Index: An Informant-Defined Experiment in Social Structure

This paper describes an informant-defined experiment, designed to answer the questions "whom does any informant know, and why?" Each of 50 informants was allowed to ask an unlimited number of questions about each of 50 target persons (all mythical, each with a created life history). When informants felt they had enough information, they told us which of their acquaintances. was most likely to know the target person, or, more precisely, could serve as the first step in a chain of acquaintances. The data show thatfour basic questions (location, occupation, age, and sex) account for more than 50 percent of questions asked, and a basic collection of six or seven questions would suffice for most circumstances. Less often used questions tend to be employed only when the basic set produces no useful infonnation for that infonnant. The (verbal) reasons given for a choice could be succinctly defined using no more than four concepts. Analysis of these reasons shows that the basic pattern occurs throughout: location is important when the target is near, in a big town, and has a low occupation rating; occupation is important when the reverse holds. In this paper we report the results of an experimental investigation of social structure. For our purposes, understanding social structure requires two essentially different kinds of information. First, we need to know, on average, how many people are known to each person in a group (such as in the U.S.), and who they are. This would provide a description of social structure. Second, we need to know how people think they are related to the people they know. This would provide an explanation of the description (at the cognitive level, at least).