably beyond; e.g., van den Haak and Wilterdink 2019 about the Netherlands). However, these truths have different legitimacy in different social situations. Our interpretation is that Jarness’ and Skarpenes’ interviews echo contrasting social situations outside of the interview, that is, in the lives of the interviewees.

Both cultural capital and egalitarianism matter to many middle-class people, and combine in intricate ways. These are two aspects of the self that they perform differently depending on the social situation. One possible (but insufficient) interpretation would be that Dag will hold up egalitarianism (‘tolerant front’) in public situations, and that this is how he justifies his opinions, by reference to egalitarianism as a Norwegian ‘common good’. Privately, the argument would go, he shares his ‘real’ thoughts about networking only with people like himself. Although Jarness discusses such methodological issues in detail (Sølvberg and Jarness 2019), to conceive of the interview as backstage or frontstage is simplistic. What is frontstage and backstage in interviews is far more complex than the dichotomy.

Situational Awareness Beyond the Interview Encounter

Our interpretation is that Skarpenes’ and Jarness’ interviews addressed different situations with different ‘imagined audiences’ and ‘imagined communities’ in their participants’ lives. Without direct inspiration from microsociology and situationalism, this would not be easily detected. Like the contrasts between inter- and intra-class encounters, an interviewee presents different aspects of herself—draws on different repertoires—if she sees the situation before her as relating to society in general, or for example to other parents, in her children's school, on the touchline at one of her daughters’ soccer games, or at the other daughter's orchestra rehearsal. She draws on different features of her personality and different values if she imagines her colleagues (and different if superiors, equals or subordinates), best friends, parents, or extended family. Moreover, depending on communities and audiences, ‘moral stakes’ will be frontstage in some interviews. In other interviews, ‘cultural stakes’ are frontstage, leading the interviewee to uphold an ‘knowledgeable front’, to which other cultural resources than egalitarianism/tolerance offer arguments. Both are constitutive to the selves of many middle-class people in Norway (and beyond), who are capable of performing both. In short, we suggest that many contradictions in interviews are fruitfully addressed as contradictions among multiple communities with which one and the same person can identify. Different social situations engage different social communities and audiences and mobilize different cultural repertoires.

Interviewers would be well advised to engage with the insights from cultural sociology. Swidler's cultural ‘toolkits’ (1986), Lamont's work on valuation (1992; 2012; Lamont and Thévenot 2000), and French pragmatism (Boltanski and Thévenot 1999) stress how actors relate to multiple logics of valuation or rationalities. As also highlighted in Goffman's ([1959]1987) and Mead's (1934) views on the self, actors are both capable of and prone to drawing on different parts of themselves, different memories, different values, or logics of reasoning depending on the situation and with whom we interact.

Deploying such reasoning to interviews means that, yes, the working consensus of an interview is constructed by interviewer and interviewee, tied to that situation. And yes, with Tavory (2020), we can approach this situation as a refracted context for identifying ‘landscapes of meaning.’ Yet, a crucial task for situationalist interviewing, is to make substantiated interpretations of which situation(s) elsewhere to which the interview situation refers or corresponds, and thus also of which imagined community/audience that is at play in the interview. A situationalist interpretation of the interview must strive to identify the type of situation that the interview points to, to which can ‘belong’ different repertoires. ‘Landscapes of meaning’ (Reed 2011; Tavory 2020) is a useful concept, but imprecise. We suggest going a step further, and theorize how different situations in the world direct people to different parts of these landscapes.

As our examples showed, interviews are sometimes ‘backstage’, sometimes ‘frontstage’. With Skarpenes and Jarness, apparently contradictory findings on egalitarianism vs. cultural capital seemingly related closely to whether interviewees understood the interview situation as resembling a private situation (where one has fun with like-minded people) or a public one. Yet, as we also showed, these issues are far more complex than the simple frontstage/backstage metaphor, or a simple private/public distinction can account for. Actors continuously move across shifting situations, to which ‘belong’ different repertoires, and actors competently navigate these shifts. These insights hold important, but neglected, lessons for how we plan for, conduct, and interpret interviews.

Inference to Landscapes of Meaning

This also relates to Tavory's typology of open, closed, and refracted contexts. As stated, our examples are of the refracted contexts for inferring about imaginations and repertoires; albeit these imaginations are fruitfully addressed analytically via situations elsewhere. However, Tavory's metaphors are weighted towards the singular. If we build on our notion of ‘imagined audiences’ in the plural, which reflect ‘realities temporarily believed’ (Collins) in the plural, we should think of ‘windows’ in the plural too. Taking methodological situationalism seriously would mean that an interview can concurrently open several windows; to different and even contradictory aspects of an interviewee's life. Following from that Skarpenes’ and Jarness’ interview situations apparently engaged contrasting situations in middle-class life, we can say that refracted light came into their interviews through separate windows overlooking different ‘restrictive areas’ of the landscapes. In Skarpenes’ interviews, the window overlooked an ‘egalitarian area’ of the landscape of middle-class imaginations; Jarness’ interviews opened yet another window onto an ‘intellectual’ area.

