ABSTRACT
Surveys are perhaps the most familiar method of social research, and for many people, they are the most ‘scientific’ form of investigation. They produce numerical descriptions of the social world (eight out of ten cats prefer Whiskas, 20 per cent of the population believe x, 50 per cent have never done y) and surveys often receive such wide publicity that their findings become part of people’s daily knowledge about the world. Surveys on sexual behaviour are no exception. Readers of this book will probably have heard it said, for example, that on average young men think about sex once every eight minutes, and that the average erect penis is six inches long. It is the survey method which produces this type of numerical ‘fact’, for the survey method generates a very particular sort of data. Where other research methods produce unstructured data (perhaps in the form of field notes written by the researcher, or in the form of taped conversations) the survey method produces a structured or systematic set of data which de Vaus (1991) calls a ‘variable by case data matrix’. The researcher begins by listing the variables with which she or he is concerned, and then systematically collects information about each of these variables.
