Participant Observation

Participant observation is a method of data collection in which the investigator uses participation in an area of ongoing social life to observe it. The investigator may or may not already be a natural member of the social setting studied. The classic form that such participation takes has been relatively [Page 798]long-term membership in a small face-to-face group, but it can also extend to the anonymous joining of a crowd; the boundary between participant and nonparticipant observation is thus not a clear one. Raymond L. Gold's (1969) distinctions among possible field roles—from complete participant to participant as observer/observer as participant/complete observer—have been widely used to indicate the range of possibilities. What participant observers actually do can vary considerably and may include elements sometimes counted as separate methods in their own right such as unstructured interviews, observation of physical features of settings, or the collection of printed materials.

It has become conventional to classify participant observation as a qualitative method, but its data can sometimes be quantified, even if the circumstances may often make the collection of standardized material difficult. Equally, it is conventionally assumed that the observer does not set out to test hypotheses but enters the field with an open mind to see what life is like. The mode of data collection as such does not prevent the testing of hypotheses, but the assumption is that adequate hypotheses cannot be formed without close familiarity with a situation and the meanings of members in it. Participant observation has mainly been used for studies of groups in work settings or of informal groupings in small communities or the social life of groups of friends; these are situations in which patterns of face-to-face interaction develop, and a single researcher or a small group of workers can gain access and establish a role. However, there also, for example, have been studies of football crowds, right-wing political groups, and schools. The method is not well suited to the study of large populations or to the collection of representative data about causal relations among predefined variables; the logic of the case study rather than of the random sample applies.