1. Introduction
[1.1] In their 2014 paper "Desperately Seeking Methodology: New Directions in Fan Studies Research", Evans and Stasi argue that the field of fan studies has been remarkably resistant to discussions of methodology. As an interdisciplinary field, fan studies attracts scholars from a range of disciplinary backgrounds, including media and cultural studies, literature, literacy studies, digital humanities, library and information science, and many more (Jenkins 1992; Byrne and Fleming 2018; R. W. Black 2006; S. Black 2018; Price and Robinson 2017). Fans, too, are far from a homogenous group, and so fan studies may cover groups as diverse as sports fans, popular music fans, media fans, fans of theatre, musicals, or the news; practices ranging from attending commercial events and fan-run conventions, cosplay, or creating transformative works; and research settings as different as wrestling venues and social media platforms (Esmonde, Cooky, and Andrews 2015; Garde-Hansen 2010; Gray, Sandvoss, and Harrington 2007; Lamerichs 2015; Lothian, Busse, and Reid 2007; Phillips 2018; Morimoto and Stein 2018). By definition, there cannot be a single method that would allow us to address all the research questions we may ask about fans. Yet there are also some distinct traditions and trends within the field that make discussions about methodology both necessary and useful. Fan studies has a long tradition of the use of ethnography in multiple senses: two foundational texts in the field are ethnographies of fandom(s) in the classical anthropological sense (Bacon-Smith 1992, Jenkins 1992), but a much broader range of research relies on the use of ethnographic methods, such as participant observation and interviews, without necessarily producing ethnographies. For the subset of fan studies researchers using ethnographic methods to study media fandom—and fans who create transformative works in particular—the transition from the kind of zine- and convention-based communities documented by Bacon-Smith (1992) and Jenkins (1992) to a largely online-based community (Tosenberger 2014; Bennett 2014; Bury 2005) poses a number of methodological challenges to ethnographic approaches. The tendency of fan studies scholars to also be fans themselves is another such commonality, and has received some attention over the years (Jenkins 1992; Hills 2002; Coppa 2014). Taking Evans and Stasi's (2014) provocation as a starting point, my interest lies in the methodological challenges to the use of ethnographic methods posed by the intersection of these two factors: the theoretical and practical challenges of studying a community one is already a member of, in a digital setting. While Evans and Stasi note that "digital ethnography and reflexive autoethnography is already being done" (19), their focus is on showing the necessity of a discussion of methodology, and they only briefly sketch out what digital ethnography and reflexive autoethnography may look like in a fan studies context. In this paper, I relate this kind of work in fan studies to its roots in traditional ethnography based in anthropology and more recent work developing digital and autoethnographic methods across a range of fields. I examine methodological debates in ethnography and their relevance to the particular research setting of one type of fan studies research: the online community that we ourselves, as researchers and as fans, are also members of. I highlight two key challenges: the ethnographer's positionality in this type of setting, and the construction of the ethnographic field site, better conceptualized as a journey. Finally, using the case study of my own research into sexual consent in fan fiction on platforms such as Tumblr and Archive of Our Own (AO3), I apply the frameworks of digital and autoethnography to offer one possible methodologically robust approach to the ethnographic study of online communities that the researcher is a member of.
2. The role of the ethnographer in online settings
[2.1] Evans and Stasi (2014) rightly identify representation and the hierarchical power relations often constructed between researcher and researched as a key methodological issue in fan studies. Yet their answer to this challenge largely lies in the merging of the fan and the researcher into the single person of the acafan: "Fan studies is already doing the ontological work of the crisis of representation, in which the object of study (the fan) and the researcher merge. Fan studies therefore already has the critical capacity to implode subject/object binaries as a practice of research, which has long been a concern for feminist methodologists" (14). But fan studies, and ethnography more broadly, encounters issues of positionality well before the representation stage of research. Researchers' positionality vis-à-vis the communities we study shapes everything from the research questions we can ask to how we go about getting answers, and how we analyze and interpret those answers. Representation is only one step of many in this process, albeit an important one, and while for many fan studies scholars the fan and the researcher may be one, the fan is still one among many diverse fans rather than part of a homogenous whole. Ethnography has a long history in fan studies, starting with some of the field's foundational texts (Jenkins 1992; Bacon-Smith 1992). It is in these foundational texts that we already see some of the challenges ethnography poses for researchers throughout the research process, whether they choose to openly identify as members of the community they are studying or not. Jenkins's unequivocal embrace of fandom stands in stark contrast to Bacon-Smith's attempt to distance herself from it and present herself both to her research participants and to her readers as the ethnographer. In this section, I examine questions of the positionality of the ethnographer—particularly the ethnographer who is already a member of the community they are studying, a community that lives primarily in digital and online spaces.
