Week 7 Discussion: The Interview Process

Asking Questions with Reflexive Focus: A Tutorial on Designing and Conducting Interviews

ZEYNEP ARSEL

Interviews have been a key primary data source for research published in the Journal of Consumer Research. This tutorial aims to walk readers through the de- sign and execution of interview-based empirical research on consumers and consumption.

Keywords: interviews, research methods, data collection, ethics, qualitative

research

INTRODUCTION

I f you pick up an empirical article published in this jour-nal that uses qualitative methods, you will most likely come across one that uses interview data. While there are

exceptions in cases of historical research, or those employ-

ing macro-level data analysis, interviews have been a key

primary data source for articles published in the Journal of Consumer Research. This is rooted in the epistemological tenet that consumers’ lived experiences can primarily be

understood through their expressed subjective narratives

(Thompson, Pollio, and Locander 1994). Interviews are

useful because they give voice to people’s lives and their

perceptions of experiences important to them (Belk,

Fischer, and Kozinets 2013) and allow the researcher to un-

derstand the way they see the world (McCracken 1988;

Thompson et al. 1994). While the number of nonprimary

data sources that provide such narratives has skyrocketed

(see Humphreys and Wang [forthcoming] for a discussion

on sources for consumer-generated data), the interactive,

flexible but focused nature of interviews still makes them

one of the most trustworthy and effective sources of data

about consumers. Note that not all research questions can

be answered by interview data: think market-level studies,

macro approaches, and historical methods that require less

individualistic accounts of consumers. Still, even these

studies can benefit from incorporating interviews as a sup-

plemental source of insight. The focus of this tutorial is to

walk you through the process of designing and conducting

interviews, whether they are the primary data source for

your qualitative research projects, or secondary. I will start with a somewhat controversial and contest-

able position. While it is common in this journal to refer to

interviews as “unstructured,” I suggest you shy away from

doing truly unstructured or unstandardized interviews,

even if you claim to do grounded theory. The idea of hav-

ing exploratory and fully unstructured dialogues with your

participants, where you let their experiences unfold as they

will, remains a romantic myth and would be an unfruitful

pursuit for the type of contribution this journal seeks. As

much as interviews are performative and constructivist

(Alvesson 2003), you as a researcher should enter the inter-

view with a research question, albeit a loosely developed

and mutable one. You should have in hand a set of themes

to explore while being open to the new directions presented

by each interviewee. Fortunately, you also will bring all

the theoretical baggage you have gathered as a scholar,

drawing on particular frames of understanding and particu-

lar ways of seeing. For example, if you were trained in the

Zeynep Arsel ([email protected]) is Concordia University

Research Chair in Consumption and Markets at John Molson School of

Business, Concordia University, MB 11-109 1455 Maisonneuve Blvd.

West, Montreal, Quebec H3G 1M8, Canada. She would like to thank the

editors for their invitation and their constructive feedback on the earlier

versions of this tutorial, as well as her students Ghalia Shamayleh and

Tevfik Karatop for their feedback and assistance.

This tutorial was invited by editors Darren Dahl, Eileen Fischer, Gita

Johar, and Vicki Morwitz.

Advance Access publication September 5, 2017

VC The Author 2017. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc.

All rights reserved. For permissions, please e-mail: [email protected] � Vol. 44 � 2017 DOI: 10.1093/jcr/ucx096

939 Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

Bourdieusian tradition, you will most probably see social

class in most answers your participants articulate. Treat

this as an advantage, but be reflexive regarding your own

theoretical blind spots. A contrast can be made between interviews that are stan-

dardized (you ask the exact same set of questions in each

repetition) and those that are semistandardized (each inter-

view takes its own form but follows a specific research

question around a series of themes) (Berg and Lune 2012).

The former takes a deductive approach, using data to verify

hypothesized propositions, and is better suited to mixed-

method designs where qualitative inquiry is used to fill in

gaps from other types of data. 1

Due to space restrictions,

this tutorial focuses on the latter, highlighting the induc-

tive, emergent, and iterative nature of the type of qualita-

tive research that gets published in this journal, while

underlining that you should never pursue freestyle shotgun

inquiries in the hopes that you will discover a theoretical

needle in a haystack. Why have some sort of structure and focus instead of

diving right in? Interpretive research should indeed rely on

the data to speak, but interviewing people without engag-

ing in much theoretical and methodological preparation be-

cause you have a hunch about something interesting in a

context might lead you to more noise than data. You can

end up with interviews that are disjointed, discretely idio-

syncratic, or redundant in light of literature because you

asked about issues that the existing theories already ad-

dress. Why not, instead, purposefully but reflexively let

your prior understandings and your research question(s)

guide your study design? A good model to follow is the ex-

tended case method, coined by Gluckman (1958) but later

elaborated upon by Burawoy (1998). While not really a

method per se, this epistemological approach suggests one

should challenge existing theories reflexively to recon-

struct new knowledge. For the purposes of the kind of con-

tributions that get published in the Journal of Consumer Research, a fitting way to consider interviews is, therefore, seeing them as orchestrated dialogues geared toward

knowledge extension, employing a “reflexive pragmatic

approach” (Alvesson 2003). In other words, your inter- views should reflexively and purposefully be designed

