Autoethnography: proposing a New Research Method for Information Systems Research

Emerging technologies are facilitating the production of revised and novel forms of “digital being” combined frames of meaning, experience, and desired notions of performativity that change what and who we are. A number of theoretical perspectives (e.g. Agential Realism and Sociomateriality) have emerged that seek to address this new reality but have struggled to grapple with the relationship between technology and the things that constitute a human in a psychological sense. In this paper, we argue that IS researchers have been hampered by the paucity of established research methods that are suited for the investigation of emerging digital systems and new forms of digital being. We believe that the IS community has made a serious error by ignoring repeated calls for the use of ethnographic methods in the field. We also feel that autoethnographic methods are highly suited to the task of conducting research on the digitally mediated experiences in everyday activities and in facilitating the development of new theories of digital being. The purpose of this paper is to provide a practical introduction to autoethnography and to explain how IS researchers might do and write autoethnography. The paper focuses on the challenges associated with using an autoethnographic approach in digitally mediated research settings. The paper also focuses on how IS researchers should evaluate autoethnographic research because the main challenge for ethnographic IS research has always been the evaluation of it. Keywords: autoethnography; ethnography; information systems; digital life; evaluation 1 Introduction and background Information Systems (IS) researchers have long been divided about the intellectual core of the IS field (Orlikowski and Iacono, 2001; Benbasat and Zmud 2003). A part of the problem is that technology is best understood as an equivoque: something that “admits of several possible or plausible interpretations and therefore can be esoteric, subject to misunderstandings, uncertain, complex, and recondite” (Weick, 1990, p. 1). In recent years, a growing consensus has begun to emerge that the main phenomena of interest in IS occur in real life settings, are continually evolving (Hassan, 2011) and emerge from the interaction between social and technological systems (Lee, 2001, p. iii; cf. Leonardi and Barley, 2008; Orlikowski, 2007; Kautz and Jensen, 2013). The emerging literature on Agential Realism and Sociomatierality (e.g. Barad, 2007; Orlikowski and Scott, 2008) asserts the impossibility of separating people from technology in understanding social phenomena and recognises that the technological and the social are inextricably entangled in everyday, digitally-mediated practices and that the main phenomena of interest emerge from the intra-action (rather than interaction) of social and technological systems. This has led to repeated calls for IS Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 2 researchers to develop new theory to explain the fusion of technology and work in organizations (Orlikowksi and Scott, 2009) in order to better understand how emerging digital systems have begun to facilitate the production of revised and novel forms of digital being (Chiasson and Teigland, 2014). However, IS researchers have struggled to investigate sociomaterial phenomena in a comprehensive and coherent way (Cecez-Kecmanovic et al., 2010). For instance, there have been renewed efforts to investigate the emergence of new forms of embodied (digital) identity (e.g. Schultze, 2014) but such efforts have been unable to specifically focus on the enactment of digital identity through the constitutive entanglement of actor and artefact in virtual worlds and have not yet succeeded in finding a way to evoke rather than represent the discursive, embodied, material and immersive aspects of digital identity in virtual worlds. This challenge has not yet been overcome because IS researchers still do not have a significant arsenal of research methodologies at their disposal (Rowe, 2012; Chen and Hirschheim, 2004; Myers in 1997 and 1999; Orlikowski and Baroudi, 1991) and more specifically, because of the limited capacity of existing research methods to shed light on the inextricable entanglement of actor and artefact (Hassan and Hovorka, 2011; Orlikowski and Scott, 2009). Thus, there have been repeated calls for a more pluralist approach (Hassan, 2011; Mingers, 2001; Robey, 1996; Avison and Myers, 1995; Landry and Banville, 1992) to IS research and there have been specific and repeated calls for more narrative research (Hunter, 2012) and for more ethnographic research (Schultze, 2000; Myers, 1997). After ground-breaking work by Wynn (1979) and Zuboff (1988), ethnography has frequently been discussed in the IS community (see Schultze, 2000; Myers, 1999; Avison and Myers, 1995; Lee, 1991; Pettigrew, 1985) and in the social sciences generally (Bryman, 2001). Ethnographic research relies on first-hand observations made by a researcher immersed over an extended period of time in a culture in which s/he is unfamiliar (Atkinson and Hammersley, 1994). It is one of the most in-depth research methods available and is well suited to obtaining a deep understanding of people and the broader context within which they are