More precise conception of the functions and limitations of prediction, validation, construct validity, and criterion referenced tests will not solve the problems which must be faced in using tests. It will not repel many of the current attacks on testing. But it should enable test specialists to think more clearly about some of those problems.…

The major purpose of the present investigation was to compare the validity of a work sample test and an intelligence test for the prediction of trainability. Several lines of evidence supported the hypothesis that samples of training behavior are the better predictors. (Author)