In doing research of any kind, there is an implicit assumption that we are investigating something âoutsideâ ourselves, that the knowledge we seek cannot be gained solely or simply through introspection. This is true for both the social and natural sciences, although in the latter the separation of researcher and research object may appear both more self-evident and more readily attainable. On the other hand, we cannot research something with which we have no contact, from which we are completely isolated. All researchers are to some degree connected to, or part of, the object of their research. And, depending on the extent and nature of these connections, questions arise as to whether the results of research are artefacts of the researcherâs presence and inevitable influence on the research process. For these reasons, considerations of reflexivity are important for all forms of research. Although the connection between an astronomer and distant stellar events may seem very tenuous indeed, no more than an ability to observe secondary indications of such events by means of sophisticated extensions of human sensory equipment, even astronomers take account of their relationship to these occurrences, for example in discarding assumptions about simultaneity of observation and event. And in the realm of particle physics, questions about the effects of observers on their observations are of fundamental importance. If reflexivity is an issue for these most objective of sciences, then clearly it is of central importance for social research, where the connection between researcher and research setting â the social world â is clearly much closer and where the nature of research objects â as conscious and self-aware beings â make influences by the researcher and the research process on its outcome both more likely and less predictable. These issues are particularly central to the practice of ethnographic research where the relationship between researcher and researched is typically even more intimate, long-term and multi-stranded, and the complexities introduced by the self-consciousness of the objects of research have even greater scope.
Debates about reflexivity overlap with concerns about subjectivity versus objectivity in social research, although they are not coterminous. Nevertheless, responses to the difficulties apparently raised by reflexivity frequently involve attempts to ensure objectivity through reducing or controlling the effects of the researcher on the research situation. Such attempts include maintaining distance through using observation and other methods in which interaction is kept to a minimum or is highly controlled. Alternatively, claims to objectivity â or at least to reduce the effect of researchers on their results â are also made on the basis of a very high level of interaction, based on complete participation, in extreme cases even concealing the identity of the researcher. These approaches have been identified with positivist and naturalist methodologies, respectively (Hammersley and Atkinson 1995: 16â17). However, even the most objective of social research methods is clearly reflexive. Survey research based on structured interviewing, for example, can ensure a form of objectivity through training of interviewers to reduce the effects of their individual attributes on respondents and employing technical tests of reliability. But survey researchers cannot remove another, and more fundamental, form of reflexivity that inheres in their construction of a highly artificial research situation, which is dependent upon a set of cultural understandings as to the nature of interviews, their conduct and appropriate forms of responses to them. At the other extreme, covert participation may eliminate the researcherâs influence qua researcher but it does not eliminate effects of their presence on their results and may render such effects less visible.
Reflexivity, broadly defined, means a turning back on oneself, a process of self-reference. In the context of social research, reflexivity at its most immediately obvious level refers to the ways in which the products of research are affected by the personnel and process of doing research. These effects are to be found in all phases of the research process from initial selection of topic to final reporting of results. While relevant for social research in general, issues of reflexivity are particularly salient for ethnographic research in which the involvement of the researcher in the society and culture of those being studied is particularly close. The term ethnography is used to refer both to a particular form of research and to its eventual written product. I adopt a broad interpretation of ethnography as a research process based on fieldwork using a variety of mainly (but not exclusively) qualitative research techniques but including engagement in the lives of those being studied over an extended period of time. The eventual written product â an ethnography â draws its data primarily from this fieldwork experience and usually emphasizes descriptive detail as a result (cf. Davies 2002; Ellen 1984: 7â8; Hammersley and Atkinson 1995: 1â3).
Not only the personal history of ethnographers but also the disciplinary and broader sociocultural circumstances under which they work have a profound effect on which topics and peoples are selected for study. Furthermore, the relationships between ethnographer and informants in the field, which form the bases of subsequent theorizing and conclusions, are expressed through social interaction in which the ethnographer participates; thus ethnographers help to construct the observations that become their data. In an early recognition of the need systematically to incorporate reflexivity into ethnographic research methods, Powdermaker argued that participant observation requires both involvement and detachment achieved by developing the ethnographerâs ârole of stepping in and out of societyâ (1966:19). In order to incorporate such insights into research practice, individual ethnographers in the field â and out of it â must seek to develop forms of research that fully acknowledge and utilize subjective experience and reflection on it as an intrinsic part of research. Furthermore, given the contribution of the ethnographerâs sociocultural context to the research, these contexts too must be considered. They become a part of the research, a turning back in the form of cultural critique that has moral and political implications as well.