embedded (Myers, 1999). The particular strength of ethnographic methods for IS research is that they provide rich insights into the human, social and organisational aspects of IS phenomena (Schultze, 2000; Myers, 1999; Avison and Myers, 1995). More specifically, ethnography is well suited to the investigation of phenomena that emerge from the interaction between social and technological systems in real life settings; by definition, it is focused on the practice of everyday life (Schultze, 2000). At the same time, ethnography sheds light on a plurality of (insider) views whilst simultaneously incorporating the full variety of research evidence (Myers, 1999). There are three main types of ethnography that are relevant to IS researchers: reflexive, holistic and critical ethnography (Myers, 1999; Schultze, 2000). Reflexive ethnography is particularly appropriate for IS research because it produces an awareness that there are no absolute distinctions between ‘truth’ and ‘fiction’, between the ‘self’ and the ‘other’ (Foley, 2002), and by implication, between the social and the technological, the personal and the professional. Reflexive ethnography takes into account the researcher’s self in interaction with the object of study (Davies, 2008). It is frequently written as a vulnerable (Behar, 1996), confessional (van Maanen, 2011) or autoethnographic (Ellis, 2004) text in which the author’s gaze turns and bends back upon themselves (Babcock, 1980). In this formulation, “the self is a multiple, constructed self that is always becoming and never quite fixed, and the ethnographic productions of such a self and the ‘cultural other’ are always historically and culturally contingent” (Foley, 2002, p. 473). Reflexive ethnography is also significant because it embraces new forms of authorship and attempts to subvert the restrictive social scientific writing conventions (cf. Ellis and Bochner, 2000; Ellingson and Ellis, 2008) that have limited the genres of writing that are used in the field as a whole (cf. Rowe, 2012) and that have limited its elite journals in particular (cf. Dennis et al., 2006). As such, reflexive methods like autoethnography have the potential to broaden the reach of IS research findings and to attract the interest of practitioners in particular, who are the primary audience for IS researchers (Taylor et al., 2010). Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 3 Nevertheless, ethnographic methods have failed to gain any significant traction in the IS field (Levina, 2005). A recent analysis conducted by Paré et al.(2008) found that only 2.6% of the IS papers they sampled use ethnographic methods. Given that other research methods like action research have come to be seen as core research methods in a relatively short period of time (cf. Liu and Myers, 2011), it seems that IS researchers are willing to adopt new methods but are encountering particular difficulties with ethnographic methods. This is at least partly explained by the fact that ethnographic research is more resource intensive than other research methods but it is also argued that part of the problem is that there is less clarity within IS about the evaluation criteria needed for ethnography (Schultze, 2000) even as evaluation criteria for methods like action research have become much more transparent (Baskerville and Woodharper, 1996). Based on this analysis, the purpose of this paper is to encourage IS researchers to reconsider ethnographic research methods. In particular, the paper focuses specifically on autoethnography and explains how IS researchers might do and write autoethnographic research and how IS editors might evaluate autoethnographic IS research. 2 About autoethnography Autoethnography is “an autobiographical genre of writing and research that displays multiple layers of consciousness, connecting the personal to the cultural” (Ellis and Bochner, 2000). It is a reflexive form of ethnography (see Figure 1) in which the researcher’s interaction with the object of study is taken into account (Davies, 2008) and in which the author’s gaze is turned and bent back upon itself (Babcock, 1980). Its main purpose is to link the micro and the meta (Boyle and Parry, 2007); more specifically, it is used to describe and systematically analyse (graphy) personal experience (auto) in order to understand cultural experience (ethno) (Ellis, 2004; Holman Jones, 2005). As such, it involves the production of highly personalized accounts where authors draw on their own experiences to extend understanding of a particular culture (Reed-Danahay, 1997). In autoethnographic research, the researcher uses the self as an ethnographic exemplar. In this way, the subjectivity of the researcher is seen as something to capitalize on rather than exorcise (Glesne & Peshkin, 1992). As such, the ontological foundation for truth in autoethnography is the self who was “there”. In this context, personal experience (or what might be called ‘thereness’) becomes what Britzman (1995, p. 28) calls “the great original”: it assures the reader that there is both a ‘there’ and ‘beings’ who are there. This perspective is a particularly useful when investigating new forms of digital being, where basic assumptions about existence and experience are routinely (and rather unhelpfully) challenged: “it’s not real; it’s only virtual”. Of