[2.2] Classical ethnography as used in anthropology offers several models of the role the ethnographer plays in the community they are studying, each considering slightly different sets of variables. While these models have been developed and built on over time, they are also still highly influential as reflected by the fact that they continue to be used in anthropology and wider social science methodology discussions and textbooks (e.g., Anderson 2006). Yet they are not unproblematic in and of themselves, and are additionally complicated by the mediated nature of the online setting. Gold (1958) identifies four roles available to the ethnographer, along a continuum from complete participant to complete observer. The complete participant is a covert researcher, fully immersed and taking part in the core activities of the group; the participant-as-observer is known to the group as a researcher but is also immersed in community activity; the observer-as-participant, too, is open about their researcher identity, though their interaction with the community is much less immersive in nature; finally, the complete observer does not interact with the community or make their role as a researcher known to them. While this model is a useful starting point, it elides level of participation on the one hand and level of openness about the ethnographer's identity and objectives on the other. The presentation of the model also appears to imply that covert research is completely unproblematic—something that has been questioned over time as new ethical norms have emerged in anthropology and the wider social sciences. Adler and Adler (1994) propose a different model based predominantly on levels of membership of the group: the peripheral member will "interact closely enough with members to establish an insider's identity without participating in those activities constituting the core of group membership" (380); the active member may participate in the group's central activities and even take on responsibilities within the group but may not be fully committed to the values and goals of the group; and the complete member may be a researcher studying a community they were already a member of, or may be converted to membership, and is fully committed to the values and goals of the group. In this model, covert and overt stances are open to researchers adopting all three of these roles. Introducing yet another component for consideration, Bryman (2012) builds on Bell (1969) to identify four forms of ethnography along two axes: level of openness of the researcher about their identity (which is similar to parts of Gold's model) and level of openness of the setting. Closed settings are often formalized organizations and tend to be characterized by formal barriers to access; open settings are less formally structured and may be easier to access, though Bryman does caution that neither the distinction between open and closed settings nor the one between overt and covert research are necessarily clear-cut. Bryman's approach is much more recent than Gold's and is part of a social science research methods textbook, and he does acknowledge that there are ethical issues with covert research, but he proceeds to largely gloss over them. These three traditional ethnographic models elide some of the variables that affect the ethnographer's positionality in relation to their research setting, while also considering different variables with some overlap. They also in large part fail to account for the possible ethical issues around covert research. As I show below, mediated settings such as online communities also affect how these variables operate theoretically, practically, and ethically, and it is useful here to dismantle the models into their constituent parts and consider each separately. From this, three key factors emerge that affect how the ethnographer conceptualizes their role and their interaction with the setting: level of openness of the setting, level of openness of the researcher, and level of participation. How, then, is each of these factors affected in an online setting?
[2.3] In traditional ethnographic settings, Bryman (2012) cites the formality of organizational structures as the main factor influencing the level of openness of the setting. In online settings, however, there are additional considerations, as the level of openness is at least partially determined by the technical features of the particular platform in question and the social norms that arise around them. Some social networking sites offer privacy settings that require the ethnographer to make themselves visible to the community they are studying and be actively accepted—for instance, through a Facebook friend request. On other platforms, content is generally public and "lurking" (Baym 1993, 148) or passive reading is the norm for a significant proportion of community participants and therefore, to an extent for the researcher (Hine 2015). Yet the fact that content on social media and other internet sites is publicly accessible in theory does not mean that those posting the content—whether it is fan fiction stories on AO3 or discussion threads on Reddit—necessarily view it as such, and here the ease with which covert research may be conducted may obscure the question of ethics. An online setting may also be closed in less formal ways than the barriers to entry one may find in formal organizations: content created by many online communities may well be public and freely accessible, but only to those who know where and how to find it, which in turn has implications for the researcher's positionality vis-à-vis the community they are researching. Having insider knowledge of the community may help in both finding and accessing research-relevant communities, regardless of levels of openness, but may also raise questions as to the researcher's ethical obligation to identify themselves.
