A Model for Studying the Prediction of Success in Complex Learning Tasks

Up to the present point in this chapter, a very simple model for studying the prediction of success in foreign language training has been assumed. The model assumes that success is a direct function of measured aptitude, plus errors which may exist in the tests or in the criterion; thus, for a given individual, c + e_c = f (a) + e_a , where c is the true score for the criterion, a is the true score for aptitude, and e_a and e_c are symbols for errors in c and a. Aptitude, a, may be regarded as a composite of true scores for various aspects of aptitude. The validity coefficient, i.e., the correlation between aptitude test scores and a criterion measurement, is regarded as an indication of the extent to which errors have been minimized.