The purpose of this paper is to show the effect of using a selection test on the average criterion score of the entering class. The correlation coefficient (times 100) is shown to be the percentage of improvement of using the selection test over what would happen on the average if the test were not used. A simple formula is developed for approximating the probability that selection by chance would yield an entering class as able or abler than a class selected with the help of a valid predictor. It is further shown that, for reasonably sized colleges and with reasonable assumptions, selection using even small correlations will almost certainly result in an entering class that would earn higher criterion scores than would occur without the selection test. (Author)