What would Malinowski do? Rethinking fieldwork in social media.

Annette Markham

Sep 4, 2013

[this is a work in progress for publication later this year in Qualitative Communication Research, so it will change frequently.  Comments and suggestions welcome! Also note, i have not included the works cited list yet, but eventually it will be there.]

In qualitative studies fieldwork is often used as a method of engaging with the phenomenon to gather information/data or to analyze practices in situ.  While fieldwork is the foundation of robust anthropological inquiry in physical settings, its practical methods do not fit comfortably in digital contexts.  For many researchers the activities of fieldwork must be so radically adjusted, they hardly resemble fieldwork anymore.  How does one conduct ‘participant observation’ of Twitter?  What counts as observation of a blog, and how closely should this practice resemble observation of cultural practices in traditional ethnographic environments?  When identities and cultural formations are located in or made of information flows through global networks, where do we demarcate the boundaries of ‘the field’?  In such global networks, what strategies do we use to conduct interviews?

This essay is meant to offer some provocations in the non-anthropology world of research where people do what they call (but is not really [foot]A caveat: This is a gross oversimplification of ethnography as well as a seeming generalization of ethnographic works. My audiences are people who study digital contexts, want to get closer to the people or contexts they’re studying, but know nothing or little about ethnography. They pick up a textbook or two, do some interviews and participant observation, and then call it ethnography because it has cache or because they don’t know what else to call it, or because their well-intentioned colleagues, who also don’t know much about ethnography, insist that anything that involves fieldwork is ethnography. My audiences are also those people who use fieldwork methods that don’t fit the contexts or phenomena they’re trying to study.  Rather than going back to the premises of fieldwork methods to develop innovative practices, they find themselves trying to do participant observation of Twitter or interviewing in situations where this mode of information gathering may be unwarranted or unnecessary or even distracting from other, more appropriate methods.  Arguably, the fieldwork and the findings both suffer (not to mention the researcher–it’s frustrating to try to jam a square peg into a round hole). My purpose is not to help people adopt a neo or post ethnographic frame, but to recognize that ‘fieldwork’ can be separate from ethnography.  If this argument is accepted, then the associated methods can be adapted to create better resonance with contemporary social media contexts, which would result in stronger research products. End caveat.[/foot]) ethnography and, more specifically, engage in ‘fieldwork’ of networked, digital, social media contexts where the methods of classic ethnography don’t fit very well.  I want to veer distinctly away from conversations about how to do ethnography[foot]And here, I am thinking of those ethnographers working to stress the importance of a strong ethnographic stance, such as the recent collaborative book by Tom Boellstorff, Bonni Nardi, Celia Pearce, and T.L. Taylor (2012).[/foot] toward a conversation about how we might think about fieldwork, whether or not it is ethnographic.

As an itinerant professor traveling through various European and Nordic countries and facilitating seminars and workshops, I frequently encounter situations where researchers are frustrated by the ill fit between traditional tools of ethnography and the social contexts they are studying.  One repeated phrase has prompted my thinking about this issue of fieldwork in social media: “Now, I need to go do the fieldwork part of the study.”  Sometimes different words are substituted for ‘fieldwork,’ like: “Now, I need to go do the interviews,” or “But I still need to do the focus group interview phase,” or “Well, I can’t really use their blogs as part of my study until I get written informed consent.”

On many occasions, I have blurted in response, “Why?”  When they have already engaged the phenomenon on a close level for sustained periods and produced more than enough rich data from methods A, B, and C, I can’t help but ask why they need to do X and Y?  It’s a genuine question, because I want to know why they feel these other techniques or procedures are important.  Quite frequently, the answer is something like: “Because X is the next step.”

In response I often protest, “But you have already elicited so much information from involved individuals about how they feel about the phenomenon.  Why would you need to conduct face-to-face interviews with them?” or “Why doesn’t 50 hours of in-game recorded conversations with your guild about their socialization process as a community count as a focus group?” “Why do you need to talk in person with underage girls cutting themselves and chatting about it in a virtual space?”  Most of the time these scholars end up giving good reasons to engage in their procedures and share my concerns about the utility or benefit of others.  Many of them are grappling with multiple competing challenges: As a PhD student, they want to do the right thing, scholastically speaking.  As a part of larger, possibly taxpayer-funded projects, they must answer to diverse stakeholders who may have typical or orthodox expectations of ‘what counts’ as legitimate qualitative methods.  As scholars studying complex, digital, mobile cultural contexts characterized by physical as well as digitally-mediated forms of communication and interaction, they may not have a ready arsenal of innovative methods to improvise effectively.  Sometimes the best we can do in such a situation is to fall back on traditional methods.

