ABSTRACT
In fields like education and therapeutic medicine, the processing or 'reduction' of observations is now a well-developed set of methodologies associated with principles of mathematical probability. The basic tools arose in sciences of the observatory, where they achieved a reasonably standardized form in the early 19th century. Quetelet regarded social science as allied to astronomy and meteorology, a science of the observatory and Stephen Stigler has pointed out that his social science was more nearly social meteorology than social physics. The statistical model of social order was in many respects liberal in its implications. A new book on historical sociology of censuses criticizes the familiar view associated with Michel Foucault that statistics has served primarily as a way of projecting the power of the state over populations. In relation to political forms, statistics has a rather complex legacy. The census grew up in most states as a form of public information or collective self-knowledge and a democratic ritual.