around existing theory, challenging it, and seeking to re-

vise or extend it. A focused approach doesn’t have to be deductive and

confirmatory. While situated in the context of each inter-

viewee’s life world, the interview should have an overarch- ing purpose that persistently and progressively seeks new

knowledge around an ever-evolving research question. Each data point and each iteration within the research pro-

cess should compel you to reconsider your understanding and the motivating research questions. In that sense, even

though the process of conducting interviews is presented in a somewhat linear fashion here, interviews should not be

seen as the first stage of research design, nor should the re-

search design be considered finished before your first inter- view. Interviews, like all data sources in interpretivist

approaches, are a part of an iterative circle that continu- ously moves back and forth between conceptualization,

data collection, data analysis, and theory building. In fact, never do all your interviews at once, unless you have field-

related constraints such as a temporally bound event. Even

when you do this, use your downtime to reflect on your provisionary findings. If you don’t have constraints, spread

your interviews across a few months. Do a few interviews first. Analyze, revise, repeat.

A reflexive approach means you should be mindful of the

intersubjective nature of your encounter with your research participants (Wilk 2001), the power relations between you

and your participants (Kvale 2006), and your own biases and preconceptions. This is important because while you

have questions that guide the interview, your participants should also have control over their narrative and what they

want to say and how they want to say it. Furthermore, your

provisional findings should continuously challenge your assumptions, from your choice of words in the interview

questions to the way you interpret the answers to these ques- tions. In the next section, I will first discuss the differences

between ethnographic and formal interviews. Following that, I will walk you through a four-step iterative process of

designing and conducting interviews.

ETHNOGRAPHIC VERSUS FORMAL INTERVIEWS

Qualitative research interviews can be either ethno- graphic or formal. A recent example of research that

includes both is Maciel and Wallendorf’s (2016) ethno- graphic study of male craft-beer aficionados. While the

authors primarily use participant observation, their data also includes both ethnographic and long or formal inter-

views to deepen their understanding of these observations. Maciel and Wallendorf not only use conversational-style

1 If you are doing exploratory interviews for a primarily experimental or survey-based project, you can be pragmatic. Consider as an exam- ple the highly cited scale development piece by Dabholkar, Thorpe, and Rentz (1996). The authors use interviews alongside observational methods to identify dimensions of service quality, and then consider these dimensions in light of the existing conceptualization of the con- struct. Then they develop a scale. In cases like this, many of the con- cerns raised in this tutorial might not be relevant. For example, if you are developing a scale, it would be overkill to ask participants for long life-history narratives to better situate their consumption practices into their life worlds. Or if you are using interview data to realistically complement scenario-based laboratory experiments about the psycho- logical outcomes of tragic decisions (Botti, Orfali, and Iyengar 2009), and you are not interested in how these effects pattern socioculturally, you might be fine focusing on these decisions and their aftermath, in- stead of devoting time to lifestyle questions. In these cases, your ques- tions might also be more hypothetico-deductive, aiming at affirming or refuting your assumptions.

940 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

probes while brewing beer with their participants in their homes, they also schedule separate formal sit-down inter- views with them afterward. This is ideal; however, in some cases, you might not have the opportunity for such follow- ups. Therefore, your only source of interview data might be ethnographic interviews conducted during participant observation.

Ethnographic interviews are short, in situ, and im- promptu conversations that take place within the con- straints of the field site. This type of interviewing might require more emergent design, and more spontaneous ques- tions tailored for each observed moment to make the best use of time and space restrictions. Depending on your data collection site, you might be pressed in terms of how much time your participants can afford to give you and where you can conduct these interviews. For example, if you are interested in understanding the experiences of paintball players, as were Woermann and Rokka (2015), and were participating in a game, you might do short, pointed, and informal interviews with participants during the game downtime on the site. In such cases, you will be better off focusing on questions that reflect your observations of the particular experiences of that moment, instead of covering a predetermined set of issues that aim to get a bigger pic- ture. If you can get a bigger picture, by all means do so, but you will most probably need to prioritize.

In ethnographic interviews, going through a written ethics protocol and asking your participant to sign a con- sent form might also be impractical. Still, you cannot forego informed consent regardless of how awkward you think this could be. The ethical consequences of not getting proper consent will be a disproportionately bigger problem than trying to find a good moment in the flow to tell participants that you are a researcher. Do not interview people under the guise of casual chat. Start your conversa- tion by concisely and clearly explaining that you are a re- searcher studying [the subject matter], and are interested in hearing their experiences on this subject. If you are using a tape recorder, keep it very visible so that participants have no doubts about whether or not they are being recorded (though in most cases you might not even record these interviews). As soon as you leave the field site or take a break during data collection, write down your recollections in as much detail as possible.

The rest of this tutorial is written for the purposes of conducting a study where the primary data source is prescheduled and formal or long interviews.

A FOUR-STEP ITERATIVE GUIDE FOR INTERVIEW DESIGN

Step 1: Settle with an Epistemological Tradition

There are never-ending epistemological debates on what kind of scientific truth interviews reveal and how

knowledge can be produced through interview data. This is not the tutorial to revisit paradigm wars. However, it is im- perative that you familiarize yourself with the epistemolog- ical differences between various interview traditions. Whether you are using a phenomenological approach that focuses on the lived experience of individuals as the princi- pal empirical evidence (Thompson 1997; Thompson, Locander, and Pollio 1989; Thompson et al. 1994), a neo- positivist perspective that seeks patterns and quasi-causal explanations (McCracken 1988), or a social constructivist lens that sees interviews as cultural conversation or episte- mic practice that invites both interviewee and interviewer to contribute to knowledge production (Brinkmann 2007; Moisander, Valtonen, and Hirsto 2009), you should have a clear understanding about what you expect interviews to reveal and what kind of theoretical stories you can tell with interviews. While undoubtedly shaping the ways you ap- proach your research question, design your study, and col- lect your data, a coherent epistemological position is even more important when it comes to analyzing this interview data and making theoretical claims about the social world you are investigating.

