The Relation Between Rao's Paradox in Discriminate Analysis and Regression Analysis

Rao has commented on an apparent paradox in discriminant analysis in which two variables discriminate the groups when used separately, but not in combination. It is shown that there is an exact relationship between this an the paradoxes in regression analysis. An expression for the F statistic in discriminant analysis is given in terms of the average of squares of t value to clarify this relationship.