The accounts people provide can be an important source of data for ethnographers. These may be produced spontaneously or elicited by the researcher, in the course of participant observation or in formally defined interview situations. Whatever their form, interviews must be viewed as social events in which the interviewer (and for that matter the interviewee) is a participant observer. The ethnographer may be able to play a more dominant role than is the case in other contexts, and this can be capitalized upon, both in terms of when and where the interview takes place and who is present, as well as through the kinds of question asked. In this way different types of data can be elicited, as required by the changing demands of the research. While this feature of interviews heightens the danger of reactivity, this is only one aspect of a more general problem that cannot be avoided: the effects of audience and context on what is said and done.
The accounts produced by the people under study must neither be treated as ‘valid in their own terms’, and thus as beyond assessment and explanation, nor simply dismissed as epiphenomena or ideological distortions. They can be used both as a source of information about events, and as revealing the perspectives and discursive practices of those who produced them. Moreover, while it may sometimes be important to distinguish between solicited and unsolicited accounts, too much must not be made of this distinction. Rather, all accounts must be examined as social phenomena occurring in, and shaped by, particular contexts. Not only will this add to sociological knowledge directly, but also it can throw light on the kind of threats to validity that we may need to consider in assessing the information provided by an account.