course, it is difficult to write about the self and to be an escape artist from the self at the same time (Gannon, 2006). According to Gannon (2006), “the self both is and is not a fiction; is unified and transcendent and fragmented and always in process of being constituted, can be spoken of in realist ways and cannot; its voice can be claimed as authentic and there is no guarantee of authenticity” (p. x). But by writing themselves into their own work, autoethnographers are able to challenge accepted views about silent authorship (Holt, 2003). Instead, they try to remove the ‘metarules’ of scholarly discourse, which “privilege arguments, theories, abstractions and jargon over feelings, stories, concrete events, and accessible prose” (Ellis and Bochner, 2000, p. 746), from their work. In this way, the “unique voicing” which emerges complete with colloquialisms, reverberations from multiple relationships, and emotional expressiveness – is honoured (Gergen and Gergen, 2002, p. 14). It is to this unique voice that the reader respond (Gernen and Gergen, 2002) precisely because it expresses the struggle to make sense of experience (Boyle and Parry, 2007). Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 4 Critical ethnography Reflexive ethnography Holistic ethnography An emergent process involving dialogue between the ethnographer and study participants where otherwise hidden agendas are revealed and scrutinized An approach that takes into account the researcher’s self in interaction with the object of study Initially a blank slate, the ethnographer tries to “go native”, identifying with the social group under study and soaking up the language and culture of study participants Figure 1. Locating autoethnography in the spectrum of ethnographic research methods (Derived from: Myers, 1999) 3 Doing autoethnography: Understanding the Process Autoethnography combines characteristics of autobiography and ethnography (Chang, 2008). As such, the literature borrows from the ethnographic literature on participant observation in terms of doing autoethnography and form the literature on autobiography in terms of writing autoethnography (Chang, 2008). Autoethnography, in the first instance, relies heavily on participant observation, in which researchers participate in the lives of their informants while observing their behaviours (Chang, 2008). The particular challenge for the autoethnographer is to be able use methodological tools and research literature to analyse experience and at the same time to use personal experience to illustrate facets of cultural experience, and, in so doing, make characteristics of a culture familiar for insiders and outsiders (Ellis, Adams and Bochner, 2011). In autoethnography, it is also especially important that the researcher observes their own behaviours and document their thoughts while living them during participant observation. When this is done individually, it is called “introspective self-observation” and when it is done through interaction with others, it is called “interactive introspection” (Chang, 2008). In this case, the researcher may collaboratively interview others as a way of reliving and describing previous experiences (Ellis and Berger, 2003). The power of self-observation is that it brings the habituated and the taken-for-granted to the surface and it gives access to covert, elusive, and/or personal experiences including cognitive processes, emotions, and motives (Rodriguez and Ryave, 2002, p. 3). In particular, it can be used to Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 5 uncover actions that might otherwise be inadvertently or deliberately concealed (ibid.). In addition to self-observation, the autoethnographer should also compare and contrast personal experience against existing research and examine any relevant cultural artifacts (Ellis et al., 2011). Field texts should be also written up promptly and regularly, and research materials should be reviewed and revised on an ongoing basis (Myers, 1999). In all cases, the autoethnographer must pay particular attention to the question of research ethics: in using personal experience, autoethnographers implicate both themselves and close, intimate others in their work. For this reason, ethical issues affiliated with friendship are a particularly important part of the research process and product (Ellis et al., 2011). In terms of actually generating field texts, the autoethnographer is asked to devise strategies for collecting and organising data from the start in order to better cope with the variety of texts that are likely to be generated. Techniques that may be useful include: (1) Using visual tools like free drawings (2) Inventorying people, artifacts, activities, proverbs (3) Chronicling the autoethnographer’s daily life (4) Reading and responding to other autoethnographies (5) Collecting other field texts such as stories, personal journals, letters, conversations, interviews, documents, photographs, memory boxes, and life experiences Source: Connelly and Clandinin, 2000, p. 101 Though texts have been written to clarify how online ethnography (cyberethnography or netnography) differs from traditional ethnography, there are no comparable texts available for autoethnographers interested in online forms of autoethnography. As such, the IS researcher wishing to