Step 2: Prepare an Interview Protocol

An interview protocol is an outline of your interview, listing key points of exploration, provisional questions, and planned probes and transitions. Your protocol should incor- porate three components. It should start with a brief intro- duction in lay terms, involving a description of the research project, an explanation of interview procedures, and an invitation to the interviewee to ask questions about the study and procedures. You do not tape this portion of the interview. This is followed by a procedure for estab- lishing informed consent: the interviewee is informed about the consequences of participation and is asked to provide explicit consent to be interviewed. A signed writ- ten document highlighting the procedure and consequences of participation is preferred to avoid ambiguities. Some in- stitutional ethics review boards will not allow you to get oral consent, unless justified or recorded on tape. Lastly, you should have a set of provisional interview questions.

Some researchers might tell you that they do not prepare interview protocols because they simply want to hear what their participants say. This is too idealistic and runs the risk of lacking focus. During an interview, you are not only listening to complex, compelling, and occasionally contra- dictory narratives, but also thinking about your next move by analytically deconstructing the meanings of these answers so that you don’t miss a follow-up opportunity. On top of this, you will need to use the vocabulary and natural language most appropriate for the context, instead of using the academic jargon that you use when talking to your col- leagues. For many of us, social settings with strangers can also be anxiety inducing. Navigating this situation without

ARSEL 941

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

a roadmap in hand is quite possible if you are someone with an impressive theoretical arsenal on the area of in- quiry, exceptional analytical skills that can help you inter- pret data in the moment, and sharp conservational abilities. Most of us are ordinary people and therefore could benefit from doing a bit of preparation. An interview protocol can (1) keep you focused during the process, helping you cover all the relevant concepts you have identified prior to this particular iteration, (2) give you a sense of control, (3) help you to analytically connect emic to etic (Wallendorf and Brucks 1993) by translating your research questions into natural conversations and vice versa, and (4) reignite con- versation when things get awkward, or redirect your con- versation when there is an extended digression.

Preparing Your Questions. Many resources will advise you to start each interview with warm-up questions to build rapport. As a consumption researcher, I find the term “warm-up” a bit puzzling, because the information you gather during this section of the interview is usually very important for contextualizing your participants’ responses. You are not asking to get to know your participants and their everyday lives just for small talk or to build rapport (even though this kind of conversation undoubtedly helps to warm things up); you are asking these questions because you want to understand how and why they do things. How and why are frequently tied to individual life histories and lifestyles. Therefore, the ubiquitous “Can you tell me about yourself?” opener is an important question that you should pay attention to. You should heavily probe the responses to this question, digging as deep as you can about back- ground, family, education, and current lifestyle. This is es- pecially important if your research is related to class, taste, aesthetics, community, or lifestyle practices. For example, if you are comparing low and high cultural capital upper- middle-class people as €Ustüner and Holt (2010) did, early life history, education, and family background will be ex- ceptionally important for your analysis. Of course, while getting to know your participants, you will also build some rapport. You can even throw in some small talk to put ev- eryone (including yourself) at ease, but don’t treat the first half-hour of your interview lightly.

Once you cross off the life history subsection, you need to slowly transition to the subject matter. Think about the concepts that you would like to elicit narratives about, what McCracken calls “cultural categories” (1988). For this, you need to reflect on both the existing theory and secondary data sources. For example, if you are studying roller derby grrrls as did Thompson and €Ustüner (2015), critical reflection on the representations and discourses sur- rounding this context will likely lead you to realize that gender is a central topic to cover in your questions. Note, however, that key concepts that emerged in Thompson and €Ustüner (2015)’s research, such as “edgework,” were prob- ably not immediately apparent to these researchers, but

most likely manifested themselves as they progressively analyzed their data. Once new concepts emerge, you can incorporate them in revised interview questions, carefully navigating the territory between gently steering the inter- view in this direction and putting words into your partici- pants’ mouths. More on this later.

During the early stages of your project, when you are conducting your first set of interviews, you will probably think about fewer (and more abstract) concepts. As you it- eratively build theory from data, you will be refining your interview protocol alongside your theoretical story. For ex- ample, when Jonathan Bean and I first started our taste regimes project in the context of Apartment Therapy (Arsel and Bean 2013), we immediately agreed that this would be a project on how taste is shaped by media narra- tives. From existing theories, we also knew that taste prac- tices are shaped by social class patterns. Our first-round interview questions, therefore, were heavily focused on so- cial class backgrounds of our participants, and their rela- tionship to the media brand. We also inquired about the layout and decoration of our participants’ homes. As we it- eratively analyzed our data, we realized that there was not much new to add to the existing theory if we wanted to tell a social class story. However, we were intrigued about how taste was manifesting itself in mundane everyday practices; something theoretically novel seemed to be emerging. So we started to ask even more detailed questions about objects, and specific domestic practices around these objects. While we still inquired about our participants’ so- cial class to ensure that we could locate their social posi- tions, what they said about this remained theoretically obvious.