Depending on how the interview situation unfolds, it will steer the interviewees to potentially contrasting communities and values. Depending on which ones that in the situation appear most legitimate, appropriate, or authentic, the interview will point to different parts of the surrounding cultural universe. This is a crucial point not addressed in Tavory's article, and seemingly what happened in the studies we reviewed. If the researcher is not sufficiently aware of this, s/he may make questionable assumptions about people's position(s) in the world and about the repertoires mobilized in this/ese position(s). If the interviewee saw herself as inhabiting one situation, and the researcher analyzes the interview statements as indicative of another, or simply as general statements about attitudes, disposition, and worldview, s/he may well arrive at flawed conclusions. Although the interviewer can never know for certain how the interviewee understands the situation (we do not make such claims), what we must expect from qualitative sociologists (but which is usually wanting!), is situational awareness and thus explicit arguments for how the researcher interprets the interviewee's understanding of the situation. Next, we offer practical advice for how sociologists can achieve such awareness, and plan and conduct studies accordingly.

5. Practical Advice for Situational Interviewing

Here, we distinguish between two circumstances of research: In one, we have at our disposal a sample of interviews produced without explicit situationalism. In that case, advice pertains mostly to interpretations, and the preceding examination was largely a case of this. We now attend more to planning and conducting situational interview studies, and highlight pre-studies, vignettes, photos, and ‘toying’ with audiences. The techniques are not new or unfamiliar to qualitative sociologists. They are standard measures like reflecting on interview schedules, phrasing and sequence of questions; approaching the interviewee; matching of interviewer and interviewee or not, by gender, class, race/ethnicity, age; considerations of power; location of the interview, etc. Yet, the techniques have not been discussed within the context of fine-tuned situational awareness. What is new, is how we use and think about our usual methods. We also reiterate that we do not claim to solve problems like the ABC or social desirability bias (or misunderstandings etc.). Our concern has been the value of situationalism for interviews, which helps ease these problems. At the very least, by making the situationalist interpretations data explicit to readers, analyses become more transparent. While the problem of inference remains, situational awareness helps alleviate it. Here also follows that, to speak of social desirability bias in the singular is simplistic; just like the backstage/frontstage as one dichotomy is simplistic. Different imagined audiences/communities produce different social desirabilities (plural), like we witnessed with Skarpenes and Jarness. To a degree, situational awareness also works through social desirabilities, in identifying when such are at play, and which communities they address.

In planning research, situational awareness implies careful considerations of whether the topic is likely to make interviewees see themselves in private or public situations, and which type of public (e.g., newspapers; TV; professional situations), ‘semi-public’ (e.g., social media; lectures; in-house business), and private situation (intra-group vs. inter-group; family; friends; relatives). Each comes with specific backstage-frontstage configurations.

To enable ourselves to make these considerations, there are several things we can do. One is to conduct situationally oriented pilot studies to grasp the social situations that make up the lives of people in our target group, and how these people move across them. That is, researchers should familiarize themselves with the most important real-life situations before starting the full-scale interview study. This could be pilot interviews that map the different social situations in the lives of their informants. Pilots could also include (limited) observation/shadowing. We suggested earlier that if Khan were to do a follow-up of his study of St. Paul's (2011), ethnographic shadowing of a broad range of the St. Paul alumni would be unfeasible. However, shadowing of a couple of St. Paul alumni could be a valuable pre-study for a subsequent broader interview study. Observations could be used to write up the typical contrasting situations that make up their lives; from boardrooms, family-life, offices, neighborhoods, politics. These descriptions could then be used as input in the more formal interviews with the larger sample. Snapshots of the social situations of these few individuals could be used as vignettes: ‘Do you recognize this situation?’ ‘Tell me what's happening.’ ‘Explain what is different in your life?’

Vignettes derived from pre-studies could thus describe different social situations that the researcher now knows are likely to be pertinent or recognizable to the informants. For instance, if Jarness’ journalist were interviewed in a pilot study, the excerpt could be used to construct a vignette for later interviews. We could extract the main content, use a few quotes, and then ask subsequent interviewees: ‘Who is this person?’ ‘To whom is he talking, or who does he have in mind?’ Then, we could toy with the imagined audiences, and ask the interviewee how she imagines that the journalist would speak/act in the presence of the people he refers to, and ridicules.

Toying with audiences is general advice. We should make interviewees describe the people they have in mind when they present their narratives. Then, we ask them to imagine these people replaced by others and tell us whether and how this changes matters. Like if a parent talks about picking up her children from kindergarten, and mentions that she is careful about revealing information to other parents about her work. If she is a cleaner or a hairdresser (i.e. someone of low occupational status), we should ask her what jobs she suspects the other parents have, and then make her draw a contrast: ‘Let us now replace that mother who is a professor and that father who is a CEO, with a mother who is a recent immigrant and a father who is a construction worker. What happens?’ We should also try to shift the interactional contexts (not just the people), and replace, say, the kindergarten with a workplace. This, of course, requires attentive and active interviewing, and it takes training.