This is not an uncommon situation, but it highlights the mismatch of methods to context as well as the potential ill fit between a priori assumptions about what the field will entail, methodological expectations of administrative stakeholders, and emergent concerns arising from the situation and complexity of the context.  For many young scholars there is little room for flexible adaptation.  It is easy to feel trapped by method.

The issue goes well beyond a simple choice of method.  It requires critical interrogation of the situation within which one finds oneself doing inquiry[foot]I use the term ‘situation’ in the sense Donna Haraway does, focusing on the situated nature of knowledge and science. There’s a particular geneology for the methods undergirding traditional fieldwork and, if we use theoretical insight from Foucault, Bateson, Garfinkel, or Goffman, we should be able to look back or dig deep to see the structures, ecologies, habits and frames that constrain and enable how we think about and enact something called fieldwork, within our particuarlar disciplines and drawing on particular anthropological or sociological ethnographic traditions. (see Haraway, 1991, chapter 9, p. 10, available from: http://science.consumercide.com/haraway_sit-knowl.html[/foot]. It also requires a somewhat broader exploration of the political and cultural conditions surrounding and influencing our inquiry with the aim of revealing some of the ways our habits and traditions discipline us at deep structures of discourse, creating powerful, yet seemingly neutral and natural norms for seeing the world and conveying this knowledge to others.  And how this all prevents us from using our creative, interpretive authority to “see.” Things. Differently.

In addition to exploring the larger situation within which research occurs, we can focus more directly on a frustration in social media, internet, and digital culture research: the difficulty of transferring a method from context A to context B, which in arenas of digital inquiry generally but not always coincides with offline to online.  To make such a shift requires the analogy to be close enough to make the transfer fit.  It fits, sometimes.  And sometimes not so well.  I suggest we’re asking the wrong questions.  Instead of asking, “How can we observe in Blogs” or “How can we interview in Twitter?” we might more productively ask:

“Why should I observe? What is an ethnographer supposed to get from observation?”

“Why should I interview? What am I supposed to get from an interview that I couldn’t get otherwise?”

These questions help provide the grounds for a more fundamental question about ethnographically informed inquiry: If we were to collectively forget anthropology ever existed, how might we invent fieldwork from the ground up, today? How did methods for studying human beings in social interactions become more than habits? It remains critical to continue considering the underlying premises of fieldwork, so that we can more thoughtfully adapt these methods to social media contexts. Without the baggage of thinking I ‘ought to’ interview, or participate or observe, one can more fully embrace what needs to be done in situ.  Good researchers do this anyway. Posing this thought experiment somewhat differently:

What if in the early 1900s, anthropologists had the Internet?

What would Malinowski do?

To begin this thought experiment is to question the working assumptions embedded in the use of terms associated with fieldwork[foot]At a different level, we could look also at what the ethnographer focused on, and we could explore how ethnography came to study kinship, rites, rituals, religious beliefs, medicine or medical beliefs and practices; or study movement and migration patterns, cultural memory, legends and myths. This is somewhat beyond the scope of what I address in this particular essay, since it’s more about the aims of anthropology than the practices and premises of fieldwork.  Still, it’s a useful arena to deconstruct and then reconstruct in a social media sense, to think about how this might get accomplished or what other categories of meaning might be relevant/salient in digital/networked cultural formations.[/foot]. In what follows, I only scratch the surface, hoping that these provocations might prompt continued conversations about the possibilities for what a robust fieldwork of digitally-saturated social contexts might incorporate.