[2.4] The question of levels of participation in group activities (Adler and Adler 1994) is also complicated by the nature of the online setting and the specific platforms the community uses for interaction and conducting their day-to-day activities. In a fan community context, some group members may generate large amounts of varied content—fan fiction stories, images, commentary or meta—but many restrict themselves to only reading content and leaving the occasional kudos (a single-click indication that a user enjoyed a story) on AO3. Intermediate levels of participation are also possible, such as reblogging other users' content on Tumblr, commenting on stories on AO3, or being a very active contributor within a small and fairly isolated friendship group while passively reading content elsewhere. Which of these, then, count as full participation? While lurkers may not actively contribute to the group, they are arguably familiar with community values, norms, types and genres of content, acceptable and unacceptable behaviors, and types of interaction (Schneider, von Krogh, and Jäger 2013). In multisited online communities such as those of fan fiction fandom, individuals may choose to participate in very specific and limited ways in only one of the many sites that form the community and still self-identify—and be identified by others—as full, participating members. These different modes of engagement or participation are therefore also available to the ethnographer. Arguably, and perhaps counterintuitively, each of them can be characterized as complete membership in Adler and Adler's (1994) model, as each of them mirrors the experiences of at least some community members. As I discuss below, in an online landscape characterized by networked individualism (Wellman 2001), the ethnographer's construction of a field site from multiple fragments is an experience not dissimilar to that of community members constructing and curating their own online spaces. The right level of participation, then, may be driven by the platform and the community being studied, but also by the research question and theoretical considerations. If, for instance, the experiences of lurkers are pertinent to the research being conducted, then being a lurker at least in some contexts may in and of itself be a valuable research tool.
[2.5] While lurking therefore may still constitute full or close to full membership of the community, the ethnographer's visibility as a researcher remains a question to be addressed separately from their level of participation, both ethically and practically. From a practical standpoint, Rutter and Smith (2005) reflect on the "presence and absence" (85) of the ethnographer in online communities. Unlike a conventional field site, community members are not constantly reminded of the ethnographer through their physical presence. In communities where passive presence is an accepted form of membership, covert research is easy (though not always ethical) to conduct, while establishing an overt and visible presence may be the right ethical choice but pose a practical challenge. Hine (2015) also comments on the many choices an ethnographer is faced with when establishing their presence in one or more online field sites: what and how much information to share in a profile, how to tap into the communities and networks of interest, and whether and how to participate in community activities. The exact forms of interaction with community participants, and hence the ethnographer's visibility to participants, also varies with the technical features and social uses of the platform(s) where the ethnography is being conducted, and with the relative openness or otherwise of the setting. From an ethical standpoint, the question of covert ethnographic research in digital settings needs to be weighed carefully, and fan studies as a field has begun to grapple with this. It is important to recognize that a blanket ban on covert research privileges some groups, while potentially silencing others—for instance, in the study of antifans (Jones 2016) or racism in fannish communities (Pande 2018). Online settings, as opposed to physical ones, also offer a different kind of insight into the groups, communities, and people studied, and this also varies with the technical affordances of the specific platform(s) where a study is sited. It is possible, for instance, to gain an aggregate view of large-scale trends on a platform or in a community without infringing on the privacy or dignity of individual users. Analyzing trends in AO3 tags is an example of such ethical, covert, digital ethnographic research. Other factors to consider here are users' legitimate expectations of how public any material they post may be, how much privacy control any given platform allows users, and to what extent any material posted online can be tied back to an identity that matters to the user—whether that is their real name or another identifier they commonly go by in communities they are part of. Facebook posts, for instance, can generally be traced back to the individual's off-line identity, while Tumblr posts cannot. On the other hand, Tumblr posts can (frequently) be traced back to the identity someone is using in fannish communities, whereas posts on the popular anonymous LiveJournal/Dreamwidth community Fail-FandomAnon cannot. The impact on the individuals and communities posting such material of it being used in covert research will vary. The key ethical question to ask is who is privileged and who is silenced by choosing (or not) to conduct covert research. If a covert approach is deemed ethical, risks to group members need nonetheless to be mitigated—for instance, through quoting practices (Markham 2012, Jones 2016). The three variables affecting the ethnographer's positionality, then, are interdependent and complicated by the digital setting. They pose theoretical, practical, and ethical challenges and warrant careful consideration.
