The Evaluation and Comparison of Experimental Designs for Fitting Regression Relationships

Abstract

The purpose of this paper is to show some of the many things one can learn about the adequacy of a proposed experimental plan before any data are collected. These permit one to evaluate whether a suggested design is appropriate and to compare alternative possible designs. The discussion deals with experiments involving controlled variables which are quantitative and which are to be related to the response variable by a least squares regression fit. A recently developed computer program which permits easy implementation of the procedures is indicated. The evaluations are illustrated in detail by a numerical example.