New Statistical Methods in Public Policy. Part I: Experimentation

Abstract

As quantitative methods have been applied more and more to public policy work, the whole panoply of statistical devices becomes appropriate for the policy student. Fairley and Mosteller (1977) have recently provided a book illustrating the applications of many statistical methods to problems in public policy. We cannot treat all of these here. We choose to discuss two areas in some detail. The first deals with data gathering—more specifically experimentation and its importance in policy work. We do not discuss how to do it, but we do try to give an empirical base for understanding the need as well as limitations.

Second, we treat exploratory data analysis. We try to give enough of the flavor of the methods and of the ideas behind them that readers will have a running start when they turn to more comprehensive works, and so that they will be able to follow the reasoning associated with some exploratory analyses. We have not here carried out large-scale practical applications, because we have to exposit the methods themselves and their rationale.

Editor’s Note

This article by the Mostellers has been divided into two parts. The first discusses experimentation and Part II discusses exploratory data analysis.