[2.6] The digital setting, however, is not the only factor that complicates the positionality of the ethnographer in much of contemporary fan studies research. The fact that many fan studies scholars start out as fans first before developing a scholarly interest in fannish communities and activities plays a significant role in how we relate to our research settings as ethnographers: fan studies as a field has a long history of fans researching other fans. Hills (2002) popularized the word acafan for this phenomenon, though both the practice and the term have a longer history than that. In one of the groundbreaking fandom ethnographies of the early 1990s, Jenkins (1992) talks of writing both as "an academic (who has access to certain theories of popular culture, certain bodies of critical and ethnographic literature) and as a fan (who has access to the particular knowledge and traditions of that community)" (5). The practice of fans studying fans is a staple of the field, sometimes overtly acknowledged (Hills 2002; Gatson and Zweerink 2004; Bennett 2014), and other times implied (Jenkins 1992; Bury 2005; Willis 2006; Freund and Fielding 2013). Yet even when this practice is named as autoethnography, the usages of the term within the field of fan studies vary widely. Jenkins's approach quoted above suggests the use of knowledges and insights generated through lived experience as a fan in shedding light on other fans. Hills's (2002) use is much more limited and centered on the ethnographer themselves, reflecting on one's own life history and experience of fandoms and putting those into a wider theoretical context. One of Hills's main methodological concerns is that in asking fans directly about their passions, fan studies scholars risk provoking a defensive stance in fans, pulling them out of the affective experience of fandom and into a need to justify themselves and their experiences. Autoethnography as Hills proposes it ostensibly allows the fan studies scholar to continue their interrogation of their own experiences beyond that justification impulse, and identify and question some of the structuring absences in their account of their own fandom. In some ways, this appears to contradict Hills's own call to "treat self and other identically, using the same theoretical terms and attributions of agency to describe both" (81). Moreover, Hills's use of autoethnography appears both rather limited and limiting, focused as it is predominantly on the question of why fans are fans of the things they are fans of. It is therefore worth examining traditions and approaches to autoethnography from outside of fan studies, and considering other kinds of research questions that this approach can be applied to, particularly in a digital setting, in order to fully grasp the potential and the pitfalls of autoethnographic methods in fan studies.
[2.7] The multiple meanings and usages of autoethnography within fan studies reflect a similar multiplicity in other fields. Reed-Danahay (1997) gives an outline of the history of autoethnography, highlighting the multiple meanings the term has across different disciplinary communities. Authors such as Van Maanen (1995), Strathern (1987), Pratt (1994), and Dyck (2000) focus predominantly on the ethnographic components of autoethnography, broadly defining it as the practice of studying one's own culture. This poses a range of challenges that more traditional ethnography does not. Dyck, for instance, discusses the anxieties and uncertainties that accompany the transition from being a social participant in a setting to "becoming anthropologically attentive to becoming an 'out' researcher" (43). Relationships and activities that have been purely personal acquire a scholarly and professional dimension, and decisions need to be made about how to treat information one may only be privy to because of personal involvement in the field rather than as a researcher, thus circling back to the question of the openness of the setting. Jenkins's (1992) fandom ethnography, for instance, would fall into this category. The second major tradition of autoethnography is that focused on life writing (e.g., Deck 1990, Denzin 1997) and the self and life story of the individual as an object of ethnographic interest. Hills's autoethnography in Fan Cultures (2002) is much closer to this life writing approach.
[2.8] The life writing or evocative turn in autoethnography has been challenged as being too narrow and self-involved, but attempts have also been made to reconcile the two traditions of autoethnography. Anderson (2006), for instance, proposes an analytic approach with five key features: analytic autoethnography is (1) conducted by a complete member researcher but (2) in dialogue with other community participants; it is characterized by (3) analytic reflexivity and (4) a commitment to theoretical analysis; and (5) the researcher's self is visible in the narrative. Voloder (2008) builds on Anderson in her attempt to reconcile the ethnographic and autobiographical traditions in autoethnography by examining the concept of analytical distance between researcher and participant in her study of the Bosnian community in Melbourne. She argues for a separation of the insider/outsider dichotomy on the one hand and the concept of distance on the other, using distance instead as an analytical tool to compare her own and her participants' experiences and therefore the wider social and cultural factors shaping them. Even as an insider to the community one is studying, one may still achieve analytical distance while using insider knowledge as part of one's research. In this way, Voloder casts herself in the roles of both ethnographer and informant without making her own experiences the sole focus of her research.