Regarding social desirability, vignettes that describe other people in social interaction, might allow interviewees to feel less responsible for the moral aspects of the inter/action described. The interviewee is now to account for the thoughts and actions of another person resembling herself rather than those of her own. We have introduced a ‘third person’ in the interview, an imaginary person via the vignette. We have suggested elsewhere (Vassenden and Jonvik 2022) that photo elicitation (Harper 2002) can work this way; photos can shift interview situations by being what actor-network-theorists term ‘third actants’. Regardless of technique, the value lies in switching interaction; we can conceive of this as changing the interview from dyadic to (imaginary) ‘triadic’ situations (via vignettes and/or photos). This may bring some relief from the possible face-saving rituals (Goffman [1967]1982) in the dyadic encounter of interviewer and interviewee. An object like a written text or a photo also provides a pause from eye-contact. Such shifts may also ameliorate some of the problems of interviews being situationally bounded (J-K, 2014a), because we now have what we can conceive of as two situations within one (the interview), dyadic and ‘triadic’, between which we can cross-check and validate responses (see Vassenden and Jonvik 2022). With photos, it would of course require meticulous preparation to select images that touch on relevant social situations in the lives of our interviewees. As with constructing vignettes, the selection of images should happen in dialogue with participants in a pre-study.

The point will in any case be to get a good impression of the varying situations that make up the lives of interviewees, and then work to understand how the interview as a social encounter match these situations, and which ones at any given time.

This then becomes crucial knowledge for analyses of interviews. When researchers encounter contradictions in interviews (like between the middle-class studies), these situationalist interpretations hold particularly high value. This may be especially pertinent in studies that were not originally planned with a situationalist approach.

Then, when one arrives at a situational interpretation of interviews—whether projected from the start or used to elucidate a contradiction—one should examine a broad range of contextual data (and other studies) to validate whether and how interview situations correspond to real-life situations. This can be historical knowledge, statistical distributions, observational data when they exist, as well as knowledge of the research field itself. With the middle-class contradictions, such contextual information was important in our interpretation of social duality. Moreover, one should do what qualitative interviewers should always do (but typically do not); validate one's interpretation with people in the target group, preferably both within one's own sample and others. If possible, re-interviewing the same people, to validate situational interpretations, and then interviewing them in groups and comparing the outcomes with the individual interviews, is recommended. This (like all the preceding advice) will be a fine balancing act so as not to compromise feasibility, which is after all one of interviewing's benefits vis-à-vis ethnography.

6. Concluding Remarks

Situational critiques of interviewing hold important lessons for sociologists, but for different reasons from those cited by many ethnographers, who sometimes implicitly discard the interview entirely. As shown, however, ethnography, like any method, is far from a panacea. Sociology cannot make do without interviews.

Nevertheless, the problems raised are real and we must tackle them head-on. Since meaning is situation- and context-bound, we cannot simply transfer meaning produced in one situation (the interview) to meaning and action in others (real life). These issues are more important than researchers who rely on interviews typically acknowledge.

This criticism of interviews invokes the interactionist tradition. However, to let this rich tradition serve mostly to discredit interviews is highly unproductive. By directing Goffmanian microsociology onto the interview itself, we suggest that interactionism/situationalism, which critics see mostly as problems, holds much value to interviews.

We have shown how situational awareness matters for how an interview should be interpreted, and what kind of data it provides, by comparing two studies of middle-class identity, conducted by Jarness and Skarpenes. We attribute their contradictory findings to their interviews having been different situations, which correspond to different situations in their interviewees’ lives, which, in turn, address different audiences/communities, so that participants evoked contrasting cultural repertoires. Situational awareness in interpreting these interviews suggest that middle-class people can adhere to both egalitarian norms and practices of cultural distinction, and that such duality is often significant to their class identity. Without subjecting these contrasts to situationalism, it would be difficult to make sense of these dualities and ambiguities.

Interviewers should acknowledge the importance of situational definitions of the interview, and how such definitions relate to variation across external situations—captured in our concept of imagined audiences and communities. Situational awareness should be critically addressed in designing our studies, planning for interviews, conducting them, and analyzing and theorizing them. A main lesson from our paper is also a hugely important promise: Situational interviewing helps us make sense of contradictions, ambiguities, and disagreements within and between interviews. (Self-)contradictions and ambiguities could lead some to doubt our findings or even the value of interviews altogether. Situationalism offers an alternative, which systematically and meticulously examines whether and how contradictions correspond to contrasting real-life situations. Such examination presupposes that we avoid the pitfalls of either rejecting interviews because of situationalism, or tacitly neglecting situationalism in defense of the interview. If we can successfully develop methodologies for situational interviewing, we should be able to move beyond simplistic dichotomies such as positivism/constructionism, frontstage/backstage, ‘miner’/‘traveler’, accounts/behavior, and indeed truths/lies.