exploring the relationship between social science and software development methodologies: a blog by Pascal Belouin

During the course of fieldwork, the ethnographer can take on a number of different roles, which could be classified in regards to the degree of one’s involvement with the population he or she is studying. For instance, according to Junker (1960) and Gold’s (1958)
model, the roles available to the researcher range from the ‘complete participant’ to the ‘complete observer’.

Although ‘extreme’ roles such as these can be useful in certain cases in regards to the practicalities of a particular context, they also come with their own philosophical assumptions, limitations and potential for bias, which all need to be carefully taken into account in the further steps of the research such as data analysis. Indeed, more ambivalent roles such as the ‘observer-as-participant’ or ‘participant-observer’ could allow the collection of richer data and provide a more interesting perspective for its analysis, allowing the researcher to keep a certain level of ‘marginality’ as defined in Hammersley and Atkinson (1995)and permitting the researcher to maintain a middle-range position between involvement and detachment.
This idea of a middle-course position of the researcher could also be put in parallel with Schutz’s metaphor of the stranger in the way that, in certain cases, the researcher needs to distantiate himself from his own culture (Hammersley and Atkinson, 1995, p. 8).

The ‘complete participant’ role has the advantage of allowing the researcher to have an experience as close as possible to the people he or she is studying. This role has practical advantages that could prove useful in terms of access, and could be recommended for someone practicing covert research. However, one of its main downsides is that one has to be particularly cautious not to become too close to the people one is studying in order not to lose any necessary detachment towards one’s object of study by ‘going native’ (Hammersley and Atkinson, 1995, p. 10), which could have an impact on the data analysis process, noticeably in terms of bias. This particular issue, which could also affect researchers taking on more ‘detached’ roles, is rather well illustrated by the case of Willis (1952), whose identification with the young working class boys he was studying impaired his analysis skills. Moreover, there is a chance that the researcher’s participation could influence the activities of the people being studied, therefore distorting the collected data.

Diametrically opposed to this approach, the ‘complete observer’s’ total detachment towards its object of study may seem as a good way to achieve objectivity, and could be rather appropriate in certain cases, such as Corsaro’s (1981) study of nursery school children. Furthermore, the fact that researchers taking on the role of a complete observer are removed from the social context they observe limits the impact the researcher could have on the activities of his object of study. However, one could argue that a total lack of involvement with the activities of the people being studied limits the richness of the data collected in this way. Furthermore, such an approach is likely to be associated with a certain empiricist approach to knowledge construction, a perspective that could be challenged by other theoretical standpoints such as idealism or social constructionism.

Between those two rather radical stances are more contrasted roles, namely ‘observer-as-participant’ and ‘participant-observer’. The latter requires from the researcher to establish himself first as a participant, but in the same time engaged in doing research, while the ‘observer as participant’ role implies that the researcher is first and foremost seen as an observer who could be engaged in the activities of the people being studied. One of the main practical advantages of this type of roles is that the researcher does not deceive the people being studied; he or she can therefore establish a relationship of trust with the participants, which is less easy to achieve when taking on the role of complete participant or observer. However, taking on such roles can also entail some problems: Indeed, participants knowing that they are being observed may – consciously or unconsciously – alter their behaviour in the presence of the researcher, and try to manipulate him. Moreover, strong bonds between the ethnographer and the people he or she is studying can give rise to a problem of ‘over-rapport’ as defined by Miller (1952) by altering his or her judgment and influence his or her analysis and findings.

The fact that a researcher adopts a particular role can also be strongly influenced by his or her epistemological assumptions or filiation. For instance, a researcher influenced by the positivist or, more widely, empiricist current of though would be more likely to take on the role of the ‘complete observer’, arguing that detachment is necessary in order to obtain closure and therefore objectivity. On the other hand, researcher more influenced by idealism or postmodernism, valuing subjectivity as a mean for understanding a social situation, would tend to use methods allowing a deep immersion into the social context being studied. However, one does not need to be from a particular ‘chapel’ to take advantage of the full range of roles made available to the researcher in order to cover as many aspects as possible of his or her object of study.
Indeed, one of the many interesting aspects of fieldwork is that the wide range of roles available to the researcher can be used alternatively in the same research context, as a mean to deepen one’s understanding of one’s object of study. Several examples from the literature, such as Sevigny’s (1981) study of art classes tend to suggest that it can be extremely useful to switch between different roles in order to gain different perspectives in regards to a particular social phenomenon. Changing roles could also be seen as a strategy for spending more time in the field, or to apprehend one’s object of study from different perspectives.

Aside from Junker and Gold’s broad categories are more subtle ways through which the researcher may be able to present him or herself to the people he or she is studying. Indeed, one could find the boundary between Junker and Gold’s roles rather unclear; the account of the participant observation activity presented in the first part of this essay is a good example of the difficulties one could have to situate oneself in a clear-cut manner in one of these categories, and emphasises the crucial importance of reflexivity in ethnographic research.


Corsaro W.A. (1981) Entering the child’s world: research strategies for field entry and data collection in a preschool setting. In J. Green & C. Wallat (Eds.), Advances in Discourse Processes: Ethnography and Language in Educational Settings v. 5. New Jersey: Ablex.

Hammersley M. and Atkinson, P. (1995) Ethnography: Principles in Practice, London, Routledge.

Miller, S. M. (1952) The participant observer and “over-rapport.” American Sociological Review, 17, pp. 97-99.

Sevigny, M. (1981) Triangulated inquiry: A method for the analysis of classroom interaction. In J. Green & C. Wallat, (Eds), Advances in Discourse Processes: Ethnography and Language in Educational Settings v. 5 (pp. 65-86). New Jersey: Ablex.

Taylor, S. and Smith, M.J. (2008) D844 Study Guide, Milton Keynes, The Open University.

Willis, P. (1977) Learning to Labour: how working class kids get working class jobs, Farnborough: Saxon House.

Related posts:

  1. Research proposal : A Foucauldian analysis of the evolution of the discourse about software development methodologies
  2. Conversation Analysis: a short introduction
  3. Applying social scientific concepts to domain definition: a short overview
  4. Social science for software developers – Using tools from social science to inform software design: should software developers also be social scientists?
  5. An epistemological critique of Grounded Theory
§125 · March 16, 2010 · Practice · Tags: , , , , , , , · [Print]

1 Comment to “Ethnography: a short description of the roles available to researchers in the field”

  1. Charles says:

    With regards to the complete participant idea: as an ethnographer, Ive had to deal with a lot of criticism involving bias in my work, since I approach it as not just a researcher “going native,” but as a fully native anthropologist. (I chose to study a “subculture” that I am a member of.)

    While I understand many of the risks and potential biases that go along with being a native anthropologist, I also know that the information I can receive from my subjects is also deeper and more fully realized, simply because they know who I am and trust that I am not trying to either stereotype them or twist their words. In addition, I am also well aware of the negative repercussions involved in my research population, and while I am not in the business of trying to denounce them, I also am not trying to “blow them out of proportion.” I feel, in my case at least, that with proper oversight from my department, and a keen eye towards moderation, I can still be a native anthropologist while not succumbing to the biases that go in hand with it.

    That said, I study media culture and otaku, which isn’t a closed off group to begin with- indeed, they are willing to talk to anyone. Also, since the beginning of the field work I’ve found that a lot of the biases and perceptions I’ve had towards my population are just dead wrong. So maybe being native wasn’t such a bad idea.

    I recommend to you Delmos Jones “Towards a Native Anthropology,” if you haven’t read it yet.

    PS- I think your analysis of the fieldwork roles are spot on. Most of my colleagues aren’t so well versed.

Leave a Reply