Fieldwork, as a specific intervention practice, is work done in the field, close to people[foot]I don’t mean to equate fieldwork with ethnography—arguably, one’s notion of ethnography oscillates between the ontological, epistemological, and methodological. There are other ways of thinking about ethnography more as a worldview than a set of methodological assumptions and practices.  Here, I focus specifically on those situations when researchers are using ethnographic tools to study cultural practices.[/foot]. In traditional anthropology the field comprised a geographical, physical boundary, defined by the object of study, which was typically a tribe or tribal unit.  The researcher entered the field physically with the expressed aim of getting closer to the unknown ‘Other.’  The process is embodied and active, requiring both mundane and significant adjustments on the part of the ethnographer, in terms of dress, manner, language, and other elements of embodiment that might convey a message.  Close attention to these details would be an initial step in successfully navigating through the first, most obvious set of boundaries.  Proximal access depended on the anthropologist’s ability to move past official and unofficial guards who would determine the threat level of the intruder.  The way might be paved by previous researchers, but more often with trinkets or gifts, which were perceived to supplement the cultural members’ natural curiosity and help the researcher gain access.  For those being studied, whether located in the depths of the Amazon rainforest or working class neighborhoods of Chicago, there generally would be a sense of security born from their greater numbers and being on their own turf.

In mediated contexts, where is the field? How should we enter it? What might be the equivalent of the boundary crossing that traditional anthropologists faced? Are these created by particular stakeholders, negotiated by groups, linked to geographic boundaries? Are the boundaries discursive, a verisimilitude of physical, or literal, as when we might encounter firewalls, entrance requirements, paywalls, and so forth?  What does it mean to be closer to the people one is studying?[foot]For extensive early conceptualization and problematization of these issues in mediated contexts, see, e.g., Markham, 1998; Hine, 2000; Kendall, 2002; Miller & Slater, 2001; and more recently, John Postill’s work on social media ethnography (20??)[/foot] Marcus’s notions of multi-sited ethnography (1998) and para-sites (2000) not only help us build more complex notion of ‘fields,’ but also prompt us to ask: How helpful is a term that requires so much effort to redefine so it fits the contemporary context?  Again, shifting focus to critically explore some of the historical reasons the term became so important can help us determine what is salient (or not) to our own work.  In this way we might ask, Why was the term ‘field’ important to anthropologists like Malinowski and how did the term ‘boundaries’ come to hold such importance?  What might be the value–in digitally-saturated globally networked tangles of communication and cultural interaction– of reconsidering what ‘boundary’ or ‘field’ should mean, on a case-by-case basis?  Certainly, a more flexible notion of the field is one that allows us actually to stop thinking about it as an object, place, or whole – and start thinking more about movement, flow and process.

Participant observation. This active positioning of the investigator is the hallmark of anthropological and/or ethnographic inquiry.  In fact, participant observation is/was often interchangeable with the term ‘fieldwork’ in traditional anthropological training.  Breaking apart the terms, let’s start with observation, again posing these root level questions: Why did classical anthropologists pay attention to what is being said around them?  How did they do it?  What did they hope to gain from this practice?  These questions may seem a bit silly on the surface, but when we break the practice down even further, it begins to offer more reflexive meanings.  Collecting ‘naturally occurring’ discourse was accomplished traditionally by listening and then later recalling in writing what was said, when, and to whom.  In social media, we need to ask not only how to collect this but ask more basically: What is naturally occurring discourse?  What forms does it take?  If we broaden the idea of “discourse” to mean anything that can be read as having communicative meaning, what might it include?  Why do we need to know how people talk about things?  Attempting to transfer participant observation from physical contexts to a mediated social space like Twitter simply doesn’t work[foot]for me, at least[/foot].

In social media, observation often takes the form of archiving.  In digital contexts, there’s a temptation to collect and archive everything, just in case.  This fire hose of data might indeed hold much value.  But in the past twenty years this capacity has proven to be as much a weakness as a strength: This sort of mass archiving of everything doesn’t and can’t substitute for observation in an ethnographic sense, because observation is for most ethnographers interwoven with its sister term: participation.  And as one participates in culturally-specific practices, one learns more about what is relevant and what isn’t, among other things.  So one learns through sustained engagement that it is not necessary to collect everything.  Moreover, it is impossible to capture everything that makes the culture unique, especially in a form that might be called ‘data.’  Just as global access to data doesn’t equate to global reach, large quantities of data points about a situation can never reach comprehensiveness.[foot]Nancy Baym offers an excellent argument of the literal extent to which this can (and perhaps should) be taken, noting that data gathered through social media metrics are flawed in many ways—by “algorithmic and affective skew, their partiality, the deceptive practices that may generate them, and the inherent ambiguity that arises from decontextualizing a moment of clicking from a stream of activity and turning it into a stand–alone data point” (p. 5). She concludes that at least for musicians using social media, “the data that matters most for assessing social value might not be measurable at all” (2013, p. 6).[/foot] Understanding what is necessary to attend to is an ongoing process of being situated closely to the actualities of the context. How might one make sense of what cannot be archived or sorted because it leaves no digital trace or is tangled in a dense network of unfathomably meaningful data trails?

