In the past, social research has generally been done by the powerful on the powerless, the poor, the underdog (Becker 1967; Barnes 1979). The collection of social intelligence about the contours of the population was born, with capitalism, in Britain's industrial centres, and it was quickly taken over by the state apparatus. Indeed, the very root of the word ‘statistics’ is the ‘state’ – items of information gathered by German statesmen, to help them rule; the connection between statistics and mathematical theory developed much later. These three historical features of social statistics – its concentration on the powerless, its close link with the capitalist mode, and its centrality to the state – have all had important consequences for the relationship between researcher and researched. Furthermore, as changes have taken palce both in the technology available for conducting research and in the relationship of the state to both the economy and to social relations, so we might expect and indeed do find developments in the way the subjects of research are characterised.

In this chapter, I intend to trace the development of ideas about the relationship between the researcher and the person from whom the researcher collects his or her information. My argument is that there have been three discernible stages in the evolution of the relationship. The first view, predominant in the nineteenth century, was that social researchers should only interview other professionals, acting as informants, giving proxy information to the researcher.