investigate aspects of digital being and the digital self must address fundamental questions about the nature of ‘being’ and the nature of ‘there’ when the research is conducted in an electronically mediated setting. For example, it is difficult to apply the singular notion of ‘there’ to virtual worlds where users can teleport within and across different virtual worlds at the click of a button. Similarly, it is difficult to distinguish between observation, participant observation and self-observation when the research is conducted in an electronically mediated setting. If an IS researcher doesn’t really feel ‘immersed’ in a virtual world, for example, is it legitimate for them to claim to have engaged in participant observation? An IS researcher may also find it difficult to distinguish between observation and selfobservation when the self is, for example, an avatar. In addition (and this issue should not be underestimated), the IS researcher is likely to be utterly overwhelmed by the volume and diverse forms of data that can be collected and analysed in electronically mediated research settings. Further, there are few guidelines available to IS researchers wishing to incorporate multimedia into their analyses. Given the emphasis placed on generating useful field texts, this is a particularly onerous challenge for the IS autoethnographer and also a serious opportunity lost. 4 Writing autoethnography: Understanding the product Ethnography is a not an innocent practice. Our research practices are performative, pedagogical, and political. Through our writing and our talk, we enact the worlds we study – Denzin, 2006, p. 422 Autoethnographic writing must satisfy two main aims: it must make personal and cultural experience meaningful and it should be accessible to a wider audience than traditional research (Ellis et al., 2011). In achieving these aims, autoethnographers engage in creative analytic practices (cf. Richardson et al., 2008, p. 660) based on an evocative or aesthetic style of writing (Denzin, 2006). Writing autoethnography is considered to be at least as important as doing autoethnography. In the first instance, autoethnographers are actively engaged in attempting to communicate more than Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 6 traditional genres (Ellis et al., 2011). For this reason, there is significant interest in developing new techniques to increase the evocativeness, verisimilitude and appeal of research texts. In the second instance, there is a lot of controversy around autoethnographic writing (Wall, 2008). Autoethnographers are described as self-absorbed narcissists who fail to fulfil scholarly obligations of hypothesizing, analysing, and theorizing (Sparkes, 2000). Similarly, autoethnographic texts are described as self-indulgent, introspective, and individualized stories (ibid.), that fictionalise life and romanticise the self (Atkinson 1997), and have greater therapeutic than analytic merit (Gannon, 2006). Notwithstanding the significance of writing autoethnography, there are few texts available on how to do it (Tierney, 2002). Autoethnographers are generally slow to offer prescriptive advice on how to write autoethnography because such advice is seen to contribute to the generation and edification of new writing conventions and to the suppression of the very thing the autoethnographer is trying to emancipate: the voice of the researcher. 4.1 Autoethnographic writing: form and writing style I start with my personal life. I pay attention to my physical feelings, thoughts, and emotions. I use what I call systematic sociological introspection and emotional recall to try to understand an experience I've lived through. Then I write my experience as a story. By exploring a particular life, I hope to understand a way of life Ellis and Bochner, 2000, p. 737 Autoethnography, like other forms of reflexive ethnography (i.e. confessional and vulnerable ethnography), is based on writing retroactively and selectively about past experiences (Muncey, 2010). Autoethnographers seek out textual strategies that evoke fractured, fragmented subjectivities and provoke discontinuity, displacement, and estrangement (Gannon, 2006) and the writing styles used by autoethnographers fall “somewhere in the continuums between realist description, impressionist caricature, analytical description and confessional self-exposure” (Chang, 2008). More specifically, autoethnographies are often written as layered accounts that focus on the author's experience alongside data, abstract analysis, and relevant literature (Ellis, et al., 2011). These layered accounts increase the evocativeness and aesthetic quality of the text by enabling the writer to incorporate multiple voices including theory, subjective experience, fantasy, and more to convey aspects of a topic at hand that would be otherwise excluded from a more traditional format (Rambo, 2005). Thus, the truths of autoethnographic writing are liminal, dynamic, and contingent (Hutton, 1988, p. 140) and the writing writes the writer as a complex (im)possible subject in a world where (self) knowledge can only ever be tentative, contingent, and situated (Gannon, 2006). Autoethnographers have developed a range of stylistic techniques that help to bring the text to life. In our review of the literature, we identified 8 unique