The paper sets out to examine the role that ethnographic work can and should play in the development of sociological theory, focusing on the case study of differentiation-polarisation theory. It provides a detailed discussion of the work of Hargreaves (1967), Lacey (1970) and Ball (1981) and assesses the degree to which their work was ethnographic in contemporary terms. It argues that the model of theory development they offer does not need to be understood in the manner adopted by Hammersley in his account of their work as a model for theory development and testing in the sociology of education. Rather it requires the ethnographer to be more attuned towards setting and maintaining a theoretical agenda, by (a) being more preoccupied with refining existing or established theoretical ideas and concepts and (b) retaining the capacity for the fieldwork setting to inform and direct the study. (Contains 4 notes.)