Year of Publication
2019
College
Martin School of Public Policy and Administration
Date Available
8-15-2019
Executive Summary
The Graduate Record Examination (GRE) is the most widely used graduate level admission test in the world, yet conflicts exist across the findings of many studies of the ability of the GRE to predict the success test takers will have as graduate students or in professional life. Additionally, most of the studies that exist on the GRE’s ability to predict graduate student success use data from a previous version of the GRE that may not be applicable to the current version, thus rendering their use for policy making among makers of admission decisions limited and flawed. These studies also tend to focus on one specific degree program or type of degree, leaving no guidance for how the GRE might be useful for other programs.
In an attempt to provide a basis from which more comprehensive analyses can be modeled in the future, this quantitative study examines how well the three sections of the GRE predict success for graduate students at the University of Kentucky (UK), using the Graduate Grade Point Averages (GGPA) of 2,349 active graduate students in degree seeking programs as a measure of intermediate success, as is common practice in the literature. Two linear regressions are reported, one which includes only UK graduate students in non-STEM programs, and one which includes only UK graduate students in STEM programs. STEM or non-STEM is used to separate these students from one another because it allows an overall picture of how GREs might be useful for admission decisions at UK while still providing results that can be relevant to programs with key differences, as “the economic and social benefits of scientific thinking and STEM education are widely believed to have broad application for workers in both STEM and non-STEM occupations” (Gonzalez and Kuenzi 2012).
Results show that the three sections of the GRE General Test are predictors of success for non-STEM graduate students at UK, while two of the three (Verbal and Quantitative Reasoning) are predictors of success for STEM students. Further, the Verbal Reasoning section is a better predictor of success for non-STEM students than for STEM students, while the Quantitative Reasoning section is a better predictor for STEM students than non-STEM students. The Analytical Reasoning section is found to predict success only for non-STEM students.
These findings do not evaluate whether GRE scores are the best way of predicting the success of graduate students, nor should they be used exclusively to make admission decisions for applicants to an academic program. Instead, it is recommended that GRE scores be used as a portion of a holistic review of such applications, a recommendation which the owners of the GRE also make. According to some of the literature, the GRE puts some minorities and women at a disadvantage for admission to graduate programs. With this in mind, other studies have found the undergraduate GPA of a student is the best predictor of how well they will do in a graduate program, thus it is recommended that programs considering revision of their policies concerning GRE requirements for admission offer waivers for students that have or exceed an undergraduate GPA they deem appropriate to succeed in their program in lieu of removing an existing requirement for the purposes of increasing diversity or attracting more students. Additionally, programs without GRE requirements should consider adding it to the list of what an applicant needs for admission (with or without waivers), as it provides a standardized score that can assist in gauging the cognitive abilities and potential successes of its takers. Lastly, programs that do have GRE requirements and are classified as STEM should place more emphasis on the Quantitative section of the GRE than the others, while non-STEM programs should do the opposite.
Recommended Citation
Floyd, Brenton A., "Studying the Graduate Record Examinations' Ability to Predict Student Success as Measured by Graduate Grade Point Averages" (2019). MPA/MPP/MPFM Capstone Projects. 313.
https://uknowledge.uky.edu/mpampp_etds/313
