The rationale behind the use of observation in sociological research is that the sociologist should become party to a set of social actions sufficiently able to be able to assess directly the social relationship involved.The degree of involvement may vary considerable from being merely a watcher on the sidelines to be deeply involved in and being a part of what is going on. The former type of observation techniques are called non-participant while the latter is called participant observation. Sometimes one way observations screen have been used to watch groups in actions that they are unaware that they are being watched and the observer cannot affect their actions by his presence. The sociologist is visibly present and is a part of the situation either as a sociologist or in another guise. Where the sociologist is merely an observer it is usually assumed that he knows enough about what the actors are doing to be able to understand their behaviour.
Any sociological observer has then to some extent be a participant observer he must at least share sufficient cultural background with the actors to be able to construe their behavior meaningfully but the degree of participation and of sharing of meaning may vary considerably. Examples of such studies are Nel Anderson's study of Hobo-Indians and William White study of Street Corner Society.