Research Questions Versus Interview Questions. Preparing an interview protocol will also help you translate your etic research questions into emic interview ones and start map- ping your data theoretically. When you prepare your proto- col, think about the concepts you are trying to evoke with each question. Then speculate on different scenarios to pre- dict what kind of probing opportunities might arise, so that you are ready—but also know that most probes will be emergent and spontaneous and will be contingent on how well you listen to your participant. Listening carefully while simultaneously deconstructing the answers as your participants speak is a skill you will build gradually, and in the meantime it doesn’t hurt to prepare a few possible probes before each interview.

In the beginning of your project, you will have a sparse set of concepts, as you are still discerning abstractions from rich subjective narratives. Though there is a strong connection between what your participants express in their answers and what you will write as a researcher—the basis of empiricism—there will also be an interpretive leap that you need to carefully build and navigate through theoreti- cal abstraction. Simply put, even though they are

942 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

connected, research questions and interview questions are not the same thing. Research questions are etic abstractions that map relationships between concepts, whereas inter- view questions seek to understand lay and subjective artic- ulations of these concepts. Your duty as a scholar is to make the connection between the two as rigorously as pos- sible, using people’s subjective narratives to thoroughly and ethically substantiate your theoretical claims. However, ordinary life worlds do not always come in clearly defined stories around academic constructs. They come in metaphors, everyday vernacular, and subjective articulations, albeit shaped by shared sociohistoric condi- tions (Thompson et al. 1994). Your participants’ narratives will be messy, full of contradictions and contestations. Good, this is exactly what you want. If you are expecting interviewees to give straightforward and concise answers, perhaps you are not asking a very novel research question. Your job as a researcher is to find a previously untold theo- retical story in this mess and uncover what is not immedi- ately obvious. To do this you need to continuously navigate between emic and etic.

When preparing your protocol, reflect on whether you are really asking what you want to ask about. Sometimes, when you are trying to understand macro- or meso-level issues such as market-level processes, you will be tempted to ask your participants questions about them. But lay theo- ries about the social world articulated by your participants do not necessarily correspond to what is really going on at the macro or meso level. They simply represent how your participants see the world around them. Unless you are studying “consumer lay theories on issue X,” this type of questioning might not be helpful. For non-individual-level questions, you might require a different level of analysis, and different type of data. For example, when I interviewed “indie consumers,” my participants repeatedly stated that they were mistaken for hipsters (Arsel and Thompson 2010). This led me to question the connection between in- die and hipster to better understand why (a) my participants found this problematic, (b) they felt misrecognized as hip- sters when they did not self-identify. Although asking them, “How do you think indie became associated with hipster?” could reveal their own hypothesis, and might bring some preliminary insights, a more appropriate way to tackle this question is not seeking an answer from consum- ers themselves, but looking at market-level data (such as historicizing the term “hipster”). In sum, you will need to carefully think about what subjective experiences your par- ticipants can reveal—and what they can’t—and how answers to interview questions relate to your research questions.

Let’s take two articles to illustrate the difference be- tween a research question and an interview question fur- ther. If you are studying veiling as stigma, as did Sandıkcı and Ger (2010), you will probably want to ask your partici- pants about the moments they experienced this stigma.

However, you cannot always expect your participants to use this sociological term, or unambiguously identify or ar- ticulate this complex experience. Instead, you will need to ask about their experiences wearing a veil in public, and with luck you might discover, as Sandıkcı and Ger did, that the “gaze and judgments of uncovered women” (19) are an important part of your participants’ everyday experiences. If you had asked directly about whether they felt stigma or not, you might have faced one of three undesirable con- sequences. First, your participants might not have directly identified or expressed their experiences through this word, and therefore might have been confused about what it meant. Second, you might have gotten a binary “yes or no” answer, with an assumption that whatever you meant by stigma was what they subjectively experienced. Third, it might have turned out that what you were studying was not about stigma at all—but since you asked about stigma, your informants obligingly talked about it.

My second example is Weinberger (2015), who studied how dominant consumption rituals affect interpersonal identity goals. When conducting her interviews, she did not ask her informants, “Can you tell me how dominant consumption rituals affect your interpersonal identity goals?” because that is not how people talk about their ev- eryday experiences (unless they are social scientists). She most probably said something along the lines of, “How do you feel when everyone around you is celebrating Christmas, and you are not?” By asking the question this way, the researcher can first empirically substantiate what she means by a dominant consumption ritual (Christmas as it is experienced by a noncelebrant) and then further in- quire about interpersonal identity goals. If you impose the concept of “dominant consumption rituals” on your inter- viewees, you will most likely capture your own preconcep- tions with interview answers, building a study around a tautology (McCracken 1988).

Interview Protocol as an Ever-Evolving Document. An interview protocol is not a survey instrument that you need to follow religiously and consistently across participants, nor it is a fixed one. During semistructured interviews, you will probably not follow the planned order of questions be- cause the conversation will take on a life of its own. You will also change the wording of questions to tailor them to each participant, and you will add new probes as you listen to their answers (Berg and Lune 2012). Treat your ques- tions as checkpoints, or prompts (McCracken 1988) within the flow of the dialogue, but try to have a natural conversa- tion. Let your participant steer the interview a bit to the spaces outside your worldview. This is particularly impor- tant during the first few interviews where the research proj- ect starts taking shape. Allowing your participants to freely talk will allow you to break down your preconceptions and revise your theorizations. Unless you are particularly pressed for time, don’t be afraid if they digress a bit.

