The objective in the project reported here was development of methods for sampling and manipulating everyday conversation. Scientific study of social interaction bridges the domains of communications engineers,sociologists, psychologists. Each discipline has its own purpose, but all face a common problem: how can our communicative processes be validly sampled and assessed? In this thesis, the author has suggested a theoretical background against which empirical approaches can be viewed. Previous work has been surveyed to illustrate the types of method which have been used to date. Alterations in method were suggested and demonstrated in the form of an experiment. In this experiment, a method of collecting samples of everyday conversation in the laboratory was compared with one traditional method of stimulating interaction in a dyad (game-playing). The two methods produced measureably different types of looking, speaking,and pause behaviour. Degree of acquaintance of the conversants,and their compatibility on the FIRO-B personality measures, and television as a medium of conversation also affected the interpersonal behaviour of the conversants.