Document Type
Honors Project - Open Access
Abstract
This paper estimates the predictive power of the SAT for college success. Most studies use GPA as a proxy for college success, but GPA is subject to differing grading standards across instructors and departments. Using individual course grades of Macalester College students from 1996 to 2005, this paper uses fixed effects to control for students sorting themselves into particular departments and courses. After controlling for sorting by major, a general trend emerges: the Math SAT is more predictive for science and math-related social science majors while the Verbal SAT is more predictive for language majors. After controlling for sorting by course, the predictive power of both the SAT and the Math SAT increases with statistical significance. The results suggest that the Verbal SAT does not carry the same predictive power as the Math SAT. Furthermore, the predictive ability of the high school GPA drops by at least 10 times. The result implies that the SAT is a predictive and reliable indicator for college grades even though the SAT does not predict each major’s grades equally. As a result, the SAT should not be the only emphasis in admission process, but neither should it be ignored.
Recommended Citation
Liu, Yuen Ting, "Does Sorting make a difference in the SAT’s prediction of Course Grades and Cumulative GPA at Macalester College?" (2009). Economics Honors Projects. 19.
https://digitalcommons.macalester.edu/economics_honors_projects/19
© Copyright is owned by author of this document
