PREDICTION OF GRADES IN COLLEGE MATHEMATICS COURSES AS A COMPONENT OF THE PLACEMENT VALIDITY OF SAT‐MATHEMATICS SCORES

Abstract

ABSTRACTBecause the Scholastic Aptitude Test Mathematical Score (SAT‐M) is a measure of general mathematical aptitude that is not closely linked to the specific skills that are prerequisites for a particular course or series of courses, it should not be the only information considered in making placement decisions. Nevertheless, if it can be shown that the SAT‐M score is a reasonably good predictor of success in particular mathematics courses, then it may have a role to play as one of the pieces of information to be considered. The question of the predictive validity of SAT‐M performance was addressed by collecting grades in freshman mathematics courses from 10 colleges. Compared to tests that were specifically designed for placement purposes, the SAT‐M score was a relatively poor predictor of grades in most courses. Even after correcting for the considerable range restriction that may occur when within‐course scores are analyzed, coefficients were typically only in the mid‐.30s (compared to corrected coefficients for a local placement test that ranged from the high .40s to the low .60s). Nevertheless, the SAT‐M score significantly improved predictions from high school grade‐point average alone, especially for calculus courses.In courses at all levels, grades of males and females were very similar, but the SAT‐M scores of males were significantly above the scores of females. Gender differences for predicting grades in particular mathematics courses generally could be eliminated or greatly reduced by considering high school GPA together with SAT‐M scores. Three subscores (Algebra, Insight, and Routine) were generated from the November 1985 SAT‐M test but they had no differential utility in prediction. No practically significant trait‐treatment interactions were demonstrated.For algebra versus precalculus course placement decisions, a weighted composite of SAT‐M score and self‐reported high school grades and courses taken correctly predicted 64 percent to 71 percent of the actual placements. For precalculus versus calculus decisions, about 80 percent of the actual placements could be predicted.