textual strategies that have been used by autoethnographers to date: The use of figurative language (e.g. simile or metaphor) to illustrate unfamiliar concepts The use of single or multiple plot line to create interest The use of foreshadowing techniques The use of an additional narrator other than the author The use of mixed tenses to create drama or suspense in the text The use of different fonts to distinguish different voices in the text The use of different indentations to distinguish different components of the text The use of icons, symbols and other tactics to clarify the purpose of each element of the text. When electronically mediated settings are increasingly immersive and interactive, the IS researcher has much to gain from using these techniques to better represent these settings. For example, it has always been difficult to effectively communicate the nature of being in virtual worlds to audiences who are unfamiliar with them. Autoethnographic writing, which is performative by nature, can help those who would otherwise struggle to describe their experience (Mitra, 2010). For example, a virtual worlds researcher could effectively describe their experiences using a screenplay format. In the first Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 7 instance, this format would allow the researcher to incorporate textual or even graphical descriptions of scenes, to construct stage directions to describe action, interaction, layout and appearance, to write dialogue between avatars and bots etc. The researcher could also use this format to describe the actual interface between the human being, the keyboard, the screen and the avatar. This kind of writing is much needed because one seldom sees a comprehensive analysis of the multiple forms of embodied being that coexist when one interacts in a virtual world. That is to say, one seldom sees meaningful analyses of the relationship between the human being, embodied in flesh and interacting with a machine, and the avatar, embodiment in pixels, inhabiting in 4-dimensional space (a space of length, breadth, height and duration) and interacting with other avatars (synchronously, asynchronously, in ‘person’ or otherwise) and with the environment itself. As this illustration seeks to illustrate, the IS researcher wishing to investigate aspects of digital being and the digital self can easily leverage these techniques to bring electronically mediated settings to life in their writing. 4.2 Autoethnographic writing genres According to Boyle and Parry (2007), the main forms of autoethnographic writing include: personal essays, poetry, short stories, journals, stream of consciousness, detailed unstructured interview narratives, visual presentations, conversation stories, polyvocal texts and other forms of fragmented writing. In Table 2, we present a more complete and systematic analysis of the autoethnographic writing genres that have already been developed. Genre Description Exemplar Layered account The authorial voice of the narrative is altered such that formal writing is ‘spliced’ with ethnographic and autobiographical materials, self-reflexive texts, and other researchers’ arguments Ronai (1995) Dialogue The narrative is written as an (possibly fictitious) interview or dialogue or even as a conversation between two or more people Ellis and Bochner, 2000 Vignette The narrative is structured around epiphanies which are presented as short, self contained evocative vignettes. Vignettes are vivid portrayals of the conduct of an event of everyday life, designed to enhance the contextual richness of ethnographic research (Miles and Huberman, 1994, p. 83). A well written vignette will make an otherwise esoteric concept tangible Humphreys (2005) Co-created narrative Each author first writes his or her experience around an epiphany and then shares and reacts to the story written by the other author Toyosaki and Pensoneau (2005) Extreme case The narrative is structured around surprises (comparing the actual experience with expectations) or contradictions. In this style, the narrative zones in on surprises and contradictions. That is to say, the researcher seeks out disjunctures and jarring moments, focusing on contrasts, paradoxes, and irregularities rather than patterns, themes, and regularities (cf. Coffey and Atkinson, 1996) Mitra (2010) Table 2. Taxonomy of autoethnographic genres ordered in terms of frequency of occurrence Autoethnography: proposing a new method for IS research Twenty Second European Conference on Information Systems, Tel Aviv 2014 8 5 Evaluating autoethnography One of the reasons why ethnographic methods in general and autoethnographic methods in particular have not been widely diffused in the IS field is a lack of clarity about evaluation criteria (Schultze, 2000; Levina, 2005). To address this problem, we have reviewed the main criteria that are typically used to evaluate qualitative (cf. Tracy, 2010; Lincoln and Guba, 1985), ethnographic (Myers, 1999; Golden-Biddle and Locke, 1993), and reflexive ethnographic (Schultze, 2000; Richardson, 2000) research. From this analysis, we have synthesized a list of four criteria that should be used to evaluate autoethnographic research. Our analysis, summarised in Table 3, explicitly links each of the four criteria to established concepts in the literature.