Archival Experiments with College Admission Policies

Abstract

Colleges and universities in the 1970's will face a problem they would have never expected a decade ago. The problem is that of too few students. Many admissions officers, who for years have needed to screen out students may now seek additional students. The latter, while previously inadmissible, would nevertheless succeed in college. This requires experimenting with alternative admissions procedures. The goal of the present investigation was to estimate the effects of selecting students on the basis of their high school grade point average as compared to the present method of using high school average (HSA). (Academic classes are English, Mathematics, Natural Science, History, and Foreign Language.) The general logic underlying the use of academic HSA is twofold. First, it is believed that academic classes are more similar in content to college classes than are such nonacademic courses as automobile shop or home economics and, therefore, better predictors of college success. Second, high school students receive higher grades in nonacademic classes (as compared to academic classes) which might imply that the regression equation for college GPA on total HSA would have a lower intercept than for academic HSA. Hence, the expected college GPA for a given HSA should be lower for total HSA than for academic HSA. It is true that HSA is usually the best single predictor of college GPA with a correlation coefficient of at least .5 usually reported (e.g., Baird, 1969; Brown & Wolins, 1965; Humphreys, 1968; Lewis, 1966; Richards & Lutz, 1968). However, there is a scarcity of studies which investigate the validity of the components of HSA, specifically grades achieved in the different types of classes taken in high school. While HSA is a composite which can be computed in alternative ways, few studies even mention how this composite was obtained. Gelso and Klock (1967) report several studies in which total HSA was compared with academic HSA as a predictor of college performance. These indicated that total HSA was as good a predictor as academic HSA. In several investigations, high school grades in specific subject