[2.9] It is this kind of analytical autoethnography, in the sense of studying a community one is already part of but also leveraging one's own insider understanding of that community, that meshes particularly well with the digital setting, allowing digital ethnographers to overcome the challenges and shortcomings of more traditional approaches. Hine (2015) highlights the increasingly individualized experience of the internet, as social media offers tailored content and users curate their own experiences by choosing who to connect with and which links to follow. This makes it impossible for the ethnographer to reconstruct a holistic picture of life online (see also Cook, Laidlaw, and Mair 2009), but at the same time this uncertainty and ambiguity mirrors participants' own experience of the environment. This mirroring makes a partially autoethnographic approach to online interaction valuable, and Hine argues for "considering how connections present themselves and what choices are available for building meaning out of these diverse influences" (Hine 2015, 83). In this way, the digital and the auto of ethnography complement each other, making this approach well suited to a significant proportion of fan studies research, particularly that taking place in online settings and in the communities where we, the researchers, are also we, the fans. Our understanding of fan communities, their output, and their norms is a strength that allows us to make informed choices at all stages of our research—from finding our research questions to designing our approach, deciding how to interact with our communities as both fans and researchers, and achieving the kinds of insights unavailable to outsiders. In the next section, I examine how this home field advantage (Dyck 2000) allows us to respond to one of the big challenges of online ethnography, that of constructing our field site.
3. From site to journey: Ethnography for the internet
[3.1] Fannish communities such as those of readers and writers of fan fiction do not so much form a single community as a collection of different, loosely connected (and sometimes entirely disconnected) communities with permeable boundaries. They/we produce not only fan fiction works but also an extraordinary amount of discussion, commentary, personal updates, fannish squee (note 1), wank (note 2), and other material. It is impossible—and arguably undesirable—for a single ethnography to capture the range and volume of this material, and even a more narrow thematic focus leaves a potentially unmanageable task. The mediated, digital nature of these communities, dispersed across multiple interconnected platforms, raises further questions about how and where best to collect and select data, or what the boundaries of the community being researched are.
[3.2] The concept of the field site has been problematized by ethnographers from a range of disciplines even before the emergence of ethnography online. The specific approach to finding—or, more recently and accurately, constructing—a field site reflects (and at times obscures) a number of ontological and epistemological assumptions, as well as practical methodological considerations. In his classic challenge to traditional, geographically bounded ethnography, Marcus (1995) calls for a multisited approach, which would enable anthropologists to study and compare how phenomena of the world system manifest and are negotiated within a variety of specific, localized settings. He proposes a range of strategies for constructing such a multisited project, including following the movements of people or objects, and centering a study on a more abstract concept, such as a metaphor, a story, or a conflict. While this approach has been lauded as allowing ethnographers to break boundaries and ask questions about wider issues that more traditional approaches were unable to address, it has also been critiqued, both on epistemological and practical grounds. Candea (2007) identifies seamlessness—the idea that the global and the local are inextricably intertwined—as a key assumption of multisited ethnography and a contributor to the idea of holism within the discipline: the assumption that by studying multiple sites we can reconstruct the whole of the world system seen to exist at a level above them, "a strange hope that once we have burst out of our field-sites, we can conquer the seamless world" (174). On a practical level, Candea argues, multisited ethnography pays little attention to the process of constructing the field site, thereby obscuring key methodological choices through the overreliance on the theoretical whole. Building on this and his challenge to multisited ethnography's particular conception of holism, he suggests that the arbitrarily bounded field site, consciously and reflexively constructed by the ethnographer taking into account the ethical and political implications, is a methodological choice that allows for continued productive engagement with ideas of seamlessness and complexity.
[3.3] Theoretical considerations and the specificity of research questions also play a role in how the ethnographer's field is conceptualized. Cook, Laidlaw, and Mair (2009) build on both Candea's critique of multisited ethnography and on their own insights from the ethnographic study of religion to advocate for a conception of the field as "un-sited" (47). Taking a broadly social constructionist approach, they reject the idea of the whole outright, arguing instead that it is "possible that no such single system exists" (53) and phenomena are "intrinsically interactive,…result from processes of assemblage or arrangement of entities" (55). Instead of the multisited ethnography, they propose that field sites be constructed based on the theoretical underpinnings of research questions: arbitrary in their multivocality and following of networks of people and other phenomena and yet on solid theoretical ground. To achieve this, the authors make a distinction between the concepts of space (abstract, impersonal), place (space imbued with meaning), and field (the object of ethnographic study). The field is decoupled from space and place, constructed across geopolitical and cultural boundaries, enabling the study and comparison of phenomena of theoretical interest. The authors point out that this approach is as applicable to places that occupy physical space as it is to nongeographic imagined communities.
