SOUTHERN JOURNAL OF AGRICULTURAL ECONOMICS JULY, 1978 AN ANALYSIS OF RATES OF CHANGE IN COMMUNITY PER CAPITA INCOME BY DISCRIMINANT ANALYSIS

sis testing, and confidence statements are FUNCTION FOR TWO GROUPS based typically on exact specification of the re-sponse variates. In the applied sciences The technique of discriminant analysis is another kind of multivariate problem is com- based on the assumption that a linear function mon in which an observation must be assigned Y = BIX + B 2X 2 +... + BnX n exists which in some optimal fashion to one of several popu- will distinguish between the elements of a pop-lations. Classification rules based on an index ulation. The discriminant model utilizes coeffi-called the linear discriminant function provide cients B,, B 2,..., B chosen in such a way that a method for such assignment. the ratio of between-group sum of squares is Use of the linear discriminant function is maximized. Therefore, the index Y represents relatively new to regional economics. Previous- the optimal discriminator between the two ly it has been used in such disciplines as groups. Factors X,, X2,..., X represent quanti-botany to classify a new specimen as belonging fiable determinants of income changes. to one of several recognized species of a flower, Several computational approaches are avail-in educational psychology to develop rules for able to use in the discriminant procedure [61. I