Do the social sciences create phenomena? The example of public opinion research

This paper is an investigation into the philosophy and the history of the social sciences. Some philosophers of the social sciences have suggested that a key feature of the natural sciences is their capacity to create phenomena, and that the social sciences do not meet this criterion. We suggest, to the contrary, that the social sciences can and do create phenomena, in the sense of new ways of describing and acting that have been used to produce all sorts of effects. Like the natural sciences, the social sciences create their phenomena through the procedures that are established to discover them. But the creation of phenomena is a complex, technically difficult and contested process and its success rare. Historically, this argument is developed through a case-study of the development and evolution of public opinion research in the USA and Britain. We argue that by the 1950s public opinion produced a version of the world that had entered 'into the true'. Special attention is given to technical considerations in the development of public opinion research, especially the genealogy of a particular research technology, that of the representative sample. Whilst we are not concerned with demarcation criteria, we argue that there are some important differences between the social and the natural sciences; that the former have a less concentrated 'spatial mix' and a slower 'tempo of creativity'. None the less, in this particular case, the social sciences have played a key role in the creation of opinioned persons and an opinionated society.