A study was conducted to investigate the validity of the Armed Services Vocational Aptitude Battery (ASVAB) and of educational data for Air Force technical training, to investigate the unique contribution of both educational data and test data in predicting Air Force technical training success, and to assess homogeneity of prediction equations for sub-groups defined by race and sex. Data were collected by using ASVAB-Form 3 for all Air Force nonprior service enlisted accessions from September 1973 through October 1975. The analyses included 43 clusters of enlisted training courses (for example, intelligence, audiovisual, and weather). Both test data and educational background data proved useful for prediction of Air Force technical training performance; moreover, when used in combination with each other, more accurate predictions were possible than through the use of either alone. Generally, test data alone provided more accurate predictions than did educational background alone. In many instances, separate race or sex group prediction equations were not homogeneous (i.e., the sub-group equations differed from each other enough that added accuracy in prediction could be achieved by using a separate equation for each sub-group). Predictions based on educational information were more susceptible to race bias than those based on test data. (Tables of data appear throughout the report.) (LMS)