An Analysis of Rates of Change in Community Per Capita Income by Discriminant Analysis

Abstract

Statistical methods for estimation, hypothesis testing, and confidence statements are based typically on exact specification of the response variates. In the applied sciences another kind of multivariate problem is common in which an observation must be assigned in some optimal fashion to one of several populations. Classification rules based on an index called the linear discriminant function provide a method for such assignment.Use of the linear discriminant function is relatively new to regional economics. Previously it has been used in such disciplines as botany to classify a new specimen as belonging to one of several recognized species of a flower, in educational psychology to develop rules for admitting applicants to college programs, in routine banking to aid credit officers in evaluating loan applications, and in agricultural economics to determine producer plans for changes in hog marketings and to identify factors associated with watershed development.