ABSTRACT
The original occasion for this chapter was a reflection on the use of sample survey methods in the field: that is, the use of structured interview schedules, probability samples, etc., in what is usually thought of as a participant-observation study. The simplest events are customarily described in statements predicating a single property of a single object at a particular time and in a particular place. The filed methods are classified into only three broad classes: participant-observation, informant-interviewing, and enumerations and samples. Certain combinations of method and type of information may be regarded as formal prototypes, in the sense that other combinations may be logically reduced to them. An "average" of a distribution is sometimes obtained not by asking for an enumeration by the informant, nor even by asking a general question concerning what people typically do; sometimes it is obtained by treating the informant as if he were a "representative respondent.".
