Statistical Thought

In this unique monograph, based on years of extensive work, Chatterjee presents the historical evolution of statistical thought from the perspective of various approaches to statistical induction. Developments in statistical concepts and theories are discussed alongside philosophical ideas on the ways we learn from experience. Suitable for researchers, lecturers and students in statistics and the history of science this book is aimed at those who have had some exposure to statistical theory. It is also useful to logicians and philosophers due to the discussion of the problem of statistical induction in a wider philosophical context and the impact of developments of statistics on current thinking The book is divided into two parts: Part I (Chapters 1-4) entitled 'Perspective' deals with foundations and structure and Part II (Chapters 5-10), explores the 'History'. In Chapter 1 statistics is characterized as 'prolongation of induction', and its philosophical background is briefly reviewed. The special features of statistical induction, the two roles (as input and output) the theory of probability plays in it, and the different interpretations of probability are discussed in the next two chapters. Chapter 4 distinguishes broadly between four different approaches to statistical induction (behavioural, instantial, pro-subjective Bayesian, and purely subjective) that have been developed by taking different interpretations of probability as input and output, and considers their comparative characteristics, advantages and disadvantages . Part II traces the historical evolution of statistical thought in the perspective of the framework described in Part I and specifically considers the origin and development of the different concepts of probability and their application to the formulation of the different approaches to statistical induction. After some reference to the prehistory of the subject, the contributions made by the principal contributors in probability and statistics in the 17th-20th centuries are outlined (beginning with Cardano, Pascal, Fermat, Huygens and James Bernoulli and proceeding through Laplace and Gauss to Karl Pearson, Fisher, Neyman, E.S.Pearson,Wald, and their successors). Throughout, the emphasis is on concepts - factual details and technicalities are introduced only if they are unavoidable.