[3.4] The online environment that many fandom scholars find ourselves in further complicates ethnographic field site construction through a number of practical challenges. Hine (2015) posits an internet that is embedded, embodied and everyday, particularly in the Western world. A plethora of apps on mobile devices, a range of social media platforms, and highly individualized engagement patterns give rise to uncertainty, complexity, and the feeling that something is always being missed. Hine argues, however, that this uncertainty is not only a challenge for the ethnographer but an intrinsic part of the experience of internet users themselves, and so "experiencing and embracing that uncertainty becomes an ethnographer's job, and pursuing some form of absolute robust certainty about a singular research object becomes a distraction, and even a threat, to the more significant goal of working out just how life is lived under these conditions" (5).
[3.5] Building on Wellman's (2001) concept of networked individualism—the idea that our experiences are increasingly tailored and individual to us and that the focus of networks has shifted from connecting places to connecting people—Hine argues that ethnography's key strength is its adaptability to new environments and advocates for a certain methodological eclecticism in its conduct. She also makes a case for utilizing the tools of everyday internet use in the construction of the field: search engines, social media platforms' tagging and filtering systems, trending topics, and other similar features shape users' everyday experiences of the internet, and can therefore be productively used by the ethnographer as tools for field site construction. Marres and Weltevrede (2013) go as far as suggesting these features may not simply be part of our research methodology but form part of the object of study itself. Beaulieu and Simakova (2006) use a similar approach in their study of the temporal dimensions of hyperlinks. Building on multisited approaches to ethnography and previous ethnographic studies of online phenomena, boyd (2008) also foregrounds the network aspect of online environments and proposes a networked ethnography that "involves finding different entry points into a phenomenon, following different relationships between people and practices, and making sense of different types of networks and their relation to one another" (54). Similarly to the unsited field Cook, Laidlaw, and Mair (2009) propose, here the components of the field site are constructed in relation to each other and the research question. Using the everyday tools of internet navigation to construct a networked field site, however, does ultimately reach a limit or boundary, even if only in the practical sense. Hills (2002) notes in his discussion of using cyberspace ethnography in fan studies that the mass of data available online makes it clear that selection and the construction of a boundary is required as "no a priori meaningful or internally coherent corpus can be identified" (174). It is therefore also important to pay close attention to how that boundary is constructed in ways that may be both deliberate and arbitrary at the same time (Candea 2007).
[3.6] Following boyd's (2008) networked ethnography approach of multiple entry points and tracing connections is an experience not dissimilar to that of discovering an online fandom community for the first time, highlighting Hine's (2015) argument that the experiences of the ethnographer mirror those of participants in online spaces. An internet search may lead to a fan fiction archive; users of the archive may include hyperlinks in the paratexts of their stories, pointing to their presence on more interactive social networking sites; social networking sites in turn may be structured in a way that aids discovery of other users with similar interests; finally, as those other users' interests are unlikely to be limited to the one fandom or piece of content originally used to find them, they may lead to other fandoms, groups or discussions. Any node in this network may be an entry point in its own right, as participation on social networking sites may lead to the discovery of fan fiction, or a trusted online or off-line friend may bring someone into the community by sharing fannish content with them. The fan and the ethnographer move through networked fannish spaces in remarkably similar ways. While networks are by no means static, the concept of network does not foreground either movement, or a temporal dimension. Yet fans frequently express a sense of their own fannish history, which involves moving both through time and through different networked fannish spaces. Additionally, both the network itself and the fans moving through it change over time: platforms and accounts are abandoned, groups drift out of touch or split due to conflict, new trends and tropes emerge. For that reason, digital (auto)ethnographies of fannish spaces are not so much ethnographies of networks as they are journeys. For the ethnographer, this journey, underpinned by the theoretical foundations of their research questions, becomes the field site.
4. Follow the trope
[4.1] For a fan studies scholar who is also already a fan, the familiar activity of navigating their way around the digital fannish landscape becomes a key tool for field site construction, and for managing their own participation in and interaction with the community they are researching. In this section, I will use the case study of my own research into the treatment of issues of sexual consent in erotic fan fiction to suggest how some of the challenges raised by the digital research environment and a scholar's prior involvement with the community they are studying may be addressed in a way that is both productive and theoretically robust.