ARSEL 943

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

Sometimes more interesting stuff comes up where you least expect it. For example, when I interviewed participants for my dissertation with the intention of learning about the meaning of indie, my participants spontaneously started to talk about why they are not hipsters instead (Arsel and Thompson 2010). This prompted me to rethink my research questions around market myths and reposition my work.

This is particularly important for novice interviewers, but if you are rusty, or if your sample is one that is hard to reach (i.e., finding participants is difficult due to the nature of the particular subject matter you are studying), test your ques- tions before interviewing your first real participant. Pick a friend who is not familiar with consumer research (or two, if the first one doesn’t go well) to test your questions and their wording. Researchers tend to use overly theoretical wording in interviews, and people without the same level of expertise often get puzzled about the funny way we talk about every- day things. Ask for your friends’ feedback to see if you were clear and whether you can improve your questions.

Knowing Your Participants. While I don’t intend to get into a discussion on sampling or how to recruit partic- ipants,

2 I would like to add a note regarding the way your

sample affects your interview design. Sampling in qualita- tive research is rarely random (for an exception, see Holt [1998], who used a random sample from the phonebook). As a researcher, you purposefully seek out specific people based on the topic of your inquiry. Are you interested in a community of consumption? In hobbyists? In people with a specific social class or gender position? This is important because these parameters will eventually determine the vo- cabulary of your questions as well as the questions them- selves. Are you familiar with the context you are investigating? If not, getting familiar with the ethos and language of the people will be necessary. For example, for his master’s thesis on communities of co-creation, Martineau spent some time familiarizing himself with the Threadless community, the online forums, and the way co- creation worked in this community (Martineau and Arsel 2017). He first gained an understanding of the task and the vocabulary used in the community before talking to partici- pants, so that he wouldn’t act like a Martian anthropologist (Belk et al. 2013) asking the obvious. Understanding the culture of your context will also help you to determine the proper dress, tone, and vocabulary to manage power distance.

Interviewing Other Market Actors and Multi-Actor Studies. In addition to consumers, you might decide to interview other market actors. These could, for example,

be managers, gatekeepers, experts, media people, pro- ducers, or service providers. You might need to interview them to understand the backstage of a consumer story, un- pack market-level processes, or contextualize your con- sumer data. For example, if your primary research revolves around Harley-Davidson owners (Schouten and McAlexander 1995), you might wish to triangulate your data with the marketers’ perspective by interviewing store managers and staff at the corporate office of the company. Some of these employees might be busy people with time and access restrictions (Harvey 2011; Thomas 1993). You do not need their life histories, and your questions can be more purposeful and pragmatic, aiming to fill in the gaps that have emerged from the consumer data. It is fine if these interviews are shorter than typical consumer interviews.

If you have a multi-actor study, you might also need to determine if you need individual protocols for specific sub- groups. Some studies necessitate that you interview both marketers and consumers (Dion and Borraz 2017). This type of study design will warrant two types of protocols, as you are investigating two interrelated facets of the same phenomenon, experienced by distinctly different actors. In other cases, where the unit of analysis is not the indi- vidual but a group, such as the family (Epp and Price 2010; Epp and Velagaleti 2014), a single protocol is likely adequate. However, when conducting studies where groups are the unit of analysis, researchers need to be attuned to the relational nature of identities that are in play, both concerning the topic of study (Epp and Price 2008) and during the interview process where different members of the group may be present, co-constructing a relational narrative.

Ethics in Asking Questions. In some cases, you might be interviewing people who are marginalized, disenfran- chised, or simply in a relatively lower power position than you. As Spivak (1988) discusses, even attempting to under- stand or represent another person’s experience is an act of power and when there is a power differential between the researcher and the participant, the effects can be oppres- sive. Consider in advance the potential power dynamics of your interviews and how knowledge construction and sub- jectification could be a form of dominance and control (Kvale 2006). Be aware of the fact that just because you think you can freely talk about something, it does not mean others can also express themselves on the same topic without any concerns or reservations. Understand that what could be an innocent question to you might con- vey symbolic violence (Bourdieu and Wacquant 1992) or epistemic violence (Spivak 1988) toward your inter- viewee. Publishing answers to your questions without carefully safeguarding your participants’ identities also can put the person in a vulnerable situation. If you are not sure if you might be harming a participant by asking a

2 Recruitment is always a challenge. As much as this will sound not very helpful, I admit there are really no hard and fast rules for making contact with potential participants other than being persistent and en- suring that you follow your institution’s ethical guidelines. Use social media, personal connections, and community organizations, and do not get discouraged if you get turned down a few times.

944 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

particular question, or printing a particular answer in your manuscript, check with your institutional ethics review board.

