Inconsistency in the predictive value of a battery of tests

Abstract

The Schools of Dentistry selection battery showed some predictive value when studied for a class admitted in 1948, but practically no predictive value for the 1949 freshman class. This study illustrates the need for more careful scrutinizing of the relationship between admission tests to an educational program and subsequent grades. Factors to consider: (1) correlations between scores and grades may be zero although the instrument is valid for selection purposes; (2) the process of selection may control the size of validity coefficients and should be properly controlled in any experimental validation of a test. (PsycInfo Database Record (c) 2025 APA, all rights reserved)