Estimating the Contribution of Experimental Tests to the Armed Services Vocational Aptitude Battery

Abstract

Abstract : The only way to accurately determine how much an experimental test will add to the Armed Services Vocational Aptitude Battery (ASVAB) in predicting a subsequent criterion would be to administer the experimental test along with ASVAB in the operational setting. This simply is not feasible to do on a routine basis. Ordinarily, experimental tests are administered to individuals already in the service and an attempt is made to determine how much predictive efficiency these tests add to ASVAB subtests administered some months previously. This approach has been justly criticized because abilities could have changed between the time of ASVAB administration and experimental test administration. Because of the availability of ADVAB test-retest data on over 4,000 subjects, it was possible in the present research effort to address the ability changes in terms of a meta-analysis of differences between test and retest validities for grades in 69 technical school courses. It was hypothesized that the retest scores should have higher validities to the extent that abilities had changed. This follows from the fact that retest scores measure abilities at the time of course entry. With the exception of slightly higher retest validities for ADVAB measures of technical knowledge, the test and retest validities centered at about the same level. It was concluded that reasonable trust can be placed in estimates of the contributions of experimental tests to ADVAB in the prediction of subsequent criteria, even though the ASVAB scores are collected some months prior to administration of experimental tests.