This mentions only one of many ways that translating the procedures of observation (participant or not) from physical contexts to online contexts frequently leads to the collection of too much information.  I know researchers who have automated the capture of game play onscreen, game play off screen, chats between players, actions and movements of avatars, infrastructural rules and norms of the game in process during the observation period, using multiple cameras to capture different angles on the physical gamer sitting in front of a console/computer. What does one do with hundreds of pages of field notes, 2,000 snapshots, hundreds of screenshots, 50 hours of video, 30 hours of onscreen video capture, 60 pages of video transcriptions, and 30 pages of onscreen video capture transcriptions?[foot]Yes, this is a real example of a single study.[/foot] This sort of material is, from a qualitative standpoint, literally unmanageable.  It also misses the essential point of ethnographically-oriented observation.  After all, understanding culture has never been a matter of collecting everything and then analyzing it later.[foot]Add to this the complication that because of the size of the collection, both management and analysis of this data is often automated or computer-aided. Even such basic management of data requires collating, coding, and/or curating, forms of organization that require a priori selection of categories. So from the moment of collection, the data is never raw. Of course, it never was, as Levi Strauss noted (date), but in an era of data-fication of everything, the illusion of having raw data is beguiling (Gitelman, 2013, Boellstorff, 2013, Markham, 2013).[/foot].

Participation. What prompted ethnographers to participate and what did they gain from this?  What was required to participate, and how did they do it?  It has been long assumed that firsthand relations with those studied provides valuable means to get closer to their understanding of their own everyday lives.  Participation allows one to get closer to the experience him or herself.  The white European ethnographer in the Amazon stood out as a foreign entity.  Participation was a level of engagement that needed to be negotiated carefully and generally over time.

In social media contexts, participation is an issue, but it doesn’t necessarily take the same form as studies in intensely physical or face-to-face contexts.  Lurking is, for the most part, a common and socially acceptable form of non-active participation.  So is participation always necessary? If so, how does it help?  Certainly, if one is joining a guild in World of Warcraft, participation seems essential. Yet if one is researching bloggers, participation may not take the form of announcing one’s presence on a blog, but rather understanding and engaging in the community of practice called ‘blogging.’  Talking with other bloggers may be necessary, of course.  But the actual length of time and level of participation that are called for?  These matters are negotiated, as with each other element of fieldwork, on a case-by-case basis.  Reflexivity can facilitate not only more logical procedures but also more thoughtful ethical stances[foot]Eva Svedmark, a Swedish researcher studying extreme behaviors online, describes in poignant detail the harm that can result from calling attention to oneself as a researcher, which flies in the face of regulatory statements regarding ethical conduct that normalize the practices of announcing one’s research intentions and obtaining informed consent in advance of a study. When her research subject realized she was being researched, things took a turn for the worse, in Svedmark’s estimation. The girl was blogging about her indecision about whether to become more anorexic or recognize her practices as an eating disorder. Eva asked her if she could study her style of blogging and immediately, the content of the blog changed. The girl expressed enthusiasm for her “new audience” and jumped off the fence, announcing her intentions to work harder to become even thinner.[/foot].

