The Validity of New Tests for the Performance of College Students With High‐level Aptitude

Abstract

This is a validity study of 15 tests tried out in an attempt to improve the capacity of the College Board tests to predict the college performance of highlevel students. 'llhe tests were intended to challenge such students in a way that a mere increase in the difficulty of present tests probably could not do. The experimental tests were named as follows: Six of these tests were given at each of 11 very high-level colleges, and their validities for freshman grade average and individual freshman grades were compared with the validities of the SAT, achievement tests, and high-school rank.Logical, though imprecise, corrections were made to avoid distortion of the observed validities coming from variabHity in the predictability of the criterion at particular colleges where a test was administered, and corrections were made for the extreme restriction in range which occurred at these highly selective colleges. The validities were averaged across colleges, and multiple The recommendation is made that Insightful Reasoning, Data Sufficiency, General Infol~lation,and Short-Item Information shotUd be considered for occupying substantial sections of the SAT, since this would increase validity, while not raising administrative problems or adversely affecting teaching in the schools.-iii-

Multiple-choice questions and student characteristics.

“…These same reviewers also examined the research relating high aptitude test scores and grades in college and graduate school (e.g., French, 1963;Holland, 1959;Lannholm & Schrader, 1951;Whitla, 1962). They observed that when the differing quality of school awarding the grade was taken into account, academic performance was usually moderately related to test score.…”

The Use of Multiple Moderators in Academic Prediction1

“…Three predictors were used: undergraduate grade-point average (U-GPA; converted into standard scores within each law school to a mean of 500 and a standard deviation of 100), the Law School Admission Test (LSAT), and French's (1963) Insightful Reasoning test.…”

Multiple‐choice Questions and Student Characteristics1

“…These same reviewers also examined the MULTIPLE-CHOICE QUESTIONS AND STUDENT CHARACTERISTICS 233 research relating high aptitude test scores and grades in college and graduate school (e.g., French, 1963;Holland, 1959;Lannholm & Schrader, 1951;Whitla, 1962). They observed that when the differing quality of school awarding the grade was taken into account, academic performance was usually moderately related to test score.…”

“…They observed that when the differing quality of school awarding the grade was taken into account, academic performance was usually moderately related to test score. In two of the studies (French, 1963;Whitla, 1962), the regression of grades on SAT scores was determined for the freshman classes of a group of highly selective colleges. Contrary to the critics' position, these regression lines did not taper off or decline for students with the highest SAT scores.…”

“…For the SAT-V items the regression de-'1 It is possible to gain some information about the effeets of reliability on the differences between these correlations by using a lower bound estimate of grade reliability from research by French (1963). For his study of the validity of a new set of aptitude tests, French acquired freshman course grades as criteria.…”