Prediction of second-term grade-point average at the U.S. Naval Postgraduate School

Abstract

In July 1948 twenty aptitude and achievement tests were given to members of the incoming class at the Naval Postgraduate School. These tests, previously validated against the criterion of, first-year quality-point ratios, have now been cor­related with grade data for the first two terms of the second year at the School. Scores on all of the tests previously found to have high effectiveness for predicting first-year academic performance were found to be significantly related to second-year grades. Using a technique developed by Horst, an estimate of' .43 was obtained for the value of the multiple correlation between tests and criterion. Since the second-year validities were consistently and appreciably lower than those for the first year, various possible explanations for the decreases in validity from first to second year were suggested and examined. It was concluded that part of this decrease was due to an observed lower reliability for second­ year grades, and that part was due to the effects of selective drop-out. It was also concluded that in part this decrease was due to differences between the first- and second-year curricula (with abilities required for success in the second year which were perhaps not measured as well by the tests as the abilities required for success in the first year), and to real changes which might have occurred in the students during the elapsed time interval, for example, changes in motivation. Since several of these factors which tend to reduce the test-criterion are in no way faults of the tests, it was concluded that the second-year validities may be regarded as fairly satisfactory. However, further work in the area seems desirable.