In addition to these fundamental aspects of fieldwork — the field, participant observation, observation, and participation — there are many more concepts with premises worthy of exploration within the larger rubric of fieldwork.  I briefly mention a few here:

Interviews: Why did ethnographers decide to talk to people more formally, or to engage in contrived, planned discourse?  What did they hope to achieve through this practice and what did it yield?  Why did the distinction between open-ended and closed-ended interviewing become important?  Was this distinction salient or were questions improvised as needed, depending on the situation?  In social media contexts, how might we consider different elicitation techniques?  What is elicitation, anyway?  What does elicitation yield and how might it be creatively encouraged in media that match the characteristics of the specific context?  How might we envision elicitation in a form other than question/answer?  What are other reasons for interviewing besides elicitation?  Could these goals be accomplished using other means, especially in digital contexts that are not conducive to traditional interviewing techniques?[foot]Much work has been done in this area in the past 20 years. Adaptation of interview techniques to digital media like instant messaging, game spaces, avatar-based environments, text-based social spaces like MUDS and MOOs (see, e.g., Hine, 2000; Markham, 1998; Orgad, 2005; Mann & Stewart, 2000; James & Busher, 2009) and real time audio/video based environments (see, e.g., the edited collection by Salmons, 2011)[/foot]

Fieldnotes: Why did ethnographers write fieldnotes?  What was included?  What did they do with them?  When did they write fieldnotes and if they stopped writing them, why?  What did they hope to achieve through this practice?  As Geertz among many others have noted, when we write fieldnotes, we engage in a process of cultural inscription (1973, p. 19).  This process of selecting from the context and then writing it elsewhere is one that simultaneously abstracts from lived experience and restricts attention to particular aspects that will be read textually.  What other methods are being used now, with what impact on how we conceptualize the documentation of our activities in situ? Photography and sketching were traditional methods, but how can we reimagine these activities in social media contexts?  How might we combine screenshots or video screencaptures with our own notes to build robust visual renderings of what is happening over time?  Should we save every altered version of our notes on our laptops? For that matter, should we change them at all? Should we be trying to find ways to capture the equivalent of a coffee stain on a notebook, which might trigger a flood of useful information about a particular event or moment?[foot]see Jackson’s wonderful piece on the liminal qualities of fieldnotes[/foot] What should we think about the growing use of the term ‘curating’ as a way of thinking about cataloging one’s audio, visual, and written engagements with the context? Does this frame change what the activity means?

Why did ethnographers often find it necessary to have an informant?  What did the informant provide that otherwise would remain invisible or hidden?  What were the problems with informants?  What might be a similar action to “get an informant” in digital contexts?

Why were artifacts important to ethnographers?  What did they do with these artifacts?  What constitutes an artifact in a digital context?  If artifacts leave digital traces, is it necessary to collect all of these for examination?  What would that yield and why would we need to see everything to make sense of the situation?

Why did ethnographers take pictures?  Was this a separate activity from fieldnotes, or separate from observation?  What is the role of the visual in social media contexts?

Why did ethnographers do other things that were ignored as irrelevant, marginalized, or suppressed as taboo, and therefore don’t get much play in institutionalized discussions of ethnographic methods, unless in the negative “don’t do this” form[foot]Such as: (don’t) go native, (don’t) have sexual relations, (don’t) be deceptive or covert, (don’t) share the corpus of data, (don’t) intervene or interfere in the culture, (don’t) do the wrong thing, etc.[/foot]? What is considered taboo for investigators in social media contexts?

Once we challenge taken-for-granted terminologies like this and start to ask: “What would Malinowski do?” we can reveal some of the hidden frameworks that shape and delimit our gaze, the process of gazing, and the objects of our attention.  Tracing the archeology of traditional practices of fieldwork practice returns us to key premises rather than procedures.  These may not be seen on the surface level of method, especially as this is described in research reports, but operate at a level both above and below method: At the epistemological or ontological level and also in the everyday practices of inquiry.  Exploring these premises allows for a systematic reconsideration of what might constitute flexible and adaptive fieldwork suitable for studying 21st Century networked communication practices and cultural formations.

These may seem very basic considerations to many researchers.  But for someone adopting traditional procedures associated with ‘fieldwork,’ without any training in the epistemological or methodological frameworks of anthropology or ethnography, the questions are valuable precisely because they allow the researcher to consider different ways to accomplish these same goals in digitally-saturated contexts, where it might not make sense to conduct ‘participant observation,’ or where ‘interviewing’ might not be the most effective strategy for elicitation.  I consider this to be a critical conversation for any qualitative researcher studying digitally-involved contexts.  Our fieldwork will only grow more robust when we focus on the level of practice, which can lead to a greater appreciation for what these fieldwork practices were trying to accomplish, both within the anthropological traditions and without.