When converting multiple linear regression results to discriminant-function values, the form of coding used in the dummy Y variable will affect scale and location constants in the conversion factor. This note provides two additional conversion procedures to those five given in a previous article. They are for [1,—1] and [1,2] coding of the vector Y, respectively. A computational example is given.
Get full access to this article
View all access options for this article.
References
1.
CramerE. M.Note: equivalence of two methods of computing discriminant function coefficients. Biometrics, 1967, 23, 153.
2.
GockaE. F.Relating two-group discriminant function and multiple regression results for classroom demonstrations. Perceptual and Motor Skills, 1978, 46, 971–975.
3.
HealyM. J. R.Note: computing a discriminant function from within sample dispersions. Biometrics, 1965, 21, 1011–1012.
4.
LiC. C.Introduction to experimental statistics. New York: McGraw-Hill, 1964.
5.
TatsuokaM. M.Multivariate analysis. New York: Wiley, 1971.
