The investigator will undertake research on the history of statistics during the first half of the twentieth century. The research will examine major lines of statistical development through the works of Karl and Egon Pearson, Ronald Fisher, William S. Gosset, Jerzy Neyman, Harold Hotelling, Harold Jeffreys, Jimmie Savage, and Abraham Wald. The emphasis will be on the development of concept and on the interplay of theoretical development and applications in the physical, social, and biological sciences. Particular lines to be pursued include the development of parametric estimation, the emergence of mathematical statistics as a specialization, the relationship of Neyman-Pearson theory and Fisherian statistics, the relationship of the theory of games and statistical decision theory, the development of multivariate analysis, modern Bayesian statistics, and categorical data analysis. The research will be based upon detailed study of original sources, both published and unpublished. It is expected that the research will yield a greater knowledge of the historical development of modern statistical methods, and thereby a deeper understanding of the properties and limitations of these methods. For example, by studying the spread of multivariate analysis from physical science to social science, where the same numerical algorithms are employed but with an entirely different conceptual framework, some light will be shed on the current debates on the appropriateness of such applications, including the use of regression to assess evidence of discrimination in hiring, the evaluation of educational programs, and the estimation of census undercount. By understanding the role of statistics in past questions of public policy, including temperance, immigration, and military affairs, we will be better able to assess the statistical component of modern policy questions.