[4.2] While online ethnography complicates the three key factors commonly identified as shaping the role of the ethnographer within the research setting in a number of ways, the autoethnographic aspect of many fan studies scholars' work provides ways of mitigating those complications. My approach to studying online fan fiction fandom was shaped significantly by the fact that I was already a member of the community before I began my research, giving me home field advantage (Dyck 2000) and access to what is only a semiopen research setting. It was also shaped by different traditions of autoethnography: the insider ethnography (Jenkins 1992) and to a lesser extent life writing approaches (Hills 2002) of fan studies, the reflexive autoethnography mirroring participants' experiences of the internet environment (Hine 2015), and the uncomfortable but productive tension between my roles as researcher, fan, informant, and analyst (Voloder 2008; Anderson 2006). My work was autoethnographic in the sense that my field was also my home and I was studying a community I was also a member of. It was also partially autoethnographic in the sense that I adopted the role of informant and used my own experiences of online fan fiction fandom to inform my data selection, collection, and analysis while situating these experiences firmly "within a story of the social context in which [they occur]" (Reed-Danahay 1997, 9). My level of participation (Gold 1958) or membership (Adler and Adler 1994) in online fan fiction fandom did not change as I became a researcher, though the modes of engagement did. One key challenge was becoming visible as a researcher in an environment where I had previously been a fan, and where presence and interactions (particularly on Tumblr) were largely ephemeral (Rutter and Smith 2005; Cho 2015). My choices about the level of openness about my research were driven by the nature of my participation in the community. My main forms of engagement were through posting content on AO3 and interacting with other members through a Tumblr blog. Neither of these offer a stable presence from the point of view of individuals seeing my occasional posts on their Tumblr dashboard or browsing AO3 for stories about a particular pairing or character. Here, my community membership and familiarity with community norms allowed me to identify ways to out myself as a researcher—for instance, through posting occasional reminders about my work in contexts where they were appropriate by community standards, such as in response to memes asking about work and personal information.
[4.3] My membership of the fan fiction community was also a key asset in constructing my field site. To collect and select material for analysis, I supplemented boyd's (2008) networked ethnography approach with my own experience, in-depth community knowledge, and cultural competence (Hine 2000) as a fan and the theoretical underpinnings of my research questions (Cook, Laidlaw, and Mair 2009). This led to what I call a follow-the-trope approach to my data collection and analysis. To understand this approach, it is first necessary to consider some of the tools and practices available to fan fiction community members in their day-to-day interaction with and navigation of the vast amounts of fan fiction and other community output available in online spaces. This will shed light on "how connections present themselves and what choices are available for building meaning" (Hine 2015, 83) within this particular community context.
Figure 1. A work of fan fiction as presented on AO3.
[4.4] The preservation of fannish history and output is a central concern to many fan fiction communities (Versaphile 2011). Archiving fan works and making them easily accessible and searchable plays a key role in such preservation efforts, and was a central motivation in the creation of AO3 (astolat 2007), which was one of the central sites for my research. AO3 has a range of features intended to make fan works easily discoverable and filterable on a range of criteria. Community members make use of these features on a day-to-day basis to navigate AO3 and choose works to read. The creation, management, and use of metadata is particularly important for this purpose. Figure 1 shows how a fan fiction story is presented on AO3. The story is preceded by a block of metadata, progressing from the generic to the specific. These are the kinds of technical features of AO3 that fans use to organize and search for fan fiction stories and navigate fannish spaces, and they can be used just as effectively as an entry point to constructing a networked field site (Wellman 2001; Hine 2015; boyd 2008). My theoretical interest in issues of power and sexual consent, my familiarity with trends and tropes in fan fiction, and my competence in navigating spaces such as AO3 were key in allowing me to identify the parts of fannish metadata that would most easily lend themselves to finding and narrowing down the kinds of content that would be suitable for my research. Archive warnings such as "rape/non-con" and "underage" are specifically designed to highlight consent issues to readers, and this feature became part of my analysis of fannish spaces as structured around consent. Ratings such as "Mature" and "Explicit" allowed for a narrowing down of works under consideration to those with at least some sexual content. And my immersion in fan fiction allowed me to identify freeform tags for tropes such as "dub con," "consent issues," or "fuck or die" as the kinds of spaces where issues of power and consent might be explored—hence, follow the trope. The sheer volume of material, however, continued to present a challenge. Here, too, fannish practices of filtering, sorting and selecting stories to read proved helpful. Sorting search results by the number of comments or kudos, for instance, allowed me to easily identify stories within a certain trope or pairing that were widely read and therefore considered influential by the community.