On the other hand, this turning back, or self-examination, both individual and collective, clearly can lead to a form of self-absorption that is also part of the definition of reflexivity in which boundaries between subject and object disappear, the one becomes the other, a process that effectively denies the possibility of social research. This outcome is closely related to various postmodernist and poststructuralist critiques which, in their most extreme forms, are essentially destructive of the enterprise of social research. Nevertheless ethnographers must seek to utilize creatively the insights of these postmodernist perspectives â insights that encourage incorporation of different standpoints, exposure of the intellectual tyranny of meta-narratives and recognition of the authority that inheres in the authorial voice â while at the same time rejecting the extreme pessimism of their epistemological critiques. In this book I develop epistemological and methodological foundations that encourage and incorporate genuinely reflexive ethnographic research while maintaining that such research can be based on a realist ontology, which assumes a social reality independent of our knowledge of it. In developing this perspective, I draw heavily on the work of Roy Bhaskar and other critical realists (cf. Archer et al. 1998) and also on pragmatism, particularly the insights of G. H. Mead.
From this perspective, the purpose of research is to increase our understanding of social reality by developing explanations of social forms and events, as well as critically examining the conceptualizations used in these explanations. However, critical realism also accepts that social research is inextricably tied to questions of meaning and interpretation due to the self-conscious nature of its subject matter. Unlike the natural sciences it is involved in a âdouble hermeneuticâ, that is, it is answerable both to the scientific community and to those being studied (Sayer 2000: 17â18). This implies that social research is a conduit that allows interpretations and influences to pass in both directions, and final products thus may take a variety of forms and be addressed to different audiences. Nevertheless, I will argue that the results of anthropological research based on ethnographic fieldwork, informed by reflexivity and assessed by a critical scholarly community, are expressive of a reality that is neither accessible directly through the actions and texts of those being studied nor simply a reflection of the individual anthropologistâs psyche. This means that both good and bad research are possible, and the development of criteria to recognize the difference should provide the basis of anthropological authority. My principal aim is thus to consider critically the actual activities of research in the context of an epistemological basis for anthropological knowledge that fully incorporates reflexivity while being rooted in a realist ontology.
Before looking in detail at various ethnographic research processes and methods, therefore, several other topics need be addressed. First, a more careful consideration will be given to the forms reflexivity assumes and its relationship to questions about ways of knowing and the nature of knowledge. Second, I briefly discuss the ways in which anthropologists have viewed reflexivity and its changing relationship to actual research practice and consider the implications of various postmodernist critiques for the practice of ethnographic research. Finally, I develop more fully a critical realist perspective that allows for the continuing practice of meaningful social research, while also benefiting from sensitivity to issues of reflexivity and responding to postmodernist and poststructuralist critiques.
Reflexivity and knowledge
Reflexivity in social research is not a single phenomenon but assumes a variety of forms and affects the research process through all its stages (Roberts and Sanders 2005). Babcock (1980) enumerates a series of dichotomies to describe varieties of reflexivity: private/public; individual/collective; implicit/explicit; partial/total. Some of these various dimensions can be placed along a spectrum: at one extreme is the relatively private, individualist and hence partially reflexive activity of the fieldworker keeping a journal â what has been termed âbenign introspectionâ (Woolgar 1988b: 22); further along is the public, collective activity of traditional rituals which display a form of âsocial reflexivityâ (Turner 1981; also cf. Rappaport 1980). But even examples of this level of social reflexivity must still be judged far from total in their implicitness (lack of self-awareness of their reflexive nature), in contrast to the journal writer. Total reflexivity requires full and uncompromising self-reference. Thus, it is argued, no process of knowing is fully reflexive until it is explicitly turned on the knower, who becomes self-conscious even of the reflexive process of knowing â what has been termed âradical constitutive reflexivityâ (Woolgar 1988b: 22). In this fullest form, reflexivity, in spite of its unavoidable and essentially desirable presence in social research, becomes destructive of the process of doing such research; as researchers we are led âto reflect on our own subjectivities, and then to reflect upon the reflection in an infinitude of self-reflexive iterationsâ (Gergen and Gergen 1991:77). It will be helpful to follow this process through by reviewing briefly the various levels of reflexivity and the ways in which they influence social research.