Here, I acknowledge that there is a great deal of cross- cultural variation in how much academic institutions undertake responsibility in governing research ethics and how ethical principles are enforced. For example, most European, Asian, and Latin American universities will assume that their employees bear their own respon- sibility in protecting the rights and well-being of re- search participants, whereas universities in Australia, Canada, New Zealand, and the United States institution- ally regulate all research involving human participants, some even demanding to approve your interview ques- tions before you embark on data collection. Beyond requirements laid out by local laws and institutional pol- icies, I suggest you never forget two universal tenets: make sure your participants understand that their answers to your questions will be used for research (par- ticularly important if you are dealing with minors, peo- ple with diminished capacities, or those with limited literacy skills), and do no harm. There are, however, exceptions for doing covert studies (Berg and Lune 2012). If you do not have an institutional ethics review board to consult, you might want to ask your peers for a second opinion.

Step 3: Conduct the Interview

Beginning the Interview and Building Rapport. There is really no magic trick I can reveal here to alleviate your fears about whether interviewees will trust you with their intimate thoughts and feelings. But here is what I recom- mend for increasing the chances of building good rapport: unless you have time constraints, I suggest you establish rapport before turning on the tape recorder. Clarify roles by explaining the interview procedure and what you expect from your participant. Explain the study, tell a bit about yourself, and say why you are interested in this project. Make yourself human. Most of us pick our topics not be- cause there are material incentives to study them, but be- cause we are interested in learning and saying something about these subjects. Sharing your personal story regarding the project with your participants will be transformative in terms of building a trustworthy relationship. This is also the point where you will need to decide how much to re- veal to your participants. Qualitative interviewers do not need to provide a cover story so that they can get a con- trolled and unprimed effect in an experiment, and, conse- quently, neither do they need to do debriefings. But you will want to refrain from talking too much about your preconceptions. If you tell your participants in advance about your emergent findings or your speculations, their whole narrative might be shaped to either support or negate this explanation because they might try to help you, or

show how sovereign they are. Give them a broad idea about your research that focuses on the context rather than theory (something like “I am trying to understand your ex- perience with [research context]”), and let your concepts naturally come out during the interview instead. To ensure that you are not leading your participants, the rule of thumb is asking questions from broad to specific and not invoking key concepts from your end. This will allow participants to spontaneously emphasize the most important issues first themselves and will reduce your chances of writing up a tautology that simply reaffirms your own assumptions (McCracken 1988).

Probing. Probes are the most important type of ques- tion in an interview, and the most difficult to master. While your participant is answering your question, you should be carefully listening to the answer to identify opportunities to dig deeper. Sometimes you will hear inconsistencies in a participant’s narrative that puzzle you. Do not try to justify these inconsistencies, or theorize about them, or judge your participant; instead, work with the interviewee to delve deeper into these inconsistencies. Like most people, your participants live in a complex world of contradictions, so they will also be interested in reflecting on these incon- sistencies. Yet, however brilliant and insightful your probe is, do not interrupt your participants. Let them speak as much as they like, and make a mental note on what to fol- low up on later. If there is a natural break in the conversa- tion, reflect back on the original answer and add your probe. If you do not trust yourself with remembering your next probe, have a piece of paper in hand and inconspicu- ously jot down your next questions. It is perfectly fine to revisit your questions during an interview; in fact, it is de- sirable. Circling back to earlier topics is a way to gain depth and fill gaps (Belk et al. 2013).

When probing, avoid questions that will close off the conversation. These are questions that can be answered with a simple yes or no. You are interested in “yes,” but you are also interested in the how, when, and why of this yes, and asking, “Do you do X?” will not yield the richness and complexity you seek. Thompson et al. (1989) argue that you should not ask “Why?” questions because it causes participants to start theorizing about their own expe- riences. But I think the issue of “Why?” is more complex than a blanket ban conveys. We should instead consider the tone, context, and delivery of how we ask “Why?” for a couple of reasons. First, asking participants to reflect on their own behavior, thereby giving them agency and power in constructing and shaping their own narratives, is impor- tant. As Spiggle (1994) says, we are not trying to read minds, but rather we should be translating the participants’ experiences by “drawing upon our stock of previously grasped meanings” (499). We, as researchers, do not hold the key to the wisdom that our participants lack, nor are we smarter than them in that we can understand their own

ARSEL 945

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

behavior better than themselves. We might have access to a better repertoire of interpretive tools and texts than an av- erage nonacademic due to our specialized training, but our participants own the right to their own narratives and explanations. Second, I have yet to find a clearer, more di- rect, and more user-friendly question than the variant of asking “Why” that entails empathically following up on some statement with “Hmmm. . .can you tell me a bit more about why you say that?” Here, “Can you tell me a bit more” is the key, as well as your delivery and body lan- guage. There is a difference between showing curiosity about why people are doing a certain thing, and dryly and antagonistically asking them to justify their behavior from a position of power.

As important as empathy is, you can show too much. Novice interviewers rush to complete their participants’ sentences, whether because they want to actively listen since it is a social convention, or because they take naı̈ve pleasure in thinking that they finally solved their theoretical puzzle. Imagine you have the following exchange with an interviewee in a study on homeyness in third places:

Interviewee: When I entered the coffee shop, it felt good.

You: So, you felt at home?

This interchange, even when due to an unnecessary in- terviewer eagerness, ruins your data. It is also ethically questionable because your response forces into your data the concept of “home,” which should have emerged natu- rally through your participant’s own volition. A better way to probe and open up the conversation would be, “So, tell me more about what was good. What do you mean by ‘good’?” You then hope that the notion of home, perhaps through metaphors, will come up in the conversa- tion. If it doesn’t, it might be time to rethink your concep- tualization or figure out why this particular person does not invoke home, despite existing theories about third places.