[4.5] My journey through this fannish landscape was guided by my theoretical interest in issues of sexual consent in fan fiction. I was particularly struck by the discrepancy between early fan studies approaches to slash as a genre of gender equality (Lamb and Veith 1986; Russ 1985; Jenkins 1992; Kustritz 2003) and my own observations of the emergence of tropes and subgenres within slash that seemed to deliberately explore unequal relationships (note 3). This also chimed with my theoretical focus on the operations of power in sexual relationships and the impact of power on individuals' ability to meaningfully negotiate consent. Immersion in several fandoms, awareness of trends and popular tropes, exploration of the relevant metadata identified above, and cultural competence within the community allowed me to choose two popular and sometimes controversial tropes or subgenres within slash fan fiction— omegaverse and arranged marriage, both characterized by significant power disparities between the characters—to explore further. To find specific fan fiction stories for my analysis, I retraced the steps a fan would use to find and select works featuring specific types of content. I used AO3's tagging functionality to search for works containing the relevant tags. I narrowed this search by focusing on a single, well-represented fandom and pairing within each trope. I then proceeded to read a wide selection of popular and impactful—as indicated by the AO3 kudos functionality—stories featuring the pairing and trope in question to understand common features and approaches. Acknowledging the networked nature of the community I was studying (boyd 2008), I also followed links and connections to and from additional material, such as fannish commentaries (meta), introductions to popular tropes or pairings (primers), fan-curated recommendation lists (recs), and histories to help me situate the material within a wider cultural context.
[4.6] The ethnographic follow-the-trope approach cannot, by definition, present a holistic or universal picture of the fandom community: different community members have different entry points, different interests, different networks, and different experiences (boyd 2008; Hine 2015). While the stories I chose for analysis are popular and impactful within the communities in which they circulate, there are hundreds of other stories—some similar to these, some different—I could have chosen, and my analysis of them would potentially have yielded results that were similar in some ways and different in others. Like Candea's (2007) both deliberately and arbitrarily bounded field site, the follow-the-trope approach does, however, present one possible, specific, and situated view of how some parts of the fan fiction community engage with issues of sexual consent. These parts of the community, and these engagements, may and do coexist with others, as well as with an absence of engagement in many cases—the community is far from homogenous. Importantly, however, the engagements I was able to demonstrate in my research are found in some of the community's most popular and celebrated contemporary output and can therefore be said to have an impact on the wider fan fiction community; and these insights could not have been generated without being a member of the community I was studying, in an online setting.
5. Conclusion
[5.1] For a fan studies scholar interested in particular aspects of fan fiction, follow the trope is a robust ethnographic approach to data collection and analysis in an online setting and in communities the researcher may already be a member of. It allows researchers to use the theoretical considerations of their research questions as a guide in making their journey through interconnected online spaces. In the case of my research, my interest in the operation of power in questions of sexual consent was a key driver in identifying specific fan fiction tropes for closer analysis. It makes visible, rather than obscuring, the choices we as researchers make on that journey in finding and selecting material, as well as in our approach to analyzing it. In my research, it allowed me to select a small number of fan fiction works for in-depth analysis while keeping that choice, along with its advantages and limitations, front and center in my analysis. By blending elements of analytic, reflexive autoethnography with insights from digital ethnographic methods, the researcher is able to leverage their own insider knowledge of both the communities they are researching and the online spaces where those communities live throughout the research process, from the identification of research questions, through data collection and analysis, to questions of representation and power imbalances between researcher and researched in any outputs. It was my in-depth immersion in fannish spaces and resulting understanding of fan fiction tropes and trends that allowed me to identify the tropes most relevant to the theoretical considerations of my research questions. The emphasis on the researcher's own experiences alongside those of community members encourages reflection on positionality and enables the researcher to consider key aspects of their role as an (auto)ethnographer in light of the practical and theoretical challenges of the online setting. My understanding of fannish spaces and norms allowed me to make ethical and practical choices in how I presented myself in fannish communities, what material I selected for analysis, and how I presented that material in the outputs of my research (note 4). While acknowledging the challenges of conducting ethnographic research in digital spaces and communities we are already members of, follow the trope ultimately seeks to leverage those factors as strengths in our research.