In its most transparent guise, reflexivity expresses researchersâ awareness of their necessary connection to the research situation and hence their effects upon it, what is sometimes called reactivity. This has often been conceived in terms of the subjectivity of the researcher, with attempts being made, especially from a positivist orientation, to ensure objectivity. For example, in conducting interviews, techniques are promoted (such as standardized wording of questions and controlling interviewer responses) so as to limit the effect of the interviewer on this particular social encounter. In ethnographic research, fieldworkers have adopted various strategies to make themselves inconspicuous and hence reduce the dangers of reactivity. They may rely on literally being an inconspicuous bystander; or they may take the opposite approach and reduce reactivity by participating as fully as possible, trying to become invisible in their role as researcher if not as human participant. Nevertheless, the impossibility of controlling the social encounters that provide the ethnographerâs data during fieldwork based on long-term participant observation has long been implicitly recognized in that claims to objectivity in fact came to be based less on the nature of the research encounter than on the objectifying rhetoric of reporting forms (Crick 1982a; Grimshaw and Hart 1995). Fuller recognition of the role of reflexivity eventually moved researchers beyond naive attempts to objectify the research encounter and towards an acceptance that in social research, âthe specificity and individuality of the observer are ever present and must therefore be acknowledged, explored and put to creative useâ (Okely 1996b:28). A developing critique of objectifying forms of ethnographic writing has accompanied this recognition (Marcus and Cushman 1982; Rosaldo 1993 [1989]).
Reflexivity in this form, while clearly calling attention to the nature of research as a social process, is still very much focused on the individual researcher. Yet even at this individualist level, considerations of reflexivity are compelled to move beyond the notion of the researcherâs effect on the data and begin to acknowledge the more active role of the researcher in the actual production of those data. Thus, âthe ethnographic enterprise is not a matter of what one person does in a situation but how two sides of an encounter arrive at a delicate workable definition of their meetingâ (Crick 1982a:25). Steier (1991b) goes further in viewing the research process as one in which researcher and reciprocators (not respondents) are engaged in co-constructing a world.
Ethnographers first came to consider the collective social dimension of reflexivity through identifying reflexive processes among the peoples that they studied. This perspective has been particularly useful and prominent in studies of ritual and performance. Perhaps the most frequently cited example is Geertzâs interpretation of the Balinese cockfight as âa Balinese reading of Balinese experience, a story they tell themselves about themselvesâ (1973: 448). Such social reflexivity may be explicit, a deliberate and conscious reflection of a people upon themselves, but it is more commonly presented as fully revealed only through the interpretative insights of the ethnographer. However, social reflexivity, especially in this latter form, preserves a privileged, and essentially non-reflexive, position for the ethnographer (cf. Watson 1987).
When the insights of this sort of social reflexivity, especially those that are grounded in a relativist and/or interpretivist perspective, are combined with the reflexivity of the individual researcher in recognizing that data are very much a cooperative product, then they tend to stimulate reflexivity of a more searching and critical form which encompasses the knowledge claims of social researchers themselves. Why should this be so? If we argue that the activities and texts of our informants are really expressing not their obvious surface message but an underlying one about the nature of their society, then, in a reflexive displacement of this analysis, we may question the researcherâs (our own) activities in producing a text about these others. Are researchersâ activities and results also really carrying a deeper message, not about those they study, but about themselves and the nature of their own society? Gudeman and Penn (1982) argue, for example, that the so-called local models developed by ethnographers are no more local than is the interpretative model through which their analyses are constructed. This latter model (which they call a Euclidean model) is simply another local model, one based in the Western cultures of the ethnographer, but one with universal pretensions. The question of how proponents of a local model develop and sustain such pretensions to universality is clearly a political one, having to do with differential access to power. In this light, the research process is more clearly perceived as an encounter in which knowledge is constituted in ways which reflect and maintain various power relations, a process with ethical implications to which I return in Chapter 3.
This more radical reflexivity thus contends that the activities and results of social research are constructed from and reflect both the broader sociohistorical context of researchers and the disciplinary culture to which they belong. It must be accepted that âanthropology is a part of itself. Any statement about culture is also a statement about anthropologyâ (Crick 1982b: 307). However, the fullest expression of reflexivity in research is realized when the âalsoâ in the above statement is dropped, and it is argued that social research is essentially about itself. At this point, it ceases to be research or to promote the fieldwork activities usually taken as constituting ethnographic research. We do not undertake to travel great distances, to situate ourselves among other social groupings, to talk to other individuals simply to learn about ourselves and our own cultures. Such activities might be pursued haphaza...