Identifying regions of significance in ANCOVA problems having non‐homogeneous regressions

Abstract

Although the Johnson‐Neyman problem has been well defined in the literature, computational problems have restricted its use by researchers. The computational problems can be simplified considerably by defining the analysis of covariance model in terms of the general linear model in which the contrast matrix for assessing group effects holds the unknown values to which groups are equated. A solution for the unknown values or boundary for the region of significance can be obtained by using the symbolic processing, or three‐dimensional graphing and contour plotting capabilities of the Mathematica computer software package.