Thomsen, S., Straubhaar J., Bolyard, D.  (1998)  'Ethnomethodology and the Study of Online Communities: Exploring the Cyber Streets', a paper presented to IRIS'98  (Internet Research and Information for Social Sciences) [online]

This is largely a methodological piece discussing the ways in which social scientists might use research techniques to investigate social interactions in online communities. The piece begins by arguing that online cyber communities are 'a primary form of social interaction for the growing number of individuals who often spend hours each day surfing the Net' [nb page numbers for quotes are inappropriate for online sources]. Some examples are given of virtual communities are shared by disabled internet users, nurses, specialized newsgroups, and public relations practitioners. Clearly, the problem is whether it is appropriate to use techniques designed primarily to research interactions in 'real' communities -- what would count as participant observation, or immersion in the community, for example?

It is first necessary to insist that online communities are as real as a face-to-face ones, despite the traditional primacy given to face-to-face. Calhoun is cited as a person who would disagree and would argue that online communities are merely imagined/imaginary. However, other writers argue that online communities offer the nearest equivalent to gemeinschaft in modernity where sociality is centred around either work or home. In these circumstances, the feeling of participating in an online community can be the same as actually going into a pub or cafe for a chat.

However, it is important to rethink past assumptions about human interactions. Specifically:

Social interaction needs to be redefined. It was thought that face-to-face was essential to have a warm and rich interchange with somebody, but, of course copresence alone does not guarantee intimacy  (as in crowds, in clubs and so on). On the other hand, electronic communication can feature considerable self disclosure and intimacy -- the example here is a father of seriously ill child confiding in fellow sufferers. Another example takes on the tone of a family letter, the authors suggest.

Social bonds are also not only formed in face-to-face situations. One study of members of news groups found that they had formed personal relationships with each other, that these had developed over time, and had been continued through more familiar forms of communication such as phone and mail. It may simply be that building a relationship online takes longer and requires more effort to actually find out about the other person. One correspondent to a news group found that the electronic interaction had broken the ice prior to a face-to-face contact and had established his credibility.

Empirical experience may also reflect these new forms of relationships. Virtual encounters often spread into actual encounters and affect interaction in the face to face world. Electronic networks become simply another option to supplement social networks.

Methodological implications follow. For example, ethnography is classically associated with rich encounters face-to-face, where the researcher learns the particular way of life, and is able to write about it from the inside  [US references to ethnography are provided here, including the legendary Geertz]. It is true that textual materials are available in the form of online lists and messages, and that ethnographers have long analyzed texts. But it is only texts that can be studied, which is both good and bad -- 'there is less for the ethnographer to miss', but textual analysis clearly has limits.

One way to proceed might be by performing a content analysis to code the large amounts of text that might accumulate. This might be fairly simple initially, and it clearly requires training and rigour. One approach might be to try to code messages  'on the basis of form, function and content', for example, which would indicate the types of messages that are common and how they are used. Other techniques might be to total up the genders of the participants, and look at the topics that are initiated, although, of course genders can be concealed  [Herring has performed some work which suggests that gender characteristics are not that easily concealed in online work -- see this article].

More qualitative possibilities arise from seeing online communication as dialogue. Listening to dialogue is classic ethnography, of course. Again, several methodologies are available, including those that look at patterns, rules and procedures in conversations. Thus he might be possible to see if jargon or groups speak emerges, whether groups develop their own emotive cons, how groups might be stratified between newbies and veterans, and how conversations might emerge over time, using the facility that many online discussions have to trace threads. Thus it might even be possible to research classic concerns such as turn taking  [see Button and Casey for an analysis of telephone conversations and how they proceed]. The language a newsgroup uses can also reveal certain codes, such as  'the root metaphors used in conversations'.

How might the validity of any research online be established? Classic ethnography relies on techniques such as immersion, empathy, and a rather mysterious way of claiming that one can observe the world as a member of the group does. Other ethnographers insist on criteria which mention different kinds of validity, including being able to make plausible and consistent reconstructions of accounts, or performing 'member tests' [where actual respondents are told what ethnographers make of them and can then agree or disagree]. Prolonged engagement is usually required, especially if one wishes to become a kind of reflective member of the group. Online research requires similar tactics, to try and enter the group, get to know people, create then test analytic models and so on  [while avoiding the twin opposite perils of generating observer effects on the one hand or going native on the other -- online communication seems quite promising in that it is quite possible to lurk unobtrusively, although sometimes lurkers are rounded on?]. It might be necessary to avoid the temptation to simply access all written records in a short a time as possible -- but it is still necessary to learn the codes.

Persistent observation is needed to achieve this social learning. Isolated exchanges alone are insufficient in revealing the past history and context of the exchange  [their indexicality as ethnomethodologists might call it -- mind you, this is a problem for face-to-face ethnographers as well]. Thus experienced newsgroup participants suffering from the same disease rapidly develop their own particular jargon, medical abbreviations, and highly coded references to medical procedures. The authors find that what looks like neutral descriptions of these procedures or conditions carries particular meaning for those in the know, so that describing the size of a tumour clinically is also a way of indicating whether or not one has a reasonable chance of survival, what sort of support one would like from the group and so on.

It is quite possible for an ethnographer to conceal their identity effectively on line, as it is for any other participants, but the authors find that  'online community seem to guard membership identity as closely as any. The authors have all experienced 'hostility... when someone figured out that we were studying the group, even if we were participating effectively as members'. It is still necessary to build trust [and to face all those dilemmas like how much one discloses to the group and so on]. Clearly, some well-bonded groups could easily identify outsiders. One obvious technique here is to use a key informant to guarantee and facilitate access. One way of detecting such a person might be to see if particular members of the group dominate the interactions and discussions.

The paper ends with noting that this is a large area requiring further discussion with other researchers. In the short term, they 'propose a multi-method approach... that involves the use of text and discourse analysis, a prolonged commitment to involved participant-observation, and the use of qualitative interviews with group members'. It is still necessary to develop the skills of being able to penetrate the community, and become part of it, and this is as difficult with online communities as it is with some face-to-face ones. Qualitative interviews can also be used as a form of triangulation -- in the example cited, some professional practitioners claimed that their main interest in online participation was purely professional, but qualitative interviews picked up the importance of increasing self-esteem and self validation as well [nice idea here -- online participants can try out their professional identities, and gain confidence before they meet face-to-face. I have met distance education students who have followed the same career]. Member checks might be used as a test of validity.

Finally, there are ethical issues, arising from the ambiguous status of newsgroups -- are they public or private discussions? Perhaps analysing conversations like this is unethical. However, full disclosure can produce an observer effect  [again I'm not sure that this ethical issue is confined only to online observation, of course].

back to key concepts