Bridges the gap between correlation coefficient and client general expectancy tables

Abstract

Readers are reminded of a "universal expectancy table" developed by Schrader from properties of the bivariate normal distribution. The authors point out some practical utilities of this table in training clinical practitioners - specifically, in developing probability statements about criteria given degree of relationship and place on predictor. Also presented is a modification of Schrader's table based on five levels of z rather than on three levels of percentile rank.