REPORT ON THE STUDY INTERCORRELATIONS OF ENTRANCE TESTS AND GRADES, CLASS OF 1947 AT THE U. S. COAST GUARD ACADEMY

Abstract

ABSTRACTA table was constructed giving the intercorrelations of seventy‐six measures made on the class of 1947 at the II. S. Coast Guard Academy. The Academy and the Educational Testing Service cooperated in the completion of this table. The study covers grades, entrance tests, and some personality ratings. The table is so enormous that it has not been reproduced with this report, but two copies have been transmitted to the Academy. In general it shows a high correlation among grades and good prediction of grades by academic entrance tests. The non‐academic entrance tests, however, do not correlate with grades or personality ratings.The recommendation is made that the data on the Class of 1947 should not be used for the computation of suitable weights for the selection of an entrance battery: selection based on the mathematics tests was carried out to such a degree that their correlation with grades is so greatly reduced that any true correlation bsoomas submerged in chance variations. Two suggestions, however, are given: (a) Using a more difficult mathematics test for entrance purposes would probably make it possible to evaluate the test in spite of selection. (b) Although precise weights cannot be determined, the English tests and scientific aptitude tests should be weighted more heavily relative to mathematics.A discussion is included of the correlations of each group of measures used.