At other times, you will try too hard to be on your partic- ipant’s side because you want to maintain rapport. You can still make empathic statements without reflexively agree- ing with your participants or putting words into their mouths. For example, instead of “I completely understand how you feel when your favorite café closes” (which closes down the dialogue, and also is presumptuous), say some- thing like, “You said you were disappointed with the clo- sure of your favorite café. Can you tell me more about this disappointment?”

Also note that during interviews, there will be silences (unless you are living in a Gilmore Girls episode). You can deal with a silence by seeking to understand its meaning, assessing whether it is just a pause in the conversation for a little break, or whether your participant is bored, frustrated, in pain, disinterested, tired, confused, or upset. Silence can also be effectively used as a probe; waiting a few seconds

while looking confused or curious might encourage your participants to elaborate (Belk et al. 2013; Berg and Lune 2012). Be aware of body language—yours and theirs. Note, also, that silence is data (Poland and Pederson 1998). Sometimes silences are opportunities to redirect the con- versation from the awkward or inappropriate to a new and more fruitful direction. Once in a while, you will also meet a participant who, despite all your efforts, is not opening up. Unless you think there is a particular problem with the way you are conducting the interview (such as your im- plied power positionality with respect to your participant, or your inadvertent use of words that intimidate, disturb, or offend them), let it go. Some people find it uncomfortable or unnecessary to express themselves in long, quotable sen- tences that academics love. Accept this, and move on to the next participant.

Lastly, end every interview with “Is there anything I have not asked regarding your experiences that you’d like to tell me?” This will give the participant an opportunity to raise issues that you might not have considered yourself. Some of my own interviews took a new turn after this question and added upward of an hour of unforeseen dia- logue that deepened and enriched my understanding of my participants’ experiences.

Step 4: Iterate

Once you have finished interviewing your participant, it is time to reflect back on your experience before mov- ing on to the next interview. Did all questions work as expected, or did you face a disinterested person who didn’t care about your questions or didn’t know what you were trying to get at? Maybe it is time to rethink your questions or their wording. Is there anything unusual, un- expected, or contradictory in your participant’s narra- tives? This is nothing to be alarmed about; in fact, most interesting findings start to emerge when your partici- pants are not telling you what you thought they would. Consider this an opportunity to revise your interpretations and rethink your question. Position your interview in your whole data set, and reflect on how all the puzzle pieces fit now that you have one more. In other words, analyze your new data in light of the existing data set. Repeat until you have an analytically robust, empirically rich, and theoreti- cally original story.

CONCLUSION

My goal for this tutorial was to walk you through one component of the research process, focusing on inter- view design and execution. While I tried to isolate fun- damentals concerning these two skills for pedagogic purposes, remember that the practice of doing qualitative research is not compartmentalized, nor is it sequential. Conceptualization, research design, data collection, data

946 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

analysis, and theory building are inseparable components

that continuously affect each other. At this point, you

might also wonder how you are going to analyze all this

data. Unfortunately, my mandate ends here. For that I rec-

ommend you consult two books I frequently refer to in this

tutorial and use in my teaching: one by Belk et al. (2013)

and the other by Berg and Lune (2012). These two books

are not the sole authorities, of course, but they are both ex-

cellent starting points and they include extensive bibliogra-

phies for you to discover more resources.

REFERENCES

Alvesson, Mats (2003), “Beyond Neopositivists, Romantics, and Localists: A Reflexive Approach to Interviews in Organizational Research,” Academy of Management Review, 28 (1), 13–33.

Arsel, Zeynep and Jonathan Bean (2013), “Taste Regimes and Market-Mediated Practice,” Journal of Consumer Research, 39 (5), 899–917.

Arsel, Zeynep and Craig J. Thompson (2010), “Demythologizing Consumption Practices: How Consumers Protect Their Field- Dependent Identity Investments from Devaluing Marketplace Myths,” Journal of Consumer Research, 37 (5), 791–806.

Belk, Russell, Eileen Fischer, and Robert V. Kozinets (2013), Qualitative Consumer and Marketing Research, London: Sage.

Berg, Bruce L. and Howard Lune (2012), Qualitative Research Methods for the Social Sciences, Boston: Pearson.

Botti, Simona, Kristina Orfali, and Sheena S. Iyengar (2009), “Tragic Choices: Autonomy and Emotional Responses to Medical Decisions,” Journal of Consumer Research, 36 (3), 337–52.

Bourdieu, Pierre and Loı̈c J. D. Wacquant (1992), An Invitation to Reflexive Sociology, Chicago: University of Chicago Press.

Brinkmann, Svend (2007), “Could Interviews Be Epistemic?” Qualitative Inquiry, 13 (8), 1116–38.

Burawoy, Michael (1998), “The Extended Case Method,” Sociological Theory, 16 (1), 4–33.

Dabholkar, Pratibha A., Dayle I. Thorpe, and Joseph O. Rentz (1996), “A Measure of Service Quality for Retail Stores: Scale Development and Validation,” Journal of the Academy of Marketing Science, 24 (1), 3.

Dion, Delphine and Stéphane Borraz (2017), “Managing Status: How Luxury Brands Shape Class Subjectivities in the Service Encounter,” Journal of Marketing, forthcoming.

Epp, Amber M. and Linda L. Price (2008), “Family Identity: A Framework of Identity Interplay in Consumption Practices,” Journal of Consumer Research, 35 (1), 50–70.

——— (2010), “The Storied Life of Singularized Objects: Forces of Agency and Network Transformation,” Journal of Consumer Research, 36 (5), 820–37.

Epp, Amber M. and Sunaina R. Velagaleti (2014), “Outsourcing Parenthood? How Families Manage Care Assemblages Using Paid Commercial Services,” Journal of Consumer Research, 41 (4), 911–35.

Gluckman, Max (1958), “Analysis of a Social Situation in Modern Zululand,” Bantu Studies 14 (1), 1–30.

Harvey, William S. (2011), “Strategies for Conducting Elite Interviews,” Qualitative Research, 11 (4), 431–41.

Holt, Douglas B. (1998), “Does Cultural Capital Structure American Consumption?” Journal of Consumer Research, 25 (1), 1–25.

Humphreys, Ashlee and Rebecca Jen-Hui Wang (forthcoming), “Automated Text Analysis for Consumer Research,” Journal of Consumer Research.

Kvale, Steinar (2006), “Dominance through Interviews and Dialogues,” Qualitative Inquiry, 12 (3), 480–500.

Maciel, Andre F. and Melanie Wallendorf (2016), “Taste Engineering: An Extended Consumer Model of Cultural Competence Constitution,” Journal of Consumer Research, 43 (5), 726–46.

Martineau, Eric and Zeynep Arsel (2017), “Managing Communities of Co-Creation around Consumer Engagement Styles,” Journal of the Association for Consumer Research, 2 (2), 179–95.

McCracken, Grant (1988), The Long Interview (Qualitative Research Methods, Vol. 13), Newbury Park: Sage.

Moisander, Johanna, Anu Valtonen, and Heidi Hirsto (2009), “Personal Interviews in Cultural Consumer Research–Post- Structuralist Challenges,” Consumption Markets & Culture, 12 (4), 329–48.

Poland, Blake and Ann Pederson (1998), “Reading between the Lines: Interpreting Silences in Qualitative Research,” Qualitative Inquiry, 4 (2), 293–312.

Sandıkcı, €Ozlem and Güliz Ger (2010), “Veiling in Style: How Does a Stigmatized Practice Become Fashionable?” Journal of Consumer Research, 37 (1), 15–36.

Schouten, John W. and James H. McAlexander (1995), “Subcultures of Consumption: An Ethnography of the New Bikers,” Journal of Consumer Research, 22 (1), 43–61.

Spiggle, Susan (1994), “Analysis and Interpretation of Qualitative Data in Consumer Research,” Journal of Consumer Research, 21 (3), 491–503.

Spivak, Gayatri Chakravorty (1988), “Can the Subaltern Speak?” Can the Subaltern Speak? Reflections on the History of an Idea, ed. Rosalind C. Morris, New York: Columbia University Press, 21–78.

Thomas, Robert J. (1993), “Interviewing Important People in Big Companies,” Journal of Contemporary Ethnography, 22 (1), 80–96.

Thompson, Craig J. (1997), “Interpreting Consumers: A Hermeneutical Framework for Deriving Marketing Insights from the Texts of Consumers’ Consumption Stories,” Journal of Marketing Research, 34 (4), 438–55.

Thompson, Craig J., William B. Locander, and Howard R. Pollio (1989), “Putting Consumer Experience Back into Consumer Research: The Philosophy and Method of Existential- Phenomenology,” Journal of Consumer Research, 16 (2), 133–46.

Thompson, Craig J., Howard R. Pollio, and William B. Locander (1994), “The Spoken and the Unspoken: A Hermeneutic Approach to Understanding the Cultural Viewpoints That Underlie Consumers’ Expressed Meanings,” Journal of Consumer Research, 21 (3), 432–52.

Thompson, Craig J. and Tuba €Ustüner (2015), “Women Skating on the Edge: Marketplace Performances as Ideological Edgework,” Journal of Consumer Research, 42 (2), 235–65.

€Ustüner, Tuba and Douglas B. Holt (2010), “Toward a Theory of Status Consumption in Less Industrialized Countries,” Journal of Consumer Research, 37 (1), 37–56.

ARSEL 947

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

Wallendorf, Melanie and Merrie Brucks (1993), “Introspection in Consumer Research: Implementation and Implications,” Journal of Consumer Research, 20 (3), 339–59.

Weinberger, Michelle F. (2015), “Dominant Consumption Rituals and Intragroup Boundary Work: How Non-Celebrants Manage Conflicting Relational and Identity Goals,” Journal of Consumer Research, 42 (3), 378–400.

Wilk, Richard R. (2001), “The Impossibility and Necessity of Re- Inquiry: Finding Middle Ground in Social Science,” Journal of Consumer Research, 28 (2), 308–12.

Woermann, Niklas and Joonas Rokka (2015), “Timeflow: How Consumption Practices Shape Consumers’ Temporal Experiences,” Journal of Consumer Research, 41 (6), 1486–508.

948 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/article-abstract/44/4/939/4104525 by Adam Ellsworth, Adam Ellsworth on 14 November 2017

Copyright of Journal of Consumer Research is the property of Oxford University Press / USA and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.