Some Statistical Problems in Small Group Research

Abstract

FOR several years at the Laboratory of Social Relations at Harvard we have been engaged in the development of a method for the recording and analysis of social interaction.' By social interaction we mean the face-to-face conversation or behavior of two or more persons as they communicate with each other. There is a rapidly growing emphasis in several of the social sciences on the microscopic study of social interaction in small face-to-face groups.2 Children at play, classroom and discussion groups, committees, planning groups, work groups, therapy groups, and many others are being studied by direct observation. Researchers face a need for standardized methods of observing, analyzing, and comparing the behavior which goes on in widely different sorts of groups under widely different sorts of conditions. Previously separated lines of research and theorizing now seem to be converging toward a coherent body of techniques and theory. The basis of research and theorizing is less and less in terms of special interest in some concrete type of group, such as a particular interest in planning committees as they operate in administrative organizations, or a particular interest in classroom groups. The conviction is rapidly growing that there